Hello,
I am currently working on my practicum project which consists of analyzing whether health measure of interest can predict SF-36 subscale scores, specifically Mental Health, Social Functioning, and Role Emotional. The data set I am using is a subset of a much larger longitudinal data set and includes 112 observations - values for each visit date of the 20 people that are in the data set. The Social Functioning and Role Emotional variables are not normally distributed and most of the responses are values of 100 which makes the data for these variables impossible to analyze using any type of statistical test. I need to find a way to either stratify these variables or censor them so I can focus on the observations in which the individuals scored anything less than 100, but I am not sure how I would go about doing that. Would I need to weight them by scores or is there a specific way to censor them where values of less than 100 are censored? These are the only two variables that I need to do this to. Any feedback would help. Thank you.
I am currently working on my practicum project which consists of analyzing whether health measure of interest can predict SF-36 subscale scores, specifically Mental Health, Social Functioning, and Role Emotional. The data set I am using is a subset of a much larger longitudinal data set and includes 112 observations - values for each visit date of the 20 people that are in the data set. The Social Functioning and Role Emotional variables are not normally distributed and most of the responses are values of 100 which makes the data for these variables impossible to analyze using any type of statistical test. I need to find a way to either stratify these variables or censor them so I can focus on the observations in which the individuals scored anything less than 100, but I am not sure how I would go about doing that. Would I need to weight them by scores or is there a specific way to censor them where values of less than 100 are censored? These are the only two variables that I need to do this to. Any feedback would help. Thank you.
Comment