I want to test for the weakening association between two variables with longitudinal data.
I have a survey that measures trust in government institutions each year between 2005 and 2020. I also have county-level data on the crime rate. The crime rate grew in most counties during this period. To no-one surprise, the crime rate is associated with lower trust in government institutions. My hypothesis is that a few years in the association between crime rate and trust in government institutions weakens. In other words, people “get used” to crime. I should thus be able to track how the association between local crime rate and the outcome variable weakens over time. How can I test for this? The low hanging fruit tactic would be to do a cross-section model of each year and compare the t-value of the variable of interest.
Remember, this is not a panel but pooled data.
