Page 1 of 1

Non Normality of an indicator. What to do?

Posted: Sun May 26, 2024 2:32 am
by IzzyMalik
Hi all,

Recently began a pls-sem analysis in a study examining the acceptance of a mental health app among university students. I assessed for univariate normality and all indicators were between -2 to +2 on skewness and kurtosis statistics. However, one of my variables (Effort expectancy) was heavily highly negatively skewed (over +2). Almost all participants found the app really easy to use (in fact, this is the most common thing we find amongst qualitative studies we've conducted).

My issue is about what to do with this variable and associated indicators before I run my pls-sem analysis. All indicators and the overall mean score of Effort expectancy are over +2 kurtosis and skewness. I'm not sure what to do. Do I remove it from the analysis? Can I transform it? (would this require transforming the rest of the indicators for other variables)?

Any help would be greatly appreciated.