The Project

I'm doing an analysis concerning a major drop in service uptake since the beginning of COVID. I have access to administrative data for those using services, including relevant demographic characteristics. We know that fewer people are using the services. The question is: what are the characteristics of those who have stopping using services since COVID.

To start, I've defined a time-period based COVID cohort. The initial plan is simply to compare aggregate stats for the COVID-cohort, and then compare to a pre-covid cohort from 2019. This would allow we to say things like, pre-covid service users were older, more likely female, etc.

The Stumbling Block

I'm trying to find an appropriate research design that will allow me to develop a multivariate profile of the people who did not use services since COVID. Is its young unmarried men without children who stayed away? Aside from doing an exhaustive series of cross tabs, I'm having a hard time identifying the right approach. I've thought about using cluster analysis to identify different groups of people who used services pre-COVID, and then see which of these cluster the post-COVID cohort fall into. The goal would be to look for pre-COVID clusters that have fewer associated people in he post period. I've also thought about doing some kind of applied propensity score matching, looking at characteristics of people from 2019 who do NOT match to people from 2020. After many hours searching around, I have not seen either of these techniques used this way, nor any other methods for this type of research question.

Does anyone have any suggestions for an appropriate research design? Do either of my ideas sound plausible? Any examples of other studies with a similar question?

Many thanks is advance, please let me know if anything is unclear.


1 Answer 1


If you have an established cohort, and assuming the demographic covariates of interest between the time points are static, why not try a multi-class classifier? In this case, you would have four classes of interest between individuals:

  • Using services T1, using services T2
  • No services T1, no services T2
  • Using services T1, no services T2
  • No services T1, using services T2

Something decision-tree based like a random forest would help accelerate the exploratory data analysis, potentially finding complex non-linearities or interactions among the available covariates.

Your point around looking how cluster prevalence changes over time brings up the possibility of something like a latent transition analysis, but I am not sure that is the most effective technique given how you laid out the original question.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.