I am trying to find the relationship between the use of defence mechanisms in early adulthood as predictors of neuroticism. Those who use more defence mechanisms will score higher on the neuroticism scale questionnaire.
To test this I have used the DSQ40 which uses a likert scale rating and the Neuroticism Scale Questionnaire which has "yes", "no" and "in-between" as the answer options.
However, I am now struggling with which statistical analysis would be more appropriate for my study?
I am thinking of using the Spearman's correlation test but would appreciate if anyone could let me know if there is a more suited analysis for this study?
Thanks in advance.
Depending on the distribution of normality of your data you can evaluate with shapiro-wilk, if it comes out significant use spearman but pearson, another more powerful model is to make a correlation through analysis of structural equations
As I understand the DSQ40 , depending upon their mean scores on this questionnaire, categorizes the respondents into 3 categories: Mature, Neurotic, or Immature.
(See: http://www.web.pdx.edu/~dcoleman/dsq.html ).
Yes, you could calculate the Spearman rank correlation.
Yet, as you most probably assume ‘defence mechanism’ is a predictor of Neuroticism (on the Neuroticism Scale Questionnaire (NSQ, https://psycnet.apa.org/record/2005-15367-002 ) ,
you could use an ANOVA (F-test) with the 3 categories of the DSQ40 scale as independent variable and the NSQ score as dependent variable,
maybe seperately for the subscales: (…NSQ …) yields part scores on Tender-Mindedness, Depression, Submissiveness, and Anxiety as well as a total Neuroticism score. See: https://psycnet.apa.org/record/2005-15367-002
Alternatively, you could calculate a chi-square test on the number of respondent in the cross tabel of
- NSQ-score categorized repondent into Neurotic vs. Other diagnosis vs. Normal (See: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.2044-8260.1965.tb00466.x It was found that the test separated the neurotic patients from the other two groups at a high level of confidence (p < 0.01).
By
- Mature vs. Neurotic vs. Immature, on the basis of the DSQ40 mean scores
You should first consider using the regular Pearson's correlation, and then check to see if the residuals of the correlation follow a normal distribution using the Shapiro-Wilk and/or Kolmogorov-Smirnov tests of normality. Under normal circumstances with college students, you should not see any deviation from normality, that is, p should be more than .05 for one correlational test. if normality is violated, apply Spearman's rank-ordered correlation. In that case, you need to rank the values belonging to each of the two sets of variables in ascending order (e.g., 1 for lowest value, 2 for second lowest value, and so on) and then correlate the two sets of ranked data. In other words, you need a separate column of ranked data for each variable.
Cristian Ramos-Vera thank you so much for your answer! It has been really helpful!!
Jos Feys I cannot thank you enough for the response you have given me, it has answered all my questions and been a massive help. I appreciate it so much, thank you!
Jimmy Y. Zhong thank you so much! thanks for such a helpful and clear response I am very thankful!
The following is a good one for your research
Elements of Forecasting,
by Francis X. Diebold.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis