For anyone who knows anything about the STAI scale.

I'm reviewing an study for a uni project. The study uses STAI to assess anxiety, gives them a self-administered intervention for 30 days, gives them a STAI test again, then gives a further 30 days treatment, then test them with STAI again. The results are all over the place!

My question is, does the STAI take into considerations determinants of anxiety? Does the STAI indicate whether they are anxious long term or just on the day of the test? Basically, how do the researchers know that it is their intervention that has changed the anxiety score (without the participants keeping a diary of their determinants of anxiety)? Also, is state and trait scored usually given separately?

Hope that makes sense! Any suggestions of research articles on the subject would be useful too!

Thanks!

Just tell 15 year olds to stay off of cigarettes and anxiety will never manifest itself in the first place.

Jenniferocious wrote:For anyone who knows anything about the STAI scale.

I'm reviewing an study for a uni project. The study uses STAI to assess anxiety, gives them a self-administered intervention for 30 days, gives them a STAI test again, then gives a further 30 days treatment, then test them with STAI again. The results are all over the place!

What do you mean by "all over the place"? As in, each patient seemed to get better and worse over time, or the average scores of the patients increased and decreased over time?

Jenniferocious wrote:My question is, does the STAI take into considerations determinants of anxiety?

I hadn't heard of this measure before so keep in mind that my comments here are based solely on the information I can find on it in books and articles that are available online. Specifically, I found "State-Trait Anxiety Inventory" and "Practitioner's guide to empirically based measures of anxiety" to be pretty useful.

From what I've read, the scale does not attempt to determine the cause of the anxiety. Instead it is simply a measure of anxiety, in the same way that the Hamilton rating scale for depression only measures the severity of depression, not the cause. It might be able to help us figure out what the cause is, as it distinguishes between state and trait anxiety which would presumably have different causes and comorbidities, but I don't think the test itself tells us what the cause is.

Jenniferocious wrote:Does the STAI indicate whether they are anxious long term or just on the day of the test?

Looking at the links I've presented above, the test is divided generally into two types of questions: how the person feels right now, and how they feel in general (which is the 'state' vs 'trait' distinction). This would give the researchers a measure for both short term and long term anxiety.

Jenniferocious wrote:Basically, how do the researchers know that it is their intervention that has changed the anxiety score (without the participants keeping a diary of their determinants of anxiety)?

Well that comes down to experimental methodology. For most research, randomised controlled trials are enough to establish causation, where they use one group as a control that receives no treatment to compare to the group that receives the treatment. If they successfully control for all other possible confounds (age, gender, social economic status, etc), then the only difference between the control group and the treatment group should be the intervention you've introduced - which means that any difference must be caused by the intervention.

There are other research designs which help us establish causation as well. For example, if we're looking at individual differences then we might not want to use average scores of groups, and instead we'd use a "within-subject" design, where the subject acts as their own control. This means that we take a baseline score, introduce the treatment and measure again, then remove the treatment and measure again, and then introduce the treatment again and measure again etc. So we're looking at how the individual compares to their own previous score, rather than comparing the scores of one group of people to the scores of a completely different group. If done correctly, you can determine what variable is causing the effect.

Jenniferocious wrote:Also, is state and trait scored usually given separately?

According to the book link above, the state version is administered prior to the trait version, and the test gives us an "S-anxiety" score and a "T-anxiety" score.

Sorry I don't know enough to provide any specific details! If you want to link to the particular article you're attempting to review, I should be able to discuss the research design more thoroughly to determine whether they are measuring what they say they're measuring. From what I could find on the STAI though, it seems to have good construct validity, test-retest reliability, etc, so it appears to be a decent assessment.

Wow, that's really helpful Mr.Samsa, thank you!

Here's the link for the article http://www.scielo.br/pdf/ape/v24n2/en_16.pdf

What I mean by 'results all over the place' is that they are up and down for all three groups at both check points, and the standard deviation is large (as you will see).

They make a lot of excuses for the non significance of their results and high drop out rate, but I personally think the STAI was not used effectively, i.e. how do we know if some of the participants (students) were having exams, perhaps in the first 2 weeks of the study but not the last 6 weeks or vice versa. The only problem is, is that my review is for Uni and they don't want my opinion, I need references to show that the STAI doesn't work effectively in this kind of study.

Thank you for all your help!

Jenniferocious wrote:Wow, that's really helpful Mr.Samsa, thank you!

Here's the link for the article http://www.scielo.br/pdf/ape/v24n2/en_16.pdf

What I mean by 'results all over the place' is that they are up and down for all three groups at both check points, and the standard deviation is large (as you will see).

Yes... Those results are bad. The results were not statistically significant, so even ignoring methodological issues, the authors should conclude that they found no evidence to support their ideas. As a general comment at the end, they could suggest that there appeared to be a positive trend that further, more rigorous, research might be able to discover, but this would clearly be identified as speculation and not a conclusion of their research.

Jenniferocious wrote:They make a lot of excuses for the non significance of their results and high drop out rate, but I personally think the STAI was not used effectively, i.e. how do we know if some of the participants (students) were having exams, perhaps in the first 2 weeks of the study but not the last 6 weeks or vice versa. The only problem is, is that my review is for Uni and they don't want my opinion, I need references to show that the STAI doesn't work effectively in this kind of study.

Thank you for all your help!

Yeah, but unfortunately the horribleness of their research and their non-results aren't overly important to you here. That is, the quality and conclusions of the research are independent of the question of whether the STAI accurately measured anxiety. As for the idea of some students having exams whilst others don't, this would be a methodological issue and presumably would be ruled out by the randomisation of subjects (i.e. if done correctly, there should be no reason why one group would have a higher rate of students with exams compared to another). The high drop out rate would also be a significant problem for the study, as we'd probably expect the more anxious students to be more likely to drop out which would bias their results, but again this wouldn't change whether the STAI was a valid measure or not. Essentially, when assessing the validity of the STAI, the question you are asking is: "Does the STAI accurately measure anxiety?".

If I were to criticise the use of the measure, I suppose I would argue that using it in a population with no apparent anxiety problems, then it may not have even been possible to statistically reduce anxiety levels to a statistically significant level so an STAI measure wouldn't help. As a comparison, even though drinking water obviously reduces dehydration, getting a population of people who already drink enough water to offset dehydration effects and getting them to drink more water at regularly intervals will probably give us the result of no difference between dehydration before and after the intervention of drinking water. But obviously we'd be wrong to conclude that drinking water doesn't reduce dehydration here. In other words, maybe you could argue that the STAI is only appropriate for clinical populations.

More directly related to the STAI itself, the second link I presented in my post above suggests that there is some concern that the STAI is a better measure of depression rather than anxiety, so perhaps you could argue that if the aromatherapy has no effect on depression, then even if there was a reduction in anxiety then the students would still score similarly at the follow-up.