Parallel-trends Assumption

[Malik Saendman]

Hello everyone,

I've conducted a difference-in-differences analysis to examine job satisfaction among workers before and after the COVID-19 pandemic. However, I'm uncertain about the validity of the Pre-treatment Trend and Parallel Trends Assumption in my study. Despite extensive reading on the topic, I'm struggling to reach a conclusion due to a minor divergence in my data. Below are the plot and a parallel-trends regression. Could you assist me in determining if my difference-in-differences approach is valid?

[Jared Greathouse]

I don't mean to be a wet blanket, but I don't think parallel trends is the highest of your concerns. Everyone was exposed to the pandemic, pretty much most labor forces were exposed to it and deeply impacted by it. So how can you argue there's an untreated group when I couldn't think of a single labor sector that wasn't affected by the pandemic?