Not Sure which statistical tests) to use

Hello everybody,

Hope you all are doing well.

We have collected some data from students at our medical college. We asked students to complete the DASS questionnaire to assess depression, anxiety and stress. Students also completed the Rosenberg self esteem scale to assess self esteem and the brief cope scale to assess the ways to coping with stress. Our intention when we planned this exercise was to got two papers out of this: one studying the effects of self esteem on measures of stress and one to assess the relationship between stress and various forms of coping.

My first question is :

Is it OK to discuss these two aspects separately in two different papers or do we have to present our findings in a single paper?

Regarding the first study i.e. the effects of self esteem on measures of stress , What statistical test(s) should be used to analyze/study this relationship ?

Similarly, what test(s) should we use to study the relationship between measures of stress and the various forms of coping?

We have collected results from around 320 students. After excluding those with current medical or psychological illness we are left with around 280 students.

Would really appreciate any and all guidance from the well informed people on this forum.

Thanks in advance.
As I see it, its not an issue if you split one bigger research project into several papers provided that all have meaningful / valuable content (and could be conceptually divided). However, as I understand, you have like 5 self-reporting scales on a number of students which u want to correlate thus I would intuitively feel that this is almost surely not going to be a "fundamental" research even if u put everything into one paper (but this might be only my prejudice towards this kind of surveys). Anyway, given your data (~5 scales), you can do correlations / regression models or maybe some kind of classification (e.g. classify coping strategies using DASS scales as predictors). Its really hard to advice you a statistical means since you was not very specific about the precise hypotheses. But generally, I would recommend you to look for the very basic statistical concepts: correlation and regression.

Best regards, good luck,
Dear Martin,

Many thanks for your informative reply. You are right in saying that the study design does not meet the criterion for true scientific research but it is ok for me as an entry level project. The hypothesis for the first study (if we did decide to divide it into two studies) is that low self esteem will cause an increased likelihood of anxiety, depression and stress in our college students. The hypothesis for the second study would be that maladaptive coping strategies will cause an increased likelihood of anxiety, depression and stress (and vice versa for adaptive coping strategies). Hope this helps to make it a little bit more clear now.

kind regards,
Still waiting for experts to weigh in on this one:rolleyes: Is there something i have missed in the question. I would be glad to add any detail if i know what is missing :confused:.


Omega Contributor

In order for use to make more specific recommendations you need to tell use more about how data are formatted. Do the instruments have a continuous score, is it bounded, are variables nominal, Likert in nature, etc. Also, do you want to control for student characteristics (e.g., gender, age, social support, year in program, etc.).

Otherwise Martin's recommendations of correlations and regressions about sums it up.
OK sure. Thanks a lot for your reply. We have used two instruments. DASS 21 is a scale used to assess stress, depression and anxiety. It has 21 items and each item is scored from 0 to 3. A score is then achieved for each student. The result is then shown as normal, mild, moderate, severe and extremely severe for each category (depression, anxiety and stress). Rosenberg self esteem scale is a 10 item scale which is scored from 1 to 4 and a score is achieved for all the students. I believe both are likert scales but am not 100% sure. We want to control for gender and study year (students included from year 1 to year 5).
I would be really grateful if it could be specified which regression/correlation method is to be used and the tabs under which I look for it in spss.

N.B. When i say control for gender and study year i mean to say that we want to find out if there is any difference between the genders and between the study years on self esteem as well as stress, anxiety and depression measures.


TS Contributor
If you do know not very much about statistics, the it is a bit doubtful that pointing
you to some SPSS menues will be of much help here. SPSS is just a software, it
annot substitute statistical knowledge. You might perhaps find this here helpful,
especially chapters 6 and 7.

With kind regards

OK. Yes i do not know much about statistics but am trying to learn. Thanks for guiding me in the right direction. In the mean while if there is a question i have not been able to answer which can make the query more meaningful then i will be more than happy to answer. :)