Hi all,
I am new to STATA and have a question regarding how to best analyze some data using STATA and multinomial logistic regression in particular.
To put my question in context, I am interested in the reasons people report for why stressful events do not occur (e.g., whether they avoided stressful events vs just got lucky etc.) and whether differences in reasons for the non-occurrence of stressful events is attributable to age. Within my analyses, my dependent variable is reason for non-stress days (categorical variable with 5 possible nominal response options) and my independent variable is age group (younger vs older adults). I planned on using multinomial logistic regression to examine age as a predictor of reasons for why stressful events did not occur.
My question regards how I handle this analysis if I have nested/clustered data. To elaborate, I have data collected over a 20 day period for each participant (i.e., each day participants reported whether or not a stressful event occurred and if a stressful event did not occur what they attributed this to), therefore my observations of reasons for non-stress days are clustered within individuals. As I am new to STATA I am trying to keep my analyses as simple as possible so I was wondering if I could use the vce (cluster clustvar) function to account for the non-independence of observations within individuals? Is this approach appropriate or should I be using a multilevel approach?
Any help/suggestions will be much appreciated.
Thank you in advance,
Victoria.
I am new to STATA and have a question regarding how to best analyze some data using STATA and multinomial logistic regression in particular.
To put my question in context, I am interested in the reasons people report for why stressful events do not occur (e.g., whether they avoided stressful events vs just got lucky etc.) and whether differences in reasons for the non-occurrence of stressful events is attributable to age. Within my analyses, my dependent variable is reason for non-stress days (categorical variable with 5 possible nominal response options) and my independent variable is age group (younger vs older adults). I planned on using multinomial logistic regression to examine age as a predictor of reasons for why stressful events did not occur.
My question regards how I handle this analysis if I have nested/clustered data. To elaborate, I have data collected over a 20 day period for each participant (i.e., each day participants reported whether or not a stressful event occurred and if a stressful event did not occur what they attributed this to), therefore my observations of reasons for non-stress days are clustered within individuals. As I am new to STATA I am trying to keep my analyses as simple as possible so I was wondering if I could use the vce (cluster clustvar) function to account for the non-independence of observations within individuals? Is this approach appropriate or should I be using a multilevel approach?
Any help/suggestions will be much appreciated.
Thank you in advance,
Victoria.
