Topic 2: Reliability {by 9/3}

Based on the text readings and lecture recording due this week consider the following two discussion points: (1) Correlation does not equal causation! Share your thoughts on why this assumption is still a common mistake, especially in the mental health field.  (2) Share your thoughts on why reliability is so (wicked) important for psychological assessments (this is a bit of a deep question – give it your best shot).

 

Your original post should be posted by 9/3.  Post your two replies no later than 9/5.  *Please remember to click the “reply” button when posting a reply.  This makes it easier for the reader to follow the blog postings.

86 Comments (+add yours?)

  1. Anne Marie Lemieux
    Aug 29, 2020 @ 21:56:13

    I think it is possible that people tend to assume that correlation means causation to prove their dogmatic theories. They do not take into account the multiple variables that can be contributing to the results. This is especially true in the mental health field. It is easy to see a trend and want there to be causation in an effort to decrease the stigma of mental illness or work to treat the cause. However, without solidifying the results they could be missing the mark. Luckily, more studies are being done around mental illness to increase our understanding of cause and effect. For example, it has been shown that diet has an impact on depression. Sometimes it is evident that there is no causation even when there is a correlation. For example, states that have MLB baseball teams, have lower rates of divorce. It is rational to believe that baseball has nothing to do with divorce but other correlations are less obvious.

    Psychological assessments are not perfect. They are attempting to measure an abstract idea where multiple variables can influence the outcome. Therefore, the higher the reliability the more accurate you can get to a true assessment of measurement. The lower the reliability the less confident you can be of accuracy.

    Reply

    • Destria Dawkins
      Aug 31, 2020 @ 15:22:00

      I agree. It’s like society tries to pull us in the direction to believe that a mental illness is built upon a specific cause, in a rush to treat a patient who is suffering.

      Reply

    • Connor Belland
      Sep 01, 2020 @ 11:57:00

      Interesting points you made Anne Marie. I agree that researchers may assume causation because that’s what they want to see when in reality there is no causation and just a correlation. Going into experiments with pre-determined assumptions and biases from the researchers will make researchers more likely to make the mistake of saying correlation is causality. I also like how you referred to reliability in terms of confidence.

      Reply

    • Beth Martin
      Sep 01, 2020 @ 22:15:16

      Hi Anne Marie!

      I enjoyed reading your response, and I always love a good “this is an absolutely ridiculous correlation” example. I see some people try to reason them out sometimes, which is always fun, and I think something that people do generally with correlation/causation! The MLB teams could possibly be linked to overall happiness (as no one has ever disliked baseball) which drives divorce rates lower… a bit of a weak link, but it’s not completely ridiculous, right? It’s something that sounds plausible at first glance, and that’s why correlation/causation is so dangerous. I’m definitely guilty of seeing information, seeing the proposed causation and trying to rationalize it out, rather than sitting down and going “wait, could it actually be xyz that’s causing this?”. Someone else has done all the hard work for us, and if they present two things that *could* be linked, I think we tend to jump to the easy-to-grasp “this causes that” because very few of us are actually inclined in going any further into researching which particular socio-economic impact of baseball is linked to divorce.

      Thanks for your post!

      Reply

  2. Elias Pinto-Hernandez
    Aug 31, 2020 @ 10:38:33

    (1) Correlation does not equal causation! Share your thoughts on why this assumption is still a common mistake, especially in the mental health field.
    There could be many factors, unknown or unmeasured, that can influence or cause variables to correlates. A variable could cause the action another but not necessarily—for example, more arrest during a full moon than the rest of the moon cycle. Does the moon influence people’s behavior, or are there more police officers during those nights? Correlation does not equal causation, and we should not infer that two or more events are causally connected because they occur together. At the same time, people experiencing two events could link them together and believe that one causes the other and could generate a disorder like the classic example (maybe too old) of the black cat and bad luck. If a black cat crosses in front of someone and something unpleasant or “bad” happened, the individual could correlate the two events and develop a phobia.

    (2) Share your thoughts on why reliability is so (wicked) important for psychological assessments (this is a bit of a deep question – give it your best shot).
    Reliability means consistency, accuracy, standard error, or precision in measurement. Its significance rests in the fact that it must have reliability as an indispensable requirement to consider the tests as adequate and scientific. As reliability increase, standard error of measurement decrease.

    Reply

    • Connor Belland
      Sep 01, 2020 @ 11:08:42

      Good answer Elias! I like your point about how many other variables known and unknown could influence something and not just some random correlation. I also like your example about arrests when the moons out. I also think the correlation/causality mistake is often made when researchers are lazy and look for the easy answers or the answer they want to get whether or not its the right answer.

      Reply

      • Elias Pinto-Hernandez
        Sep 03, 2020 @ 11:42:45

        Thank you, Connor, you see, back home (PR), people are still superstitious, and politicians exploit citizen’s fears. Now and then, I hear the governor with the secretary of any agency or the police chief in a press conference explaining events without conducting research having evidence or data. They just come up with explanations that sound (in their minds) “rationales,” or made sense, such as the correlation of full moon higher crime rate that leads to more arrests. It is evident that the root of the problem is the full moon. Stay home during those nights! Now, when it comes to researchers, I believe that more than lazy, they are opportunists who use the stats to mislead people.

        Reply

    • Bibi
      Sep 02, 2020 @ 15:25:52

      I really liked that you brought nup the other variables in your discussion of correlation and causation that can result in misleading a correlation to be taken as causation. However, I think it would have been a little more helpful to delve more deeply into what this means for mental health, especially when they are a billion variables that are resultant in mental health. Your full moon example was awesome though, I haven’t heard that one before!

      Reply

  3. Destria Dawkins
    Aug 31, 2020 @ 13:54:32

    It is easy to confuse correlation with causation. People tend to confuse the two without even thinking about it. For example, people like to say that smoking causes alcoholism. Smoking does correlate with alcoholism, but smoking doesn’t automatically cause alcoholism. People still mistake the two for the simple fact that people are not generally thinking about the two different terms statistically.

    Reply

  4. Connor Belland
    Sep 01, 2020 @ 10:38:26

    It is important to know that correlations does not mean causation when working in the mental health field, as well as anywhere. I think researchers often confuse correlation with causality because they only see what they want to see. When they have an assumption or bias going into a study then they are going to want to be right. When they are only looking for things to prove them right it is often easier to find relationships in correlations. But they don’t realize that just having a correlation doesn’t technically prove them right it just shows a correlation between the variables when their could be various other variables that have an actual effect or causation on a dependent variable. It is much harder to prove causation so researchers will just look for correlations to save time and effort. In the mental health field if a correlation showed lack of sleep having a positive relationship with depression, and then assumed that was the inly variable that influenced depression, then they would be very mistaken. There are a hundred other variables that influence depression and there is no way of knowing from a correlation if lack of sleep causes depression because its just a correlation.
    Reliability is very important in psychological assessment and just in life in general. People often want to surround themselves with reliable people, so why wouldn’t they want a reliable test. Tests need to be reliable so they can give precise and consistent scores for anyone taking them. A reliable test will be credible and fulfill the hypothesis it was testing for. If we didn’t have reliability in tests, there could be a thousand different types of tests to measure the same thing and no one would know which one to use. It is also important for an assessment to have validity or to accurately assess what it is supposed too, and you cant have validity without reliability. Reliability is just super important.

    Reply

    • Beth Martin
      Sep 01, 2020 @ 22:09:38

      Hi Connor!

      I think you make a really good point on people seeing one variable as the only variable, and it’s not something I’ve actively considered before. I’m a firm believer that we tend to just assume things that are easier for us to digest/conclude, and it makes complete sense that when we do see one variable, we’re more inclined to believe that that’s *the* variable that’s impacting things, not something else present (again, as it’s just the easier default I definitely revert to sometimes). You’re definitely right on the money about causation being harder to prove too, and it’s making me wonder how many things I’ve been told have been “proven” have just been assumed or not fully investigated. I liked your comment about reliable people too! I think it draws a nice, neat, little parallel – we *want* reliable people in our lives, but sometimes just assume our friend group etc is until proven otherwise, and that’s why actually measuring reliability is so important!

      Reply

    • Maya Lopez
      Sep 02, 2020 @ 15:52:06

      Hi Connor,
      I think you raised a very good point that researchers may be a bit blinded by their own agenda at times and as you said are just looking to prove themselves right. The scientific model of proposing a hypothesis and then doing an experiment to prove oneself right or at times, wrong can be very influential. It may be easier to just assume the findings are causing one another and not think about it further. However we know how damaging this can be because if we lived in that type of world we would be assuming every time we ate a chocolate bar we would gain a pound, just because we know that eating unhealthy is correlated with weight gain, yet that would exclude the other factors such as genetics, or amount that person exercises. I also like how you mentioned that we could potentially be using many different kinds of tests to test the same variable with no way of knowing which test is best. It’s very important to know which tests are most reliable and valid to know which one to use to yield the most reliable stable results.

      Reply

    • Anne Marie Lemieux
      Sep 02, 2020 @ 19:59:16

      I think you make important points that researchers often “see what they want to see”. It made me think about how autism was thought to be the cause of refrigerator mothers in the 1950’s. One researcher, Leo Kanner, noted a lack of warmth from autistic children’s mothers and made this wild assumption. His assumption was taken as fact and countless parents were left feeling blamed for their child’s diagnosis. It is a prime example of how seeing a correlation and proving a relationship is so important. As well as being able to use assessments that have been proven to be true and reliable. You stated “tests need to be reliable so they can give precise and consistent scores for anyone taking them”. This is a legitimate point, that not only do assessments need to be accurate in what they are assessing but need to take into account diversity.

      Reply

    • Pawel Zawistowski
      Sep 02, 2020 @ 23:06:52

      Connor, I appreciate your observation that if we did not care about reliability, then we would not know how to select our assessments to be suit the needs of our client. I think this very simple observation is exactly why reliability is so important! I also think that when discussing reliability, it is also important to mention that by making your assessment more reliable, you are limiting the amount of error interfering with your test results.

      Reply

    • Brianna Walls
      Sep 06, 2020 @ 15:04:41

      Hi connor! I agree with your statement about how researchers may have an assumption or bring in some bias towards what they are researching. It may be easier for them to believe one thing causes the other so that they can report to their clients “useful” information. It is good to note that there are multiple other variables that are not taken into consideration when comparing two things to see if one causes the other. I also liked how you threw in you can’t have validity without reliability!

      Reply

    • Laura Wheeler
      Sep 07, 2020 @ 16:06:55

      Connor, I think you made some great points and it makes me wonder what measures can be taken to ensure that our biases and desired outcomes don’t sway the actual validity of the test and results. Certainly understanding that correlation isn’t enough helps, but I wonder what other specific practices/policies/procedures are in place legally and otherwise to ensure that biases are accounted for and eliminated in every way possible.

      Reply

  5. Beth Martin
    Sep 01, 2020 @ 22:04:04

    Correlation absolutely does not equal causation, but I feel that it masquerades as a “safe and easy” assumption to make. A lot of people, myself included, follow the “if that, then this” way of assessing information e.g. if I throw this egg onto the ground, it will break. This leads to an assumption that is really easy to reach: egg shells are fragile. There are variables that very easily can affect this – is the egg hard boiled, how hard am I throwing the egg? – but we tend to take things at face/surface value, especially when we have experience of previous face values being true (egg shells are, shockingly, fragile!). Deep down, I think a lot of us know that there are variables that impact everything we witness, and for simple topics such as ruining breakfast for everyone, it’s easy to break it down into “how can this be affected by me or the environment?”. There’s a handful of simple, easily-summoned questions to ask, rather than assuming that the egg will always break because egg shell are fragile. We’re made aware of force/momentum at school, and most of us know that eggs vary depending on how they’re cooked, or even by chicken. It gets harder to come up with possible explanations for results when you aren’t as familiar with the topic, or if the topic is more complex. I know I’m a lot quicker to see correlation and causation relationships when I am not well-versed in the topic, and that’s simply because I have absolutely no idea on what variables could be affecting, let alone what the variables themselves are.

