[Imtiaz Bhayat]

Ok, thanks. I ran that script. Stata noted that the panel data set was unbalanced.
I have 200 securities, minute by minute data but only when there was a trade. So its unbalