Hi All,

I am currently working on a project for my degree using Stata 16. In my research I am attempting to examine the impact of frequency of internet use on trust. My plan is to examine whether there are any generational differences in how internet use affects trust (found to be a negative relationship), more prescisely, whether younger users experience less of this negative effect due to their technical abilities and the relative normality of using the internet to communicate as opposed to older generations.
My main dependent variable, trust, is ordinal on an 11 point scale, going from low to high trust. The main independent variable, daily internet use use is split into 4 categories; < 90 minutes, 90minutes- 3hours, 3-5hours and >5hours. I have also created 4 generational categories which are Millennials& Gen Z, Gen X, Baby boomers and the Silent Generation. Along with a range of socio-demographic controls. I have run an ordered logit model for my preliminary analysis and interpreted the odds-ratios. However, I was interested in examining the marginal effects, however for ease of interpretation I wanted to test the marginal effects of internet use on the probability of observing high trust= p(trust>7) and comparing this marginal effect across generations. I was wondering if there is a way to predict the probability of outcomes above a certain threshold in Stata, rather than predicting the probability of outcomes one at a time.

Any advice on how to carry this out would be much appreciated.

Thank you for your time,

Maya