    I think this is a common mistake in the mental health field for a similar reason. Practitioners typically have treated enough patients to be able to recognize the presence of variables in more common disorders, but it gets a lot murkier when it the less common ones. It’s natural to rely on research that’s been done by someone who has more experience with this particular disorder, or on your supervisors. But if you don’t know enough about your variables, moderators and mediators, it can be extremely difficult to automatically assume they’re involved. It’s easier, and in my opinion, human nature to take things at face value and to work off the assumption of X causing Y sometimes. I think it’s something all practitioners should be hyper aware of, though! If you’re at least aware that you don’t know something, that’s a start.

    Reliability is incredibly important for psychological assessment as it helps for consistency across the board. In theory, anyway. It helps ensure that the measure is consistent, which in turn allows people administering them to have more faith in them measuring progress. Assessment is something that needs to be done throughout treatment of all types, to a) make sure that a patient is responding to their treatment and b) note any changes in their symptoms that may need a different approach. As a person goes through treatment, the ideal is that they progress and display lessening symptoms. To get a good read of whether someone’s depression is improving or responding to treatment, we have to have reliable measures that are as consistent as possible over time. You can’t accurately measure someone’s progress if the test has poor test-retest reliability, and that makes actually assessing treatment considerably harder. As we know that the average client stays with a counselor for about 2 sessions, it’s also important to have high inter-rater reliability to ensure that a client is actually getting diagnosed with the same problem with multiple clinicians from the same assessment tools, not getting various different terms bandied around (which can be very disheartening and discourage seeking help/treatment). On the flip side, reliability also helps a clinician pick which assessment tools to use – if as assessment has a poor reliability coefficient, it’s probably not a great idea to use it! Finally, if a tool is not reliable, it can never be valid, so it’s absolutely crucial in assessment to make sure that a tool has the basis to measure what it actually says it is.

    RE: correlation/causation (I hope this is okay to post!), I always get a chuckle out of the graphs here: https://www.tylervigen.com/spurious-correlations , and it’s definitely a good visual tool for explaining that correlation definitely is NOT causation/

    Reply

    • Anna Lindgren
      Sep 02, 2020 @ 12:27:55

      Hi Beth,

      First of all, thank you for sharing that gold mine of a link. Hilarious content and drives home the point the correlation definitely does NOT mean causation! I think you’re onto something with the point you make about knowing less about a topic, and being quick to jump to the correlation equals causation conclusion. Specifically, as it applies to the mental health field, I feel like sometimes research findings can be oversimplified for public consumption through the media. People who aren’t mental health experts or used to analyzing data may draw conclusions based on correlations they see or read about.

      You make a good point about the need for reliability as well. I was mostly concerned with the clinician’s perspective in my response, but you raise a key point that it can have a negative effect on the client as well. If they are getting different diagnoses each time they take an assessment even though they aren’t feeling that different, that could break the trust that they have in their clinician that they can help them with their issues. You’re absolutely right, it’s imperative that the clinician use instruments that have proven reliability so that the therapeutic process isn’t hindered.

      Thanks for the insight and the laughs!
      Anna

      Reply

    • Abby Robinson
      Sep 02, 2020 @ 13:50:36

      Beth, I love the link you shared with the graphs! It really does a great job showing correlation does not show causation! You can take any two factors and show that the have similar patterns, but even though their relationship may be similar, one does not cause the other- I think my favorite was the number of divorces in Maine and the amounts of margarine consumed- LOL.
      Also, I really like your point on how the margin of error for testing really does need to be small and the reliability needs to be strong because the average number of sessions a client goes to is one to two. So it is important that the client is getting the appropriate test and diagnosis the first round because there may not be another chance to test them if the counselor thinks another test needs to be done to yield better results.

      Reply

    • Lilly Brochu
      Sep 03, 2020 @ 12:31:00

      Hi Beth,

      Your point about people generally taking things at face value is spot on. Researchers may use this to their advantage when presenting correlation data knowing that many may not take the time to look at additional research. I appreciated your honesty about making this same mistake in areas you are not knowledgeable in. Throughout my years of learning, many of my teachers and professors have stressed the importance of checking different sources to gather additional information about a particular topic. Similarly, with correlations, I believe that it is important to look at other sources and observe if there are other variables that may have played a role in the findings. Furthermore, you made a great point about the reliability of an assessment tool, and how the type of assessment tool you pick may have an impact on the measurement. It should be stressed that one may conclude different findings based on the type of psychological assessment they use. The psychological assessment can be reliable, but not if the assessment itself is not appropriate for the client. Great points, and I enjoyed the link you posted! 😊

      Reply

    • Destria Dawkins
      Sep 03, 2020 @ 21:15:40

      Hi Beth! You make a really good point that people like to assume that 1 thing causes another because it is safe to assume so. I agree. People don’t take the time to actually think about all of the different factors that are involved, such as in your example (an egg).

      Reply

  6. Anna Lindgren
    Sep 02, 2020 @ 11:47:04

    I think that sometimes people can equate correlations and causations because there are so many unknowns in the mental health field and it can make us feel better or less stressed by a problem if we think we can pinpoint the root cause of it. Additionally, if you as a mental health professional have a certain belief about or experience of a mental disorder, you may look for proof to back that up, and confirmation bias can come into play. For example, if you observe that your clients who do not exercise are more depressed, and then you see a graph charting the correlation between the amount of exercise and level of depression, you may want to conclude that lower activity levels cause a higher level of depression. But in this instance, like many others, we cannot prove through correlation alone which variable is affecting which, or if there are other variables at play that haven’t been considered. Are they depressed because they aren’t exercising, or are they not exercising because they are depressed? Or are there other factors that are affecting both ability to exercise and depression, like chronic pain or illness? Because there are so many factors that contribute to mental disorders, it is highly unlikely that the correlation of just two variables will give us the root cause.

    Reliability is crucial for psychological measurement because we need to be sure of what an assessment is measuring. If a client were to take a depression assessment and score very low, and just two days later took it again and scored very high, that would be cause for alarm. Barring a significant loss or crisis in that client’s life over the last two days, a score change that extreme would reflect poorly on the reliability of the assessment. In order for clinicians to accurately assess their clients and track their progress over the course of treatment, they must be able to use tools and instruments that will consistently measure what they claim to within the standard error of measurement.

    Reply

    • Abby Robinson
      Sep 02, 2020 @ 13:40:15

      Hi Anna, I completely agree with your thought on how in the mental health field we would feel less stressed by knowing that A caused or causes B. Having that strong answer as to why someone is feeling a certain way would be more comforting then just knowing there is a strong correlation between the two. I could see why this would be a common mistake that happens. But, you’re right by saying there are too many factors that contribute to that relationship in that it can show the root of the cause.

      Reply

    • Maya Lopez
      Sep 02, 2020 @ 18:16:25

      Hi Anna,
      I liked the example you used about exercise and depression because it reminded me of a post I saw the other day which was, “am I unmotivated because I am depressed or am I depressed because I am unmotivated to do things” and at times I think it can be both, it can be a cycle. Living in this confirmation bias type of thinking would neglect the cycle I mentioned as an option. One would only think one causes the other not that it could ever be a cycle so it is a very important bias to be aware of because it can narrow our mind and cloud our conclusions. I also liked the point you made about consistency in measuring and that can only come from having reliable tests to use, the importance of this is that we can look for patterns that may help us with our diagnosis or treatment plan. I wonder what it must have been like before having reliable tests to use and having to only go off of inference and observations from the clients at hand.

      Reply

    • Anne Marie Lemieux
      Sep 02, 2020 @ 20:09:45

      Your comment that researchers often look for proof to back up their beliefs made me think of Andrew Wakefield. He published a research paper claiming that autism was caused by MMR immunizations. Even though it has been debunked multiple times, he actually still stands by his research. He has also created a following of people who still believe this. Despite the fact that extensive research has been done proving it’s inaccuracy. Your comments on reliability being important as reliable assessments are used to track clients progress and regression was one I hadn’t thought of but an important point.

      Reply

    • Laura Wheeler
      Sep 07, 2020 @ 16:12:00

      Hi Anna! I totally agree, the unknowns are particularly uncomfortable in the field of mental health, and as someone else mentioned, things get more confusing when the symptoms get merky and arent as clear cut as we might always hope. I think the longer people are in the field it might become more habitual to draw conclusions when the facts might not all be there. What I mean by that is that if you are a clinician that generally works with depression and anxiety, and used to seeing a multitude of symptoms, you might jump quickly to that diagnosis. So, if a client comes to you and says they have 2 or 3 typical symptoms that you work with consistently, it would be easy to think it must be depression or anxiety- but being cognizant of that help us to be more careful and considerate in collecting all the information first to get a clear picture of the individual client before jumping to a diagnosis.

      Reply

  7. Abby Robinson
    Sep 02, 2020 @ 13:33:42

    Correlation does not equal causation is a common mistake that people often make. I think that people often mistake correlation and causation because they see the apparent pattern or relationship between the two factors, which makes it easy to assume one causes the other type of relationship. However, because people are only looking at the strong relationship between the two factors they are missing the “z-factors” or all the other factors that may influence the relationship. For example there may be a strong correlation with people who suffer from depression and also have alcoholism, but they having depression doesn’t cause an individual to have an alcohol addiction, there are other outside factors that may have influenced that individual.
    Reliability is very important for psychological assessment because if we continue to use instruments to test individuals in the counseling and mental health ‘world’, we want the testing to be appropriate and also strong in ways that yield results the counselors can use in the most effective ways. If we didn’t take into account if an instrument had strong reliability it wouldn’t be doing any good for both the counselor and the client because the results would continuously be skewed. Also, having reliability testing is helpful in this field because if you perform reliability tests on an instrument and the results show a large margin of unsystematic error, then the users of the test will know that the results won’t be as strong as if they margin of error was small. If the SEM was a large range we can then think of how to change the outside factors that may have contributed to that. For example, the test administrator rushed through the instructions and didn’t leave time for questions, or the room the test was given in was very loud, etc. It is important to understand that a perfect test with no error is extremely rare and that there will be some margin of error with testing, but it is important that the test yields results that have a little error as possible.

    Reply

    • Bibi
      Sep 02, 2020 @ 18:00:26

      Hey Abby, I really liked your explanation of why reliability was so important in this field. I also kind of touched on the fact that having reliable testing was important for diagnosis. I hadn’t even thought about how more descriptive aspects of reliability (such as SEM) has such an effect on the reliability of a measure so I thought that was really interesting to read.

      Reply

    • Elizabeth Baker
      Sep 04, 2020 @ 17:27:33

      Hello Abby,

      I liked your more realistic approach of why correlation does not equal causation. That yes, depression may cause substance abuse or vise versa, but that’s not the reason for every single case of depression or substance abuse.
      For reliability, it’s important to have good reliability within assessments because what’s the point of taking part in an assessment if you can’t accurately interpret the scores? Thankfully we have assessments that already have good reliability and are ready for the use; but it will be very important if you choose to create a new assessment in the future. As we stated earlier that correlation does not equal causation, we need to make sure we don’t come to a false conclusion by an assessment that has a poor reliability score.

      Reply

  8. Bibi
    Sep 02, 2020 @ 15:36:48

    1. I think people still struggle with correlation as causation idea because of how it is often reported. The news for example and popular media often reports correlations as though they are causations which then creates that link in people’s heads. An example of this would be that that drowning increases as ice cream sales rise. Drowning in reality has nothing at all to do with ice cream sales but you still see a spike of both at the same time. In reality both things are related to rising summer temperatures and the opening of pools and snack bars. (murders also spike around this time but for other reasons). In the mental health field specifically, I think that it the correlation equaling causation idea is often related to the confirmation bias. People are looking for a specific link and when they find it, they attribute it to exactly what they are looking for. A specific example in the mental health field is that brain abnormalities are the cause of depression (fitting with the medical model of mental disorders that all mental disorders have a biological cause). This fits the medical model and thus confirms that there is some biological link to depression. In actuality, we don’t know if brain abnormalities are causing depression, the result of chronic depression, or even the result of psychotropic medication use. However, people (and scientists) will still say that there is a biological link with depression because it fits what they are trying to say
    2. Reliability is really important in psychological testing in measurement because without it, we wouldn’t be able to tell if we are actually diagnosing people correctly. For example, if you have depression and you take a measure of depression, the measure should indicate that you do in fact have depression. If nothing else changes (you don’t seek treatment, other variables don’t change), when you take the measure again, it should say that you still have depression. If a measure is unreliable, it might say you might not have depression one time (when you actually do) or vice versa and then when you take the measure again, your score might drastically change. As psychologists, you expect that the measures you are using in order to diagnose patients are reliable in that they consistently measure accurate scores. Without that reliability, there would be no point in giving patients the measures because the score really wouldn’t mean anything.

