I have two IVs (computed scores from likert scale) and one DV (computed scores from likert scale) IVs are Depression and Anxiety scores and DV is Social Connectedness scores. These are cintnuous. I have 2 demographic variables Gender (male and female)-categorical, and Marital Status (four categories). The sample size is 332 and when I checked the normality, the Shapiro-Wilk and Kolmogorov's values for one IV and DV was 0.000 which inficates the data is not normally distributed. I want to check the difference between gender, and marital status in terms of Depression, anxiety and social conncetedness, for that I think MANOVA was appropriate but since data is not normally distributed what should i do? I also want to check the interaction effect of gender and marital status on depression, anxiety and social connectedness. Secondly, i want to check the mediating role of anxiety in the relationship of depression and social connectedness. Can it be done on non-normally distributed data using PROCESS? Thirdly, what is an alternative non parametric test for hierarchical multiple linear regression? Please help me, i have no idea what to do with non-normally distributed data. I have also tried 10Log and 2Log transformation procedures but no luck. Thanks in advance
$\begingroup$
$\endgroup$
3
-
1$\begingroup$ The data do not need to be normally distributed. $\endgroup$– Peter FlomCommented May 1 at 12:10
-
$\begingroup$ So can I run Pearson product moment correlation, the two-way MANOVA and hierarchical regression analysis? Actually i was doing it by watching tutorials on YouTube and they said it is necessary. $\endgroup$– Muhammad Muneeb -ur-RehmanCommented May 1 at 12:12
-
2$\begingroup$ Not sure why you need the correlations, but MANOVA and multiple reg. do not require normally distributed data. Don't learn stats from You Tubes. There are some good ones, but not generally trustworthy. Read good books or take reputable classes. $\endgroup$– Peter FlomCommented May 1 at 13:18
Add a comment
|