Good morning, I am writing a dissertation for my master's degree in psychology. I have to do statistical studies on data that I have collected. I want to proceed with a correlational analysis.

Hypothesis: The higher the children's autistic trait score, the greater the correlation between perceived predilection interests and perceived cognitive strengths.

I compare the results of a high autistic trait group to a low autistic trait group. I have 9 normal numeric variables and 8 ordinal and therefore non-normal variables. (13 participants)

for the moment I used a spearman correlation but I have a doubt on the choice of my statistical analysis to use because of my ordinal and non-normal variables.

Can I use the spearman method ? if not, what can I use?

Thank you for your help.

  • $\begingroup$ Welcome. What are these variables? Likert scaled ratings? Can you give examples of the data? $\endgroup$
    – AliceD
    Commented Jun 23, 2023 at 13:49
  • $\begingroup$ Yes, the ordinal variable (8 variables) data were obtained using a 5-point Likert scale. They are part of a questionnaire on perceived strengths. Do I have to transfer the data for it to be normal? And if so, how? $\endgroup$
    – lau.
    Commented Jun 24, 2023 at 11:48
  • 2
    $\begingroup$ @lau are the 8 variables you used items from the same scale? i.e. you can derive a sum score for all 8 responses? also, spearman's correlation is non-parametric, meaning there is no assumption of normality. your data can also be ordinal. if you explain a little more about what data you have, it will be easier to help you. i suspect that your small sample size may be an issue. $\endgroup$
    – faustus
    Commented Jun 26, 2023 at 6:12


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.