    Reply

    • Elias Pinto-Hernandez
      Sep 03, 2020 @ 11:40:49

      Hi Bibi,
      I agree with your pots this week. I think that deep inside, we are prompt to create links in our heads and quickly assume referring now back to Beth Martin’s words: “if that, then this.” Then we have the drug companies that pay for research, and magically the results are in favor of the pharmaceutical. However, what’s more astonishing is seeing a respected researcher publishing misinterpreting data to mislead people. Can you believe that into this week, I was under the impression that chemical imbalances were the factor in certain symptoms of mental health conditions. Like you said, “link in people’s heads.”

      Reply

    • Tayler
      Sep 03, 2020 @ 12:16:30

      Bibi,
      I think your depression example is really salient. I wonder, does reliability take into account the fact that patients do tend to get better over time? In your example, the levels of depression are assumed to remain constant. But is that the case? I feel like for the nuances in test administration that would yield a ‘depressed’ verdict one time and a ‘not depressed’ verdict another time, the standard error of measurement becomes really important. Reliability allows us to get an idea of what a range of scores might look like, just based on the inherent measure of error. So we can use that to see the ways that a client’s depression may fluctuate, like the book says. Depression, like most psychological concepts, isn’t necessarily consistent (e.g., it’s dependent on situational and environmental factors), and so being able to gauge how it might change becomes very important.

      Reply

  9. Maya Lopez
    Sep 02, 2020 @ 15:37:01

    (1)
    Correlation for sure does not equal causation and this is very important to remember in this field. Just because 2 things have a relationship whether positive or negative it does not mean 1 is causing the other. I think it can be easy however to assume since 2 things are co-occurring it may seem as though one is resulting in the other. An example in our field that is coming to mind is when someone engages in NSSI such as cutting, we know that these acts can be seen by individuals who are trying to escape severe rumination, seek punishment etc. however the act of cutting does not mean that that person is indeed suicidal. There is a correlation that those who engage in self harm may be more likely than those who do not engage in self harm to commit suicide. However the act of cutting does not directly cause need/want for suicide. (that is not to say it cannot happen or is less important, it should still be taken very seriously and be proceeded with caution)
    (2)
    Reliability is very important for psychological assessments because we use many different instruments in the field to examine the patients in our care. We must know that the scales and instruments we are using are reliable and valid. For if we knew a certain test would not produce reliable scores it would be most wise to not use it. A test that is unreliable would make it very hard to tell if patterns are occurring and patterns are good indicators of behaviors therefore it is very important to us to know what we are using to assess patients is a reliable instrument. Another reason is that we make many inferences based off of scores, so if a test is unreliable and we don’t know it, we could end up misdiagnosing the patient or engaging in treatment that is unhelpful. In order to give our best to our patients we need to have the best data from reliable sources.

    Reply

  10. Carly Moris
    Sep 02, 2020 @ 20:41:30

    Correlation does not equal causation. Correlation shows us the relationship between two things or variables. It can’t tell us if one variable is influencing the other because it doesn’t take into account if there is another outside factor influencing the correlation. For example in the summer both ice cream sales and robberies increase. While these factors have a positive correlation it does not mean that an increase in ice cream sales causes more robberies. In the summer the weather is hotter and people tend to buy more ice cream. People also tend to leave their windows open and be out of the house more in the summer, which could lead to more robberies. The correlation between ice cream and robberies doesn’t take into account the third factor weather.
    Correlation not equaling causation is a common mistake and problem in the mental health field. There is an over emphasis on a biological cause of mental disorders. Following the medical model, mental health professionals tend to see correlations between differences in brain structure or functioning and behavioral or emotional disorders as being a causal link (Linchan, 2007). This outlook ignores psychosocial and environmental factors that could influence the development of a mental disorder. Concluding that a correlation between a chemical imbalance in the brain is the cause of a specific mental disorder ignores an important question; did the imbalance cause the disorder or did the disorder cause the imbalance? This is an important differentiation because it affects how we approach the treatment of these disorders. If the imbalance caused the disorder, medication to treat the imbalance would be an effective approach to treatment. But if the disorder caused the imbalance medications would be treating the symptoms and not the cause, which would mean that therapy would be a better approach to treatment. New research highlights the fact that the correlation between a difference in brain function and mental disorders may actually be caused by medication. The drugs we use to treat chemical imbalances may actually be causing them (Whitaker, 2015). This is why it is important for us to remember the difference between correlation and causation, because while it may be exciting/tempting to draw certain conclusions from a correlation it may not be correct and in the field of mental health that can have a significant impact on people’s lives.

    Reliability is important to psychological assessment because it lets you know that the assessment is actually testing what it claims to. You want to know that the instrument you are using has a high degree of reliability for the population your client belongs to so that you know the results of the assessment accurately represent your client. There wouldn’t be a point in using psychological assessments if they didn’t tell you what you wanted to know. That being said no psychological assessment is perfect and they all have a degree of error. This is why reliability is important because a reliability coefficient can be calculated to give you an estimate of the amount of measurement error in an assessment.
    The way we tend to look at reliability is based off of classical test theory, which was established on the idea that any result from an assessment is a combination of the individual’s true score and error. Through classical theory mathematical theorists came up with the equation to be able to estimate reliability. Reliability (r )= 1- error variance/observed variance. This is where I get a bit confused because the textbook says “ A reliability coefficient of .75 would indicate that 75% of the variance is true to observed variance”. The statement ‘true to observed variance” makes me unsure if it is talking about true variance or observed variance or the ratio of true variance to observed variance, but looking at the equation I think it’s talking about true variance to observed variance? I understand that reliability is an estimate of true variance to observed variance and how I would use the reliability coefficient in the equation, but I get a bit lost on the nuances of the equation and think I may be focusing a bit too much on the original equation from classical theory.
    I do know that it is impossible to get a measure of true variance in everyday counseling which is why we use correlation to give us an estimate of reliability that is based on consistency. Correlations provide an indication of consistency between two sets of data. This relationship is called the correlation coefficient and the more consistent the correlation is the closer the coefficient will be to +/-1, with 0 indicating no relationship. There are a few different ways to calculate reliability that will depend on the type of assessment being looked at. Each of these use correlation in a slightly different way. So it is important for us to know when to use the different types of reliability and how to calculate them. We want to make sure the instruments we are using report the right type of reliability for the assessment, ideally the assessment should report multiple types of reliability.

    References

    Linchan, M. M. (2007). President’s column on being mad versus bad: cautionary
    comments on biology as a causal explanation of mental disorders. The clinical psychologist, 60(2)
    Whitaker, R. (2015) Anatomy of an epidemic: the history and science of a failed paradigm of care. The behavior therapist

    Reply

    • Carly Moris
      Sep 05, 2020 @ 13:41:25

      I just realized that I got a bit off topic in my response and didn’t explicitly state that one of the main reason reliability is important to psychology assessment is that it tell us how consistent the scores are. You don’t want to use an assessment that will give you extremely different answers every time it’s used without the content it’s testing for changing because you won’t be able to make an accurate decision based of the results.

      Reply

  11. Lilly Brochu
    Sep 02, 2020 @ 21:00:31

    Correlation does not equal causation! So many times this statement was drilled into my head during my undergrad years. Causation implies that x causes y or vice versa. A correlation shows that there is a connection or relationship between two variables, but it does not consider the several other variables that may contribute to the correlation. Additionally, this observation alone between two variables is not a substantial amount of evidence to prove that there is a causal link between the two variables. For example, there has been several claims that exposure to violent video games causes aggressive or violent behavior. However, it is important to look at the several additional variables that may play a role, such as biological factors, familial physical/verbal abuse, genetics, and so on. In my opinion, I believe people are quick to jump to conclusions in their own research studies as well as others who are easily swayed by what they are exposed to on the internet and other media outlets. The assumption that a correlation is in fact the cause between two variables is a problem, especially in the mental health field, because researchers may take a “tunnel vision” perspective that focuses solely on their own findings, and deters them from considering the research of others, and the possibility of additional factors that may be associated with the correlation as well.

    Reliability is important for psychological measurements because one (a therapist, clinician, etc.) can assume that each time an assessment is given, it will continue to give consistent results. If an assessment is reliable, it conveys a sense of credibility of the assessment and the therapist. Within psychological assessments and the therapeutic process, it is important to have strong reliability in order to gauge the accuracy of the assessment and the progress of the client. These psychological assessments should be given throughout the therapeutic process to ensure the client is progressing, responding positively to the treatment, and to assess the therapist’s skills. If the reliability of psychological assessments are not reliable or consistent, it would be difficult for the client and the therapist to come to a clear understanding as well as several other negative factors, such as an inaccurate portrayal of the client’s needs.

    Reply

    • Tayler
      Sep 03, 2020 @ 12:09:46

      Lilly,
      That’s a good point about “tunnel vision”! In my clinical psychology class in undergrad, one thing we stressed was that counselors are prone to confirmation bias. When they see their patient every few weeks, and they seem better, they might assume that it was because of their treatment. I think this is a crucial correlation/causation mix up to acknowledge. Like you say, we need to consider the research of others, and also the population in aggregate. I feel like it would be easy to focus on your clients and your small practice and forget about the larger context of the work. Reliability for assessments puts the practice back in to context, at least a little bit!

      Reply

    • Cailee Norton
      Sep 03, 2020 @ 14:39:46

      Lilly,
      I also discussed how the media can be influential in projecting the idea that correlation equals causation, when in fact it is the opposite. I like your mention of “tunnel vision” within our field, and I think that this is definitely something many of us as helpers fall into thinking. We believe that the treatment we are giving is helping a client, while not considering the other factors that come into how a person may be feeling that could be what happened to them that day, if they just got into a new relationship, or anything else that would typically make someone happy. We as a helper in this scenario are solely viewing the impact we make on that person, not this outlying factors. We believe we are the cause of that happiness, not that our treatment might be correlated with their happiness. As you mention this is something we’re exposed to within the media. The media can play a big part into perpetuating the false narrative of A + B = C. You used the example of violent video games and their impact on violence in children, I think this is one of the classic examples of the media misconstruing the studies being performed in order to neatly wrap up an explanation to give to our society. Great job on your post!

      Reply

  12. Cassie Miller
    Sep 02, 2020 @ 21:29:05

    The statement correlation does not equal causation is an extremely important one. To oversimplify it is basically stating that just because there is a mutual relationship between two variables does not mean that either variable was responsible for creating that relationship. A basic example of this could be that ice cream sales went up significantly over the summer and so did shark attacks. Here you can see that this is an obvious correlation, but it does not mean that either variable mentioned is responsible for this relationship. What I mean by this is that shark attacks may not have influenced ice cream sales and vice versa. There are many other options that may have increased both variables separately: such as an expansion of ice cream companies, or great weather over the summer (increasing the amount of time people were outside eating ice cream or swimming in the ocean).
    Focusing back on the mental health field it is important to be aware of this rule because it can become very easy to rely on a certain statistic or variable for the production of a behavioral response (when it may have had nothing to do with it). This is why it is important to know how reliable your results are and their standard of error. Furthermore, correlation is often used as a make-shift solution for a problem that is much broader than it seems at surface value. Many times we see this error in pseudoscience where marketing comes into play for a “solution” or “quick fix” to a mental health issue that is not significantly impacted by that given variable. Sure, the variables may have some sort of relation to eachother, but this can happen for a multitude of different reasons (this is why it is so important to make sure the information you are reading is peer-reviewed and scientifically significant).
    The issue discussed above is a great segue into reliability. Reliability in short, is the consistency of measurements, allowing us to determine whether the scores on multiple instruments are consistent with eachother. It allows for an estimation of the proportion of error variance, as well as, the proportion of total variance. Thus, when conducting a psychological assessment it is vital to make sure that the instruments used correlate consistently, allowing us to gain confidence in the results obtained. In addition, the higher the reliability score the better, since it means the scores have minimal variance with eachother. Without reliability we would have no way of knowing whether the information from these tests were accurate, or even consistent. This opens us back up to the correlation does not equal causation debate. Since the error variance between assessments would not be examined, individuals could easily be misinformed about data used to explain behaviors that really do not share a relationship at all.

