What statistical analysis to use? Parametric or Non-parametric?

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Vangelis55
Posts: 2
Joined: Thu Mar 09, 2017 4:23 am

What statistical analysis to use? Parametric or Non-parametric?

Postby Vangelis55 » Thu Mar 09, 2017 4:33 am

For my research in psychology, I created an assessment tool to measure symptom X in a clinical (n = 10) and a non-clinical (n = 10) sample. I want to verify whether the overall average score obtained by the clinical group on the measure differs from that of the non-clinical group.

The measure involves a 1 to 5 scoring scale which is applied four times to the same person on four tasks; as such, four scores are obtained per person from which I derive one final average score per person on the measure (i.e., scores of 3, 3, 2, 1 = 9; 9/4= 2.25, so 2.25 is the person's average score on the measure).

In SPSS, my first variable is each participant's final average score on the measure. My second variable is their group (clinical = 1 or non-clinical = 2). I think I would be using a non-parametric t-test for 2 independent samples due to how small my sample is; however, I am unsure. Can I use a parametric t test for 2 independent samples? I know I need to verify whether my distribution is normally distributed or not. I have done this doing a visual scan of a histogram for each group, and my distribution appears normal for the non-clinical sample (as per a visual scan of data); however, it is skewed for the clinical-sample. I am unsure how to proceed to answer my initial question, or whether there might even be a more optimal way to organize my data in SPSS.

Any help is greatly appreciated!

OS = Windows 8, and I have used version 17.0 of SPSS.
Last edited by Vangelis55 on Sat Mar 11, 2017 3:43 am, edited 1 time in total.
statman
Administrator
Posts: 2750
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: What statistical analysis to use? Parametric or Non-parametric?

Postby statman » Fri Mar 10, 2017 1:38 pm

Based on what I understand you to say/describe, and the case count, look into a nonparametric independence test

Also .............
See the note below

NOTE: Please read the Posting Guidelines and always tell us your OS, the SPSS version and information about your study and data!

Statman
Statistical Services
Vangelis55
Posts: 2
Joined: Thu Mar 09, 2017 4:23 am

Re: What statistical analysis to use? Parametric or Non-parametric?

Postby Vangelis55 » Sat Mar 11, 2017 3:46 am

statman wrote:
Fri Mar 10, 2017 1:38 pm
Based on what I understand you to say/describe, and the case count, look into a nonparametric independence test

Also .............
Sorry about this. I have updated the original post.

If it's alright, I have a follw-up question.. I have used the Mann Whitney Test and my results are statistically significant (p < 0.05). Given the Mann Whitney looks at medians rather than means, I am unsure how to report my results because I was originally examining differences in the mean score obtained on a measure between two groups. As such, all the data I entered into my variable reflected the average scored obtained by each participant on measure X rather one single score.

Would I report that the medians on the measure used differ significantly between my clinical and non-clinical groups; or, would I report that the mean scores obtained on the measure differed significantly between both groups? I guess I am just unsure how to report my results in terms of this detail (which I think is significant).

Thank you so much for any feedback.

Who is online

Users browsing this forum: No registered users and 2 guests

cron