    Reply

    • Viviana
      Sep 03, 2020 @ 14:07:40

      This quick fix that mentioned in your comment is often times practice in the mental health field. The first step in any treatment in getting the right diagnosis and I would assume that is not an easy task for clinicians that need to report to health insurances in order to get the payment for sessions. Given that, I wonder at what stage in treatment a counselor decides to obtain a formal psychological testing or even if a test-retest to determine reliability does get applied and under what circumstances. If a counselor applies an instrument to a client to evaluate depression on a certain day and the day before the second time this test is taken the individual gets into a verbal or physical argument with his/her partner then an unsystematic error prevails as the person’s emotional state got affected by the incident. As the reading in chapter 3 in our textbook, even if the test-restest is utilized it presents difficulties for determining reliability of an instrument. One example is known math program that children would take a test after finishing a specific level and passing this test with certain points is required to pass to the next level of math. Now, the child is allowed to take the test many times as needed until she/he obtains the passing score. Reliability is this test-retest example is very low as the child will mostly remember the right answers from the first test and could also work on finding out the results for the wrong answers. I would be interested to know how counselors get to choose what instrument to use in their private practice and if they have the professional liberty to use certain instruments. Health insurance coverage, I believe plays an important role on this. For example, health insurance masshealth in the state of Massachusetts requires clinicians to assess children and youth under the age of 21 with a specific instrument to evaluate behavioral health called Child and Adolescent Needs and Strengths (CANS). Clinicians get trained to administer this instrument and regardless if the clinician believes it is whether reliable or not, the instrument is mandatory for treatment under this insurance.

      Reply

  13. Tanya Nair
    Sep 02, 2020 @ 21:54:41

    Correlation does not equal causation because there could be many other factors that also come to play when saying that one thing relates to another. For instance, we all know about the example of a positive correlation between ice-cream sales and homicide rates. However, does this mean that ice-cream causes us to commit a crime? No. There is evidence to show that hotter temperatures cause an increase in crime and also hotter temperatures make us crave ice-cream. In this case, heat is our other factor that may be seen as a confounding variable. I think many people in the mental health field like to believe that correlation is causation because they want to prove a point and believe what they want to see. It is often easier for a mental health clinician to explain to their client that because they have trouble focusing, they have ADD instead of adding more complex terms that make it difficult for them to understand. It could also perhaps enhance the client to easily work on a goal. For instance, if the mental health clinician tells the client that they have to stop being hard on themselves so that they can change a behavior, it is easier to hear what they need to do instead of hearing multiples things that can instead make them less motivated to change their behavior because there are other reasons to justify why their personality makes it hard for them to change a behavior. In essence, I am eluding to the fact that mental health clinicians want to make it ever so easy to help the client with their difficulty. This means that mental health clinicians prefer to tell their clients that X=Y instead of X=Y=Z so that the client is more likely to understand and comply.

    Reliability is the extent to which you have constant consistency between results. Reliability is important for psychological assessments because it shows us if an assessment is accurate. For instance, if an individual took an anxiety assessment and scored higher and then took it again and scored lower this would mean that the assessment is not reliable. It is important that assessments especially are reliable so that mental health clinicians are able to diagnose and treat clients accurately. Without reliability of psychological assessments, it would make no sense to give any individuals a diagnosis because no one would have the same thing and those with similar symptoms may also not have the same thing due to an unreliable assessment. Also, if an assessment is not reliable it is not valid, which means that it is not measuring what it actually says it is.

    Reply

    • Pawel Zawistowski
      Sep 02, 2020 @ 22:32:49

      Tanya, I also used the ice cream and homicide example I think that was a popular topic among our classmates! I am intrigued by your observation of mental health experts simplifying correlations do be able to communicate with a client better. I am curious if you think this is something that occurs inadvertently or purposefully? Also, do you think mental health counselors do that with the best interest of the client in mind or for selfish reasons? Does it help the therapeutic process or does it decrease the trust between the client and counselor. My opinion is that this may occur because mental health counselors feel rushed and pressured to come up with explanations and diagnosis because of limited amount of visits and possibly because they want to provide answers for the client to increase the chances of them coming back for another visit. What do you think?

      Reply

  14. Pawel Zawistowski
    Sep 02, 2020 @ 22:00:51

    1. Correlation absolutely does not mean causation! A correlation just means that there is a relationship between the two variables. We do not know if x is causing y or if y is causing x, or if another variable that is not being measured is causing this result. All studies will have many variables that are not being measured because you can only control and measure so many variables accurately within one study. An example of this is a positive correlation between ice cream sales and increase in murder rate. It is very unlikely that buying and eating ice cream is causing people to later murder someone or that someone will feel more compelled to purchase an ice cream cone after committing murder. The more likely hypothesis is that there is an unaccounted variable that is causing this relationship such us the time of year. In the summer there is more daylight allowing for more human interaction. The temperature is also warmer in the summer time, bringing people outside to socialize and also buy their ice cream cones. As for why this is a common mistake, I am not entirely sure. I believe it may have something to do with personal agendas and bias. After putting effort into a research project or just simply looking for an explanation, people may just settle for the simple explanation or attribute their findings to making a fascinating discovery. I would like to think mental health experts would understand that life is much more complex than that and would be very careful before they make a judgment on causation.
    2. Reliability is important for psychological assessments so that we can feel confident that the assessment we are asking patients to take will produce a very similar result if it was taken again by that same patient. For example, if I take a skills test and I score 75/100 on a given subject, and retake the same tests two more times and score 73/100 and 78/100. With this result I know that I have not quite mastered the skill and there is room for improvement. I am confident that my knowledge level is just short of the 80% mark. However, in a similar scenario if I scored 25/100, 58/100, and 98/100 without taking time to study I know there are unsystematic errors that helped me improve my score and the test did not do a good job of measuring my skill level. A strong reliability assures us that unsystematic errors have been minimized. This is important in psychological assessments because when working with a client we want to produce accurate and consistent results so that when we are providing care for client, we can best understand the person, what their needs are, and how we can provide them the care they need.

    Reply

    • Lilly Brochu
      Sep 03, 2020 @ 12:32:55

      Hi Pawel,

      Thank you for your example on correlations! Some of these correlation examples are rather humorous and silly, but it does make it easier for those to understand or grasp the concept of correlations, and why they are not necessarily accurate. It would make sense that people would use correlations to further their own personal agenda, research, reputation, etc. by presenting data in a specific way to persuade others into believing their findings are in fact a causal link between two variables. However, I think it is important in research to have as much information as possible from various sources, and do not understand why researchers would not appreciate the findings of others that may advance their research. It is plausible that people may enjoy the “spotlight” of finding the very cause of something, and do not want any interference. As for reliability, it is very important that the psychological assessments we give are reliable to ensure that our clients receive the very best treatment. Additionally, it is just as important as the therapist to have an accurate portrayal of the client’s symptoms, needs, and progress as well.

      Reply

  15. Viviana
    Sep 03, 2020 @ 12:04:46

    Correlation does not equal causation because even if there is a very strong association or relationship between two variables, we can’t assume that one causes the other. It is an understanding that correlation not only identifies variables but looks for a relationship between them. And by looking into the results of these two variables and their connection, correlation indicates if consistency is happening and this consistency is an estimate if the test is reliable or not. Causation as the word indicates is causing an effect. Now, even if the results show a positive correlation between variables and just because things seem connected on the surface does not mean that they are related or vice versa, hence if they are not connected then they don’t correlate. For example, I recall an example indicated by a previous professor who said that playing violent video games and violent behavior in children were correlated but also indicated that could be that the cause of both of these was growing up in a violent home environment and both playing violent video games and the violent behavior are the outcome of this.

    Individuals, not only in the mental health field but in different professions and levels in life, are predetermined to look for relationships between all sorts of events. Mental health practitioners and other professionals in the human health services are wired to find the cause to an effect to rapidly create a label for the individual without assessing external factors that could be impacting that specific behavior. Counselors need to continue assessing individuals consistently and over time and create a clinical rapport which it will give them the opportunity to know other factors or variables affecting their client. Mental Health counselors often times measure individuals’ skills and challenges with preconceived knowledge acquired in their personal and professional life and this also creates a barrier to make the mistake and diagnose individuals based on bias and prejudices.

    Taking into consideration that there are multiple instruments to assess multiples mental health disorders, it could be challenging for mental health field to find reliable instruments. Mental health counselors not only need to understand the instrument’s reliability but to know how administered it. It is really important that an assessment is reliable as the results of this instrument could have an important effect in a client’s life. Low reliability in a test or instrument could have a significant effect in an individual’s life as the results would go to other entities that would make decisions for this individual based on those results. I’ve observed in multiple occasions, along with other factors, parents losing parental rights over their child and this decision in court is affected by an assessment made by a mental health counselor who might have or not taking into consideration if the instrument applied was reliable. Mental health counselors need to spend meticulous time reading the instruments not only to learn how to administer it but to study and determine if the instrument is reliable. Also, it’s important that mental health counselors should have an understanding of the areas an assessment the instrument is evaluating in order to determine the level of reliability. It’s important to know the client’s culture and background and find instruments relevant to the culture. For example, I remember in one of my practicum at a public school years ago where a child who immigrated from a Caribbean Island was being assessed for an Individualized Educational Plan (IEP) with an instrument with questions about the snow and what to wear during winter time. How reliable the results were going to be for a child who didn’t have different seasons during the year, yet, decisions to support this student were made based on these results. Overall, providers should always use instruments with high reliability and be cognizant of the impact it will have in someone’s life.

    Reply

    • Lina Boothby-Zapata
      Sep 03, 2020 @ 14:12:37

      Hi Vivi,
      I think you brought an interesting discussion about Parenting assessments at court when DCF is looking for terminating parental rights on biological parents. This is an example of how relevant instruments and evaluations are in the mental health field and how powerful these instruments are in our society. Another example that I would like to add and I am sure you have seen it is the neuropsychological assessments, having conversations with the neuropsychologist and reading the results of our clients are fundamental to the Department to present an Action Plan to the families. I am always very careful about these recommendations because if we look carefully about it, these are the reasons why the family is involved with the agency, meaning lacking of the neuro-psych recommendations. Another example that you also brought is the Immigrant Children being assessed with instruments designed for American Children and I would like to go a little bit further, all these instruments are export and bringing to other countries; such as Colombia. Then circumstances need to add in considerations for these instruments, such as language, cultural background, and validation of the questions. I will be interested to know this process is called Validation of the test. I am curious about what is the statistics process that need to be happened to apply for example the Personality Test in Japan and how reliability and validity play out in this process.

      Reply

  16. Tayler
    Sep 03, 2020 @ 12:04:54

    1. It is so very tempting to say that correlation equals causation, because it makes the data more intuitively comprehensible. In my experience, just saying something is correlated doesn’t make me feel like I actually know anything, but if you say it’s causal, that feels meaningful. I also think people don’t like how much research, particularly psychological research, doesn’t actually answer the questions we have about various elements of the human experience. Instead, it seems like the field just circles what might be true, and correlation is a good example of that. We think, based on the data, that these two things might be related. But without the assumption that correlation is causation, we can’t say what that means, or even why we care. It’s very frustrating! It would be nice if things could just have clear causes and meanings, and I think that’s why people love to confuse correlation and causation. It makes the data feel more meaningful, and like we actually know more than we do.
    2. Reliability, to me, is important in the way that a recipe is important. If you’re trying to make a six layer cake (let’s just say all the layers are chocolate), you want every layer to taste the same. You don’t want a layer to come out all bitter or chalky, when the other five are excellent! Reliability feels like a recipe – can you use the assessment to consistently get a tasty chocolate cake? And just like a recipe, there is always a certain degree of error that must be incorporated in. Did you spill the flour a little bit? Run out of vanilla on the last layer? Small, unsystematic errors are why reliability is so important. There is no guarantee that every assessment will go the perfect, controlled way it was intended, and in fact most of them don’t. But, for example on the SAT, when the fire alarm goes off in the middle of the test at one school, we still need to be able to use the scores from the other schools! Otherwise it makes the whole test kind of useless, if it can only work in the exact perfect conditions. For psychological assessments, we need the test to have reliability because of the huge number of differences between people and conditions. One client is not going to be the same as another, even if they have all the same demographic information. We still need to be able to reliably measure constructs for them, accounting for all the small errors that make up the testing and individual experience.

    Reply

    • Lina Boothby-Zapata
      Sep 03, 2020 @ 14:18:49

      (POST)
      First above all, let’s answer the question of what correlation is? Based on the reading and the lectures. Correlation is the exercise of examining the relationship between two or multiple scores from the same instruments or tests. This examination of analysis will allow us to understand what is the reliability of the instrument or the score, meaning what the consistency between the scores is. One characteristic is the Correlation Coefficient; the correlation coefficient is the numerical indicator of the relationship between the two sets of data. The correlation coefficient can be expressed in different types of methods, such as the Pearson-Product Moment Correlation Coefficient. The mathematical formula is utilized to be more accurate in the calculation r=∑ Z1 Z2, ÷ N. Other characteristics of the Correlation Coefficient, is the range between -1 to +1. Now the closer the coefficient to 1, the better; we don’t want a coefficient of 0 because this will tell us that there is no relationship between the scores. A simple example is a test in the class; the individuals take the same test twice. If it is negative, let’s say that the first time the individual got a score of 4.2 out of 5, and the next test fall to 3.5, then the result will be closer to -1. I am assuming that correlations can be made with a re-test for individuals and groups, and it is also viable to do repeat twice or 100 times. The third characteristic of correlations is that the Correlation Coefficient is represented in graphs allowing us to have a visual understanding of the correlation but also a more accurate interpretation. The lecture provided us examples such as; more exposure to TV programs, more aggressive behaviors meaning that there is a direct correlation between Variable and X and Y, or the more you exercise yourself, the more longevity you will have. However, multiple variables are not adding into consideration that will probably prove that there is no causation; for example, aggression can be also be correlated with socio-economic level and parents’ education.

      Thoughts about reasons why this assumption; “Correlation is Causation” is a common mistake in the mental health field. I will say that there is bias in our society, as well in the social and psychological field. There are a couple of examples that I can provide in my area. A 51A report was filed due to parents consuming alcohol and drugs. The mandated reporter stated that the child is being neglected, there is no food at home, and the kid spends hours alone in the house. A hotline response is initiated, and during the clinical team meeting (before going to the field), social worker and supervisors made assumptions around “parents substance abuse, and children’s neglect is a cause of removal”. As a consequence, the investigators have in their mind the thought, “removing the child”. Another example related with Substance Abuse is that some Counselors don’t recommend to their client’s medication such as suboxone or vivitrol to control cravens and anxiety because they think that “medication maintains the addiction”. Finally, moving into the clinical field, what I can think about is when counselors have a client with comorbid diagnoses such as PTSD and Anxiety. Clinical questions emerged about which symptoms need to be addressed first and also which diagnoses cause the second one. In summary, I believe there is a tendency to find for the origin or the cause of the illness/mental health. However, these types of bias “correlations are synonym of causation” creates a misunderstanding of the relationship between variables and also brings narrows thoughts in the clinical field while we found several variables in the mental health field that affect the individual.

      Why reliability is important for the psychological assessment field. First, I am thinking about research and investigation where variables are being examined, and hypothesis are made to confirm or no these assumptions; as an example, is: Murphy, C. M., & O’Leary, K. D. (1989) found that Psychological aggression predicts physical aggression in early marriage. The investigation measures the reliability between the following variables: Psychological aggression by self and partner, physical aggression by the partner, and marital dissatisfaction. These variables were examined as a predictor of physical aggression during the marriage”. Hence, reliability can allow us to measure these variables and find what type of correlation exists between those variables. Second, reliability/correlations can be applied in the social worker field; an example is when we utilize our Domestic Violence Specialist to do consultations. This Specialist is always in the direction of finding the variables, and assessing the relationship between them, asking during the conversation is the specific situation is substance abuse a direct correlation of domestic violence? Or this is Domestic Violence, meaning a relationship between a batterer and a victim? What any other circumstances or variables are creating these violent/physical aggression situations between these partners? After the positive or negative correlation of variables is identified, the next step is to decide as a clinical team the specific services that need to be implemented with the families. Furthermore, it will help us to determine what it is the terms of Safety Plan needs to implement with the family. I guess this is a positive example of using of reliability/correlation during the assessment period. Lastly, in the counseling/assessment period between therapist and client. A potential scenario could be if I am doubtful about a diagnosis, meaning that as a counselor, I am struggling to identify symptoms of depression; I can use the instrument Beck Depression Inventory-II and measure the reliability with the Re-test. Outcomes will allow me for this specific client to confirm or denied the possibility of those symptoms, hence to incline or no to make the diagnosis of depression.

      Reply

    • Cailee Norton
      Sep 03, 2020 @ 14:28:39

      Tayler,
      As someone who has picked up baking during this pandemic, I found your explanation of the importance of reliability within psychological assessment to be perfect! The end result of a recipe (assessment) can vary greatly if one of the steps isn’t done properly or is skipped (a construct has low reliability). By paying attention to the unsystematic errors (too much flour, not enough eggs, etc.) that impact the true score (the recipe itself) we are able to come up with something that has reliability. I think you make some valuable points, and to add on to them I’d like to say that without reliability it’s difficult to use any psychological assessment in moving forward with our clients in a helping session. The entire purpose of a psychological assessment is to take a measure of our clients in order to better understand them at some level, be it on how depressed they are, how intelligent they are (more depending on what type of helper you are), or whatever it is you’re trying to find out. If we can’t trust that the assessments we are using have good reliability, then it makes it difficult to know where to move forward from. Of course, no assessment can have perfect reliability (I will say that a baking recipe can be very perfect), but having a standard to which we either accept an assessment or reject it for being unreliable is where we set ourselves up to succeed with our clients. Great job on your post!

      Reply

    • Anna Lindgren
      Sep 03, 2020 @ 18:56:50

      Hi Tayler,

      I agree with your thoughts on correlation and causation. I think that especially in a field where virtually everyone has some personal experience, like mental health, we may want to see our own experiences that we know to be true validated through research. Things that anecdotally we accept as true, may not have been clinically proven yet and that can definitely be frustrating.

      I love your cake metaphor for reliability. Baking is a science and while there is a little wiggle room for error, if you’re missing an ingredient you can end up with something wildly different than you set out to make. It fits nicely with the idea of systematic and unsystematic errors as well.

      Reply

    • Christina DeMalia
      Sep 04, 2020 @ 14:21:26

      Hi Tayler,

      I think you make a great point when you point out the frustration some people feel with a lack of answers. When something is cause and effect, we know exactly what happened and exactly what caused it to happen. These are the type of straightforward answers we have become accustomed to with many aspects of science. In the past, psychology was similar to philosophy. People would suggest theories and make guesses, with the assumption that no one idea could be proven more correct than another. However, as psychology shifted to a scientific discipline, we started getting more concrete answers. By using the scientific method, we can now conduct studies and get results, and this feels a lot closer to having specific answers. As you phrased it, the information starts to feel more meaningful. However, most times studies really are just pointing out correlations, noting that there could be other factors playing a role. We are often desperate for explanations of causation, even if they aren’t there, because that is what other scientific disciplines have given us.

      I also love the example of a recipe for reliability. It doesn’t matter if the chocolate cake comes out great the first time if it can’t be recreated the next time you try to make it. It reminded me of my Italian grandmother who passed on only handwritten recipes. Rather than tablespoons or cups, she wrote her measurements in handfuls and “half a handful”s. Even though her meatballs were great, when we tried to recreate them, we couldn’t get the same results because the measurements weren’t reliable at all. Her hand might have been a different size, or what she viewed as a half-full hand could have been different than someone else. I am really glad you shared this comparison because I think its a great way to conceptualize the importance of reliability.

      Reply

  17. Cailee Norton
    Sep 03, 2020 @ 14:21:23

    1. The idea that correlation does not equal causation can be very difficult for people to understand. What the mental health field tries to focus on is that two things can be correlated (putting in a lot of time to study on a test can be positively correlated with doing well on that test) that doesn’t mean that one is caused by one another. The way many people understand the world around us is much to the idea of A + B = C. The media is one of the biggest ways in which information can be misconstrued. Often we see media report various studies where a new drug makes you happier. While in fact that drug may have had an impact on an individual’s happiness, we do not know what other factors played into that (people going to therapy, people being laid off before or during the drug trial, etc.), what the study itself actually says, the way the study was structured. These are the important details that the media neglects to mention that lead us to the assumptions of well this drug makes people happier, so if I’m not happy all I need to do is take this pill. Much of pharmaceutical commercials we see also lean on this belief in order to sell their product. Due to these media influences we as a society come to expect that formula of A + B = C to be true for the information we take in from the world around us. Those working in the mental health field are no different and are susceptible to such messaging and thought processes. I also think that as humans we like things to be fairly neat and orderly to explain difficult things. Therefore we want to believe that if I do A then B will happen for me. It’s important to remind ourselves that correlation does not equal causation, and we must strive to uncover the underlying factors in order to understand the correlations we make and see in our media.

    2. I think that reliability is very important within psychological assessments because the conclusions we make about topics like depression come from the assessments that are performed. If a psychological assessment on depression has very little reliability, we can’t make proper conclusions about the individual. If they were to take that assessment multiple times and receive a 30/100 the first time, a 50/100 the second, and a 80/100 we can clearly see there is something wrong with the reliability of this test. So we must ask ourselves when examining an assessment what the reliability is and how this can impact my end result of being able to potentially diagnose someone and provide treatment or seek an alternative path for exploration with that client. On the other hand, with an assessment with good reliability I know that I can use that assessment with my clients, while still maintaining an understanding of possible differences between various clients, conditions of that assessment being taken, etc. This impacts not only my assessment process, but the treatment I provide and conclusions I can make with my client.

    Reply

  18. Christina DeMalia
    Sep 03, 2020 @ 14:50:41

    (1)
    The phrase “correlation does not equal causation” is one most people have probably heard at one time or another. I would also guess that a large portion of people understand what that means, in theory. However, it is often forgotten about when it comes to talking about and analyzing results.

    I believe one contributing factor for the prevalence of this mistake is the way psychological studies are released to the general population. As psychology students, we know that the most reliable sources of information are things like peer reviewed journals. However, looking at one of those journals, it becomes very apparent why the general population would not read them for enjoyment. Instead, people are far more likely to read an article that pops up on their social media with click-bait headlines such as “Children who watch ‘Sesame Street’ may perform better at school, study finds.” A very small amount of people may actually read the published journal on the study, more will only read the article covering the study, and even more will simply read the title of the article and make assumptions. A title like that leads people to the assumption that if their children watch Sesame Street, it will cause them do better in school.

    Although the study may have found correlations between viewing sesame street and performance in class, it does not necessarily imply that is the cause of their improved performance. In fact, when reviewing the actual journal, it states that “The question that we are best able to address with our analysis is what is the impact of making a show like Sesame Street more readily available to children, not what impact does it have on an individual child watching the show,” and “our large-scale analysis finds positive impacts on the educational performance of the generation of children who experienced their preschool years when Sesame Street was introduced in areas with greater broadcast coverage.” (Kearney & Levine, 2019) The study actually just noted a correlation between areas where the show could be broadcast, and performance in school. This means that several other factors could have contributed as well. For example, it is possible that the broadcast reached cities, where there were larger more equipped public schools versus rural areas where the broadcast didn’t reach, having smaller schools with less resources. This could account for the difference in school performance, rather than actually watching the show itself, since the study was unable to look at those specific factors.

    I think another reason why this mistake may be common, specifically in the mental health field, is the way people are always looking for specific answers to the question “Why?” If someone is depressed, they may be happier with an answer such as “because of a chemical imbalance in your brain,” rather than an explanation that there could be many complicated contributing factors, and not one clear answer. The first answer seems to suggest that something wrong with your brain is the *cause* of the depression, and not that maybe the depression someone is experiencing could lead to issues in the functioning of the brain. It also excludes things like external factors, environment, previous experiences, etc. When we notice a pattern, it is easy to guess that one thing must be leading to another. However, it is very rare that we can isolate single factors to be sure one thing is the cause of the other.

    (2)
    Making sure a psychological assessment is reliable is extremely important. Many psychological assessments aim to test for things that aren’t easily measurable. Things like height, weight, speed and temperature can be measured with predetermined units. Things like depression, anxiety, personality traits, and emotions are not measured in a widely agreed upon unit, so creating assessments that test for them is already complicated. That is why so much attention must be given to ensuring the reliability of assessments that are created for measuring those types of attributes.

    Because of systematic error, no assessment can be perfect or without error. However, assessments should aim to have as much reliability as possible, given the seriousness of the things they are testing for. People undergoing psychological assessments are often having issues with their mental health in one way or another. Assessments are beneficial when they can correctly identify an area of concern, and can therefore lead to treatment. The consequences of incorrect results on one of those assessments, however, could have serious effects.

    An example of this could be someone who was struggling mentally and was given an assessment. If that assessment had a lower reliability, its results might suggest the wrong issue. Maybe the test looks for traits of depression, without proper questions asking about other characteristics. Therefore someone who actually has bipolar disorder could be inaccurately categorized as only having depression and given anti-depressants. Although this may help them during depressive phases, they might not be on the correct medication to help with a manic phase. Untreated, a manic episode could create anything from short-term issues for the client to life altering problems, depending on the severity. Hopefully, a professional would never diagnose off of one assessment but at the minimum an inaccurate assessment could lead to a delay in treatment for a patient, which has its own risks as well.

    Kearney, M. S., & Levine, P. B. (2019). Early Childhood Education by Television: Lessons from Sesame Street. American Economic Journal: Applied Economics, 11(1), 318-350. doi:10.1257/app.20170300

    Reply

    • Nicole Giannetto
      Sep 06, 2020 @ 14:50:44

      Hi Christina. I loved that you commented on the influence and power of “clickbait” headlines on social media or the internet. It does feel like no one truly reads anymore, because we have the internet that is full of short phrases that get a message across quickly without people thinking they need to read because they got the whole message already from the title. This is definitely problematic, because what is written about research is not given the proper attention from the public. I feel like people today rely on the internet to inform us quick and easy as we go throughout our day. When we do this we forfeit our chance to deepen our understanding of a topic of interest.

      Reply

  19. Karlena Henry
    Sep 03, 2020 @ 16:45:52

    I think the problem with people believing correlation equals causation stems from the simple philosophical maxim which states that “of any given set of explanations for an event occurring, the simplest one is most likely the correct one”. This is a constant issue with the residents I work with, and it leads to significant problems.

    I work at a recovery house for women early in recovery. For most of the residents, they assume that their substance abuse was the cause of the destruction of their lives, but in fact most everyone chose to abuse alcohol and drugs to cope with underlying mental health problems. Unfortunately, their belief that the addiction led to their problems leads them to assume if they quit their addictions they can just go back to normal. By accepting the simplest issue (substance abuse) as the root cause, when they leave the program, the stress from their mental health problems are still affecting them, and often they relapse. If they examined deeper into how their addictions developed, they would discover the real issue, and work on both their problems while in the program. I definitely understand why it happens this way. The motivation for most of them is to try to get out of a consequence their substance abuse caused like having their children removed by DCF or being arrested for a drug related crime. The two issues, mental health problems and substance abuse are often correlated, since you see them together a lot, but in fact they have two different purposes and two different treatments that need to be addressed independently. The significance of this mistake can be life-threatening with some people.

    Reliability is wicked important for mental health practitioners because, as you talked about in the lecture, our reliability is not typically as high as traditional scientific studies. Our data is based off of subjective information like participants answering questions or people’s observations instead of solid information.
    Another issue is how many in the scientific and general public view psychological research as inaccurate. When we make assertions from our research findings, outsiders question how accurate our results can be when there is no testable evidence. For instance, take the rorschach test. When it was developed it was accepted as a valid form of psychological measurement, but now, very few people take it seriously, and many outsiders from the field is a representation of our “research “. It lowers our reputation in the scientific community and with the general public. If people don’t take our work seriously, then it limits our effectiveness.

    Reply

    • Carly Moris
      Sep 05, 2020 @ 16:12:46

      Hi Karlena! I really like how for why correlation does not equal causation is important for counselors you used examples from your work. I think that you bring up an interesting point that it’s easier for these women to acknowledge their substance abuse than a potential underlying mental illness that may have contributed to the issues in their lives. I wonder if this is because it is easier for them to recognize an external cause (substance abuse) then an internal cause (mental illness)? Or if maybe recognizing the substance abuse as an issue in their lives is a first step in recognizing their mental illness. I agree with you that mental illness and substance abuse have a strong correlation. But this does not mean that one necessarily caused the other if we just look at correlations. We wouldn’t be able to determine if metal illness caused substance abuse or if substance abuse caused mental illness from just a correlation.

      I also think you bring up a really important point about why having reliability for psychological assessments is important. If our assessments aren’t reliable people won’t trust us or our recommendations. It is good for people to be skeptical to make sure they are getting the best care. But if there are too many unreliable assessments being used, people will begin to stop trusting even the reliable assessments.

      Reply

      • Karlena Henry
        Sep 05, 2020 @ 20:53:53

        Hi Carly,
        When I was writing the section about the importance of reliability, I had a flashback to my undergrad where I had philosophical arguments with my classmates on whether psychology should be included in scientific research. By the lack of quantifiable data we are usually working with, my biology major classmates didn’t even want to discuss it. I often wonder if the attitude by the scientific community directs society to question the value of therapy. Did you ever run into this challenge? If so, I would love to compare experiences.

        Reply

        • Carly Moris
          Sep 09, 2020 @ 21:48:54

          Hi Karlena, I have to say that I didn’t have that experience during undergrad. It never seemed to be a question that psychology didn’t belong in the scientific community. I find it odd to not think of psychology as part of the scientific community, especially considering that our program uses evidence based techniques. They couldn’t be evidence based if there was research and science to back it up. The study of psychology requires research using the scientific method, and I know I read a fair number of psychology research articles during my undergrad!

          Reply

    • Lina Boothby-Zapata
      Sep 05, 2020 @ 17:05:40

      Hi Karlena,

      thank you much for sharing your work experience at the housing recovery. As a DCF social worker, I support mothers to work towards reunification with their children while they live in these houses. I share with you the same thoughts about Substance Abuse and Mental Health. They are two different variables that need to be addressed in terms, of treatment. But at the same time, they are interconnected. Trauma history is common with parents with Substance Abuse. When we do the DCF assessment named FAAPS to these parents and create the Action Plan, I use to add independently outpatient therapy and substance abuse treatment. However, this is very difficult for the clients, in this case, mothers that they are already overwhelmed with treatment. Hence, most of the time it is necessary to decide which needs to be tackled in terms of treatment. In terms of correlations, I think will be what variable is predominant in the life of this specific individual and what type of treatment it is necessary to emphasize first mental health or substance abuse?

      Reply

      • Karlena Henry
        Sep 05, 2020 @ 20:40:15

        Hi Lina,
        The house I work at has several mother’s who have physical custody because they are in our program, so I interact with DCF often. I completely agree that dealing with the substance abuse seems to be the most immediate concern, but in my opinion, even though it’s difficult, making some headway in their mental health problems needs to be accomplished before they would be discharged. Our program attempts both, and for most of the residents, they understand the importance of dealing with both. We should discuss this some time not in a discussion blog.

        Reply

  20. Brianna Walls
    Sep 03, 2020 @ 17:27:03

    1. Correlation and causation are two terms that are usually misunderstood and are often used interchangeably. It is important to understand that there can be two things that share a relationship without one causing the other. For example, some may think that if you exercise you will live a longer healthier life, but if you take other variables into consideration like diet and smoking you will soon realize that there are multiple other variables that are not being measured that may be affecting the relationship between exercise and a healthier lifestyle. I believe some mental health professionals make the assumption that correlation equals causation for a few reasons. One reason may be that they are just drawing to a conclusion too quickly. For instance, you should take the time to find other underlying factors/variables they may be affecting the relationship between the correlation. A mental health professional might get ahead of themselves and conclude that because there is a correlation between two things that there has to also be a causation effect. It is important to note that just because it seems like one variable is influencing the other doesn’t mean it actually is. I also think some mental health professionals are eager to find a cause between a correlation. There is so much unknown information in the mental health field that they become eager to find out what the cause is to an underlying mental health problem that their bias gets in the way and they forget to include other multiple factors. They just want to come to a quick and easy answer so that they can inform their client(s).
    2. Reliability is very important for psychological assessments for a few reasons. The number one reason or the most obvious reason is you want to be sure the assessment you are providing or taking correctly measures what it is supposed to. For instance, if you wanted to measure someone’s depression you would not administer an intelligence assessment you would want to provide the Beck Depression Inventory assessment instead. Reliability is also important because for example, if you are weighing yourself the results of each weighing may be consistent but the scale itself may be off and may be giving you a false reading, therefore you cannot have validity without reliability.

    Reply

    • Alexa Berry
      Sep 06, 2020 @ 00:06:17

      Hi Brianna,

      I thought your comment about how counselors sometimes come to a conclusion without examining third/other factors was very thoughtful. It made me think about how primary care doctors often do the same thing. If you go in with a stomach ache they typically dismiss it as a minor ailment such as a virus, food poisoning, pulled muscle etc. It is important to be accurate and examine all factors in clinical settings because you don’t want to miss anything important (like an appendix thats about to burst) just because it appears there is a causal relationship between two variables. I definitely have struggled with this myself even though it is an easy enough to grasp concept. When doing research between mental health and physical health during my undergrad studies my group wanted to see if we were able to find a causal link between the two, unfortunately with the data we were using this wasn’t possible. I think you’re right when you say people in the mental health field can want to find a cause for why a specific illness it happening. It can be easier to provide a concrete answer as to why something is occurring, rather than stepping back and looking at the bigger picture.

      Reply

      • Brianna Walls
        Sep 06, 2020 @ 15:12:42

        Hi Alexa! I really liked how you took my response and gave me another perspective to look at. It is true that when you visit your primary doctor they usually only consider minor issues. To add to your response when I was younger I was diagnosed with type 1 diabetes. I was in and out of the doctors office for months until they realized I had something serious going on, but at first my doctor just assumed I had some type of bug/virus. If my doctor never took into consideration other factors/variables I might not have survived. Now this is a bit out there but I think the message is important!

        Reply

    • Nicole Giannetto
      Sep 06, 2020 @ 14:45:04

      Hi Bri! I agree with you that one of the most common misconceptions is that people often think correlation and causation are interchangeable terms. I liked how you emphasized the importance of examine all contributing factors that influence a certain behavior, feeling or thought. It is very important for clinicians to recognize this so they can be better equipped when building rapport with a client.

      Reply

  21. Elizabeth Baker
    Sep 03, 2020 @ 17:31:10

    Although having a strong correlation (positive or negative) is a preferable result when looking over data, it does NOT mean that an increase or decrease of one variable is the only cause of the increase or decrease in the other variable. For a silly example that I came across in one of my undergrad courses, let’s take the correlation of Nicolas Cage’s appearances in movies with the increase in deaths by drowning. Of course, the easy answer would be to say that Nicolas Cage’s appearances in movies are causing an increase in deaths by drowning, but that’s obviously not the reasoning here. There are external variables that affect the causation of these two variables to positively correlate. For example, the increase of drowning deaths could be due to risky actions like swimming alone without a life jacket, swimming under the influence, swimming during a storm, or when the water conditions are dangerous. There can be other external variables of course, but there are a lot of external variables that factor into the positive or negative correlation. Sometimes the list of external variables is too large to include in the research, or maybe excluded for other reasons, but it’s important to list or mention as many external variables to support or debunk a researcher’s findings.
    Now in the realm of counseling, it may be a common mistake to conclude correlation causation without considering the external variables. For example, if a client tells their counselor that they’ve been feeling down lately and that they’ve been having difficult times at work, the counselor may quickly link being upset to the troubling workdays. Without further investigating the reasoning beyond having issues at work, that can result in the counselor providing a false conclusion. For example, the client may have had an intense argument with their significant other or family member, they might have seen or received upsetting news, etc. It’s important to see if there are any other outlying causes as to why the client is experiencing upsetting times, and to not quickly assume causation between the two events.

    Reliability is extremely important for psychological assessments because we need to make sure that these instruments are appropriate for use, and that they have been used in the past and have presented strong results. Having this information will help us feel reassured when needed assessments are proven to be of use. The book discusses how it’s important to have a range of scores to present a true representation of scores, instead of basing off of one score. For an assessment to have good reliability, the assessment has to be used multiple times and produce strong results. An assessment that’s been used multiple times is more reliable than an assessment that’s only been used once. Not only is this important within the structure of assessments, it’s also important when talking with clients. For example, if a counselor has directed a client through an instrument to determine if s/he has depression and just goes off one score from the instrument; that could produce a false interpretation of the scores, or the client may not believe the interpretation of the score. It’s important to have a range of scores to fully conclude that scores of the assessment are reliable and true, and can help counselors give a more clear interpretation to clients that may initially dismiss the scores. I think that’s why reliability is important in this field.

    Reply

    • Carly Moris
      Sep 05, 2020 @ 14:01:59

      Hi Elizabeth! I liked the example you used for correlation does not equal causation, I’ve never heard about the Nicholas Cage and drowning example before! I think that you bring up a very good point for why correlation does not equal causation is important in counseling. It can be easy to remember correlation does not equal causation when looking at a graph, but we also need to keep that in mind for more informal situations. As counselors it can be easy for us to make assumptions based off of a few things a client has told us, without getting the whole picture. I agree that it is important that we make sure to take into account other variables that could be effecting the client and we try to get more information before drawing conclusions.

      I also think that you seem to have a strong understanding of reliability and why it is important for psychological assessments. I like how you highlighted the importance of using a range of scores when reporting results to a client. That a range of scores can be more representative of the client especially for traits like mood that may shift a bit depending on the day. The client may not think a single score represents them in the moment, but can understand how a range of scores can represent their range of feelings.

      Reply

    • Viviana
      Sep 05, 2020 @ 19:42:30

      This common mistake of not taking causation into consideration when diagnosing it’s not only a common mistake but also misleads clients and families to think that probably what they have is only biological. If a child is diagnosed with mental disorders, counselors may be quickly to notice difficulties consistent with that mental disorder and fail to recognize other external issues that are causing some of the traits observed and evaluated such the home or community environment this child interacts the majority of the day. Also, counselors have a difficult job of how much of this environment is affecting the child if all the information is being self-reported. I wonder if in this case it would be appropriate to use an alternate or parallel forms test with the same number of items, and the items with the same format. For example, if a counselor wants to test a child’s disruptive, impulse-control, or conduct disorder and the instrument has questions that varies in wording in order to obtain answers to assess the family environment then it would be confusing the reliability of this test would be low, hence, as our class textbook indicates in chapter 3, often times the results of an alternate or parallel forms are a less prevalent method of estimating reliability.

      I believe for counselors to be skilled at not assuming causation will have to gain experience and practice over time. With time and experience counselors should be able to develop clinical skills and judgement which requires that the scores are not only interpreted as they present but also applying some subjective attribution. Now, these clinical skills should be closely self-monitored to not become judgmental or have preconceived opinions for clients.

      Reply

  22. Zoe DiPinto
    Sep 03, 2020 @ 18:12:24

    Correlation does not equal causation! This is a very common mistake that I frequently remind myself of in day to day life. Did you know that increased sales of ice cream is significantly positively correlated with increased crime? The first time I heard this, I desperately wanted to figure out why people who suddenly consumed more ice cream were going around, killing people. Because we are psychologists, we are trained to look at behavior (in this case- crime) and ask ourselves why it is increasing in frequency given the information at hand. However, we must actively be aware of the lack of information that is spoon fed to us. It’s proven that when ice cream sales increase, crime increases. Think about the third variables– ice cream is most often sold in the summer, when people are out and about and the sun is up for longer. The data also doesn’t describe what populations are being affected, the location of sales and crime, or even the type of ice cream or type of crime! Even if it was specified that the same people who were eating ice cream were breaking into people’s homes, we still don’t know the directionality of causation. Maybe after a big score, a criminal wants to celebrate with a nice, cold drumstick cone. Maybe there’s an ingredient in choco-tacos that make young men feverish for a bank robbery. Yes, a silly example, but it raises awareness for psychologists to avoid absurd claims of causality when presented with a correlation.
    Reliability is very important for psychological testing. Without reliability, there would be no standard of testing. Say a mother walks into your office and wants to know if her son needs extra help in school. No problem, you give the son an intelligence test. However, this intelligence test is famous because no matter how many times you administer the test, children receive drastically different scores each time they take it. The test lacks consistency. A metaphor that helps me is the comparison of a reliable test to a reliable parent. If I tell my mom that I want to be picked up from school every day at 3pm, a reliable mom will be there every day at 3pm. An unreliable mom may be there at noon on Monday, 3pm on Tuesday, 10pm on Wednesday, 2am on Thursday, and 3pm without a car on Friday, only to continue to appear at random times the next week. A reliable test is something that you can count on to give consistent information without worrying about back up methods to get the right information. We need reliability in psychology to tell our clients consistent and standardized information so people can receive correct treatment, and trust their provider.

    Reply

    • Tanya Nair
      Sep 04, 2020 @ 20:50:19

      Hi Zoe, thank you for your post. I too used the ice-cream and crime example as well in my post. It seemed to be a very common example amongst our classmates. However, I did enjoy your example because you brought up a variety of things that I had not thought of before such as how the data does not describe what populations are being affected, the location of sales and crime, or even the type of ice cream or type of crime. But yes, mental health clinicians really need to stop claiming causality when they are given any correlation. I liked your explanation and examples of why reliability is important. I think including more about what this means for mental health may beneficial to your post. Maybe mental health clinicians are aware of the other variables that play a part in correlations but want to make it easier for their client to understand by just saying X=Y instead of X=Y=Z.

      Reply

  23. Laura Wheeler
    Sep 03, 2020 @ 18:15:35

    Correlation does not equal causation for a couple of reasons. The first being that it is not always certain which variable impacted which; did variable x cause the response in variable y or vice versa? A correlation alone does not answer that question. Additionally, it is highly unlikely, especially in the field of mental health for two variables to be absolutely isolated with no outside factors; more often than not, there are many additional factors that could impact the results of a correlation, both known and unknown. Causation might be assumed based on a correlation in the field of mental health because of the deceiving impact of outside variables. For example, a client could begin attending therapy immediately following the loss of a loved one. After 3 months the client could have made significant progress and no longer exhibit problematic symptoms. If this clients’ sessions and progress were illustrated as a correlation, it would appear to be a positive outcome. However, it is likely that during those three months of treatment there were other factors contributing to the clients improvement. This isnt to say that the treatment didn’t also have an impact, but it cannot be said with certainty that treatment was the sole “cause” of improvement.

    Reliability is incredibly important in psychological testing. Given that psychological testing can impact so many things, including treatment and diagnoses, it could be unsafe and highly problematic to be collecting data that is not reliable. Further, consistent data is important to research and progress in regards to diagnoses and effective treatment. If psychological tests were unreliable it is possible that clients would be misdiagnosed, or they could receive no diagnosis when one is needed to ensure services. Another reason that reliability is necessary is to ensure that the correct test is being administered; meaning that the test is testing for the correct condition/symptoms/etc. As an example, an ASD assessment would not accurately determine whether or not a client has a substance abuse problem. In order to ensure reliability, tests need to be utilized many times in order to collect a broad range of data. Much like the example that was given regarding a scale showing the same weight each time you step on, if an ASD assessment showed different results for the same client if administered more than once the results would not be reliable.

    Reply

    • Tanya Nair
      Sep 04, 2020 @ 21:37:26

      Hi Laura, thank you for your post. I like the point you bring up about mental health clinicians believing in causation because of the deceiving impact of outside variables. I think it is very true. However, I also believe that mental health clinicians believe in causation to make it easier for their clients. For instance, they are aware of the other variables that play a part in correlations but want to make it easier for their client to understand by just saying X=Y instead of X=Y=Z. It is very important that assessments are reliable so that the client is able to get the most accurate diagnosis and best treatment. Obtaining a large data set could be a good way to see how reliable an assessment is. Sometimes when using various mental illnesses such as depression, we also have to think about how it often changes and if the test still is reliable even when improvement is made. What do you think about that?

      Reply

    • Timothy Cody
      Sep 04, 2020 @ 23:49:50

      Hi Laura. You bring up some very good points in describing the difference between Correlation and Causation. Now we know how to define correlation by finding out if a test is reliable, but how does one know if correlation does lead to causation? Is there a regression test that could be studied to determine whether therapy treatment does lead to an increase in progress? Also, would it be better in this case to test to see if there is a negative correlation between therapy treatment and negative symptoms of grief? Is that not the point of therapy, to decrease one’s abnormal behavior? Or are you taking the route where you would not only wish to decrease negative behave but also increase the positive?

      I would agree that it is very important to use test-retest in order to determine reliability. But would it not be better to test multiple patients in order to determine if the test across the board is reliable? Couldn’t there be the case where certain tests resonate with certain clients and therefore could be reliable, but do not sit well with others? I am wondering if this would disprove reliability if that is the case.

      Reply

  24. Alexa Berry
    Sep 03, 2020 @ 18:37:43

    Correlation equaling causation is a common mistake that is made. I believe that this error is common, even within the mental health field, because clinicians, researchers, and other people want to be able to provide an answer as to why there is a relationship between two variables. The book states that when individuals want to examine consistency, they use correlation. It is possible that when examining correlation, individuals contribute this consistency as a causal effect, when in reality there are possible confounding, or outside factors, that are influencing this relationship. A common example that comes to mind when I consider this is the relationship between heat and crime levels. There is a positive correlation between crime and warmer temperatures, such as in the summer. There are multiple possible confounding variables present in this relationship, such as the increased tendency for people to spend time outside in warmer weather and leave their doors and windows open and unlocked. During my research seminar in undergrad, I struggled with correlation as well. Part of our assignment was to conduct research that would bring something new to the conversation of existing literature. This was challenging because my group wanted to examine the relationship between mental illness and physical health. It was difficult to not be able to say that either poor physical health caused mental illness, or vice versa. Although there is other research out there that supports a causal link between these variables, we were limited in our data analysis and had to rely on correlation, and thus could not imply causation in our research findings.

    Reliability is important for psychological assessment and research because it ensures that there are limited confounding variables that contribute to the findings of an assessment or research. If assessments were being used that were not reliable, the results would not be useful since they could vary from time to time. This is specifically relevant to psychological assessments that counselors use because reliability can also vary based on populations. For example, an instrument can have a good overall reliability but not with certain subgroups (Whiston, 2013). At times it would not be appropriate to use the same assessment across populations due to concerns with reliability- such as using the same assessment for children and adults. In addition to reliability being important in choosing and administering assessments, it is also useful in interpreting results from these assessments. These measurements can help a counselor determine where a client’s score may fall within a range. The score they get once typically will not be the same score they get every time because they can be influenced by outside factors such as being distracted or not feeling well. Finally, reliability is also important due to the fact that it is necessary for validity.

    Reply

    • Christina DeMalia
      Sep 04, 2020 @ 14:33:19

      Hi Alexa,

      I think you touched on something really interesting here when you mentioned your undergraduate research. It is something I hadn’t considered initially, but your post made me realize this is probably a huge contributing factor to the prevalence of this problem. I thought about how the general public likes straightforward answers, and therefor looks for causation in results. However, as you point out, even the researches themselves are sometimes looking for those concrete answers. When a research question is posed, and someone spends countless hours on the topic, they may want to come out with concrete answers.

      I remember looking at mental health and social media use for my own project. In the end I had results that looked like interacting with social media increased anxiety, and when I presented my project that’s what many people took away. However, those were the results I was looking to find. There was a clear correlation between the things I measured, but looking back, I viewed it as cause and effect. Although I mentioned limitations, many people still got the general idea that social media use caused an increase in anxiety. As people in the field of psychology I do think we need to be extra aware of when we fall into this mistake ourselves with our own research or interpretation of data.

      Reply

  25. Timothy Cody
    Sep 03, 2020 @ 18:46:53

    When looking at things such a test-retest reliability, it is easy to confuse correlation not equaling causation, for in this type of reliability, the individual is taking the same test multiple times, ergo, one might insinuate that the two variables are linked through causation. However, reliability only measures the consistency of such measurements and does not guarantee that one variable cause another to occur. For example, there is a positive correlation between eating healthy and living a long life, but what if there is an accident and a life ends before longevity cannot occur? We can link a person’s sudden death to not eating healthy, so the two variables cannot be linked through causality.

    Reliability is an important part of the psychological assessment because tests cannot be administered and data cannot be submitted if the group if there is an issue with reliability. The test needs to be checked to see if it is measuring what it is supposed to measure. We cannot further persist and check for Validity if we first do not have Reliability. If a test does not have further reliability, then it could lead to data that is skewed and disproportionate to the test itself.

    Reply

    • Zoe DiPinto
      Sep 04, 2020 @ 11:36:51

      Hey Tim! Your explanation of why reliability is so important to psychology got me thinking. You raised a great point from the text book about testing consistency in assessment tests we give to clients through test-retest reliability checks. In an ideal world, these assessments would be tested for reliability through a test-retest system many many times before getting cleared to give to a client. However, I wonder what the real process is like to get an assessment test approved. Is there a psychology reliability testing board? Is there a whole company dedicated to making sure these tests work, and if they don’t, who is responsible for making sure bad tests do not circulate in the world of psychology? It also made me question whether assessments that get approved are easily accessible for practitioners, whether or not they cost money, etc. And if they are not easily accessed, this could make psychologists find unapproved tests to administer to underprivileged communities without resources to access approved tests. I wonder how this works!

      Reply

      • Timothy Cody
        Sep 04, 2020 @ 23:22:28

        Hi Zoe,
        You raise some great questions! This must be some issues that psychologists face when they are a part of assessment studies. From what I understand, these types of tests and assessments would not need to be brought to the Ethical Review Board, but it would be unethical to present a study that does not prove consistent results. So if a test does not undergo certain reliability, then one could then say there is no correlation between two variables. I also would not compare results in a within subjects design, but as a between groups as well. In other words, I would administer the same test to other clients so that it is not only that one client that would yield correlating results.

        Reply

  26. Zoe DiPinto
    Sep 03, 2020 @ 20:43:06

    I tried uploading this earlier, but it doesn’t seem to be here so lets try again!

    Correlation does not equal causation! This is a very common mistake that I frequently remind myself of in day to day life. Did you know that increased sales of ice cream is significantly positively correlated with increased crime? The first time I heard this, I desperately wanted to figure out why people who suddenly consumed more ice cream were going around, killing people. Because we are psychologists, we are trained to look at behavior (in this case- crime) and ask ourselves why it is increasing in frequency given the information at hand. However, we must actively be aware of the lack of information that is spoon fed to us. It’s proven that when ice cream sales increase, crime increases. Think about the third variables– ice cream is most often sold in the summer, when people are out and about and the sun is up for longer. The data also doesn’t describe what populations are being affected, the location of sales and crime, or even the type of ice cream or type of crime! Even if it was specified that the same people who were eating ice cream were breaking into people’s homes, we still don’t know the directionality of causation. Maybe after a big score, a criminal wants to celebrate with a nice, cold drumstick cone. Maybe there’s an ingredient in choco-tacos that make young men feverish for a bank robbery. Yes, a silly example, but it raises awareness for psychologists to avoid absurd claims of causality when presented with a correlation.
    Reliability is very important for psychological testing. Without reliability, there would be no standard of testing. Say a mother walks into your office and wants to know if her son needs extra help in school. No problem, you give the son an intelligence test. However, this intelligence test is famous because no matter how many times you administer the test, children receive drastically different scores each time they take it. The test lacks consistency. A metaphor that helps me is the comparison of a reliable test to a reliable parent. If I tell my mom that I want to be picked up from school every day at 3pm, a reliable mom will be there every day at 3pm. An unreliable mom may be there at noon on Monday, 3pm on Tuesday, 10pm on Wednesday, 2am on Thursday, and 3pm without a car on Friday, only to continue to appear at random times the next week. A reliable test is something that you can count on to give consistent information without worrying about back up methods to get the right information. We need reliability in psychology to tell our clients consistent and standardized information so people can receive correct treatment, and trust their provider.

    Reply

    • Cassie Miller
      Sep 04, 2020 @ 22:15:15

      Hi Zoe. I liked your example for reliability because it was very realistic. Often times when trying to understand definitions for specific terminology we get caught up with the fancy nomenclature without truly understanding the meaning behind it. There were many times when my own parents were late picking me up from school, which does not make them bad parents, but rather unreliable. Thus, if we were going to test the reliability of them picking me up at a given time I would say that the score would be low. It is important to know this so that we do not depend so much on the “score,” as we know that it is not consistent.

      Reply

    • Alexa Berry
      Sep 05, 2020 @ 23:58:53

      Hi Zoe,

      I like how you used a concrete example to show how a correlation between two variables does not have to be causal. I think your example was the one I was trying to discuss in my own post but I realize now it was ice cream and crime, not heat and crime. It definitely makes sense that there is a correlation between these variables, but outside factors are important if people are looking for an “explanation” to this relationship. I always thought about how people leave their windows open and unlocked during the summer to beat the heat. Having open/unlocked entry into a household definitely can contribute to increased crime levels. Coincidentally, having ice cream is another activity people do in the summer to escape the heat.

      Reply

  27. Nicole Giannetto
    Sep 03, 2020 @ 22:53:35

    (1) In one of my classes in undergrad we discussed some examples that show why correlation does not equal causation. The one that I remember says that “As consumption of ice cream goes up, so does the number of shark attacks. A little frightening, I know, but if you break it down it starts to become more believable.
    Think about the time of the year where ice cream consumption skyrockets- SUMMER! Now if you set the scene in the summer time you may also see that more people swim in oceans on hot beach days. The increase of people swimming in the ocean during this summer increases the sample size that hungry sharks have to choose from for their meal. So, yes we can say that there is a correlation between increased consumption of ice cream and increased incidents of shark attacks. However, it would be incorrect to say that simply because you eat more ice cream that you will be attacked by a shark. Examining all the variables in the equation is a crucial step in doing good research.
    Now if you tie in this example into the mental health field you may say that someone who is depressed will attempt to hurt themselves. We know that sadly there are many cases where someone who is depressed will attempt to relieve their pain by engaging in self-harm activities, and even suicide. This example is a correlation and not a causation because depression is a dimensional disorder that can affect people differently, and different people can react differently to depression. The danger in assuming that everyone who is depressed will attempt to hurt themselves takes away from the other symptoms of the disorder that may not be equated to inflicting physical harm.

    (2) Reliability describes how consistent something is. When dealing with psychological assessments reliability is quite important because researchers need to know how consistent their instrument is at measuring each specific construct in order to prove its effectiveness or not. When a psychological assessment is reliable, that means that a client will have a better shot at a more successful treatment.

    Reply

    • Zoe DiPinto
      Sep 04, 2020 @ 11:26:30

      Hi Nicole! That’s so funny, the example that you gave of ice cream correlated to shark attacks is very similar to the example I was given in undergrad in which ice cream sales correlated to crime! Your explanation was very clear. I especially appreciated your direct comparison to an example in the mental health field- a comparison my explanation was lacking. You raise a great point about self-harm and depression occurring simultaneously in many patients, and as therapists we should not jump to conclusions about one causing another. This made me think about the different ways a therapist could approach treatment in these individuals. Perhaps the best way is to treat the depression and self harm as two entirely separate symptoms. Thanks for the thought provoking comparison!

      Reply

    • Elizabeth Baker
      Sep 04, 2020 @ 16:34:35

      Hello Nicole,

      I really liked your example of explaining how correlation does not equal causation. It would’ve been funny if we just left it at that, that yes more ice cream consumption increases your chances of going fin to toe with a shark. Ice cream sales would’ve gone WAY down in the summer. Maybe eating and selling ice cream in the winter would become more popular that way. But it’s super important to understand that there are other variables that come into play when two things correlate, so we don’t gain a false understanding.
      I also think you used a very good example of how correlation doesn’t mean causation in the mental health field. It’s important to understand that mental illnesses can affect each individual differently, and to not immediately believe that individuals with depression are engaging in suicidal and other self-inflicting activities. Once again, it’s important to explore the possible reasonings of why the client feeling depressed, that are beyond the event(s) they tell you. Counselors who are trying to rush the process, and hopefully none do this, might give a false diagnosis if they fall victim to the false-cause fallacy.

      Reply

    • Cassie Miller
      Sep 04, 2020 @ 22:06:34

      Hi Nicole! I love your response, especially since we used the same example for correlation does not equal causation. I’m pretty sure I learned that in undergrad as well! What you said about depression and the likelihood that someone will hurt themselves is such a good example. There are so many stereotypes about people with certain illnesses and we often just assume that they fall into one specific category without considering its reliability. Sometimes when I hear about a specific mental disorder I immediately think about a stereotype associated with it before even considering other symptoms of the disorder. We have to train ourselves to look at everything with an individualistic lens, so that we do not let these biases cloud our judgements as counselors.

      Reply

  28. Lina Boothby-Zapata
    Sep 05, 2020 @ 16:46:30

    Hi Christina,
    Ups, I tried to do the reply on Christina DeMallia but didn’t work, doing it again!!
    The clinical example that you used for reliability is interesting, about using instruments with low reliability for diagnoses and as a consequence, assigning the wrong diagnosis and worst-case scenario psychiatrists will prescribe the wrong medication to the client. Something, that it was clear for me after class is that is as a counselor when we search and select the instruments; our selection needs to be based on what we are going to measure (criteria/domain) and if the instrument is reliable. I guess that the instrument’s handbooks are going to be relevant for us. I am assuming that one of the goals of this class is to have the capacity of reading them and understand them all instruments (reliability and validity, how to apply, and results) to select it for our client. Now, let’s assume that it was the right instrument and definitely measure “depression”, next step is to provide the instrument to the client, score it, and communicate the results to the client in terms that she/he can understand. Furthermore, ss you said in your post, instruments can support a diagnosis. I think, that the instruments don’t provide the diagnoses to us, I think the results of the instruments will help us to confirm or denied the hypothesis that we have about the diagnoses. Hence, instruments are an excellent tool that helps us during the treatment, the counselor has the responsibility to know how to use it and provide it a fair value but not an excessive value.

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Adam M. Volungis, PhD, LMHC

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 66 other followers

%d bloggers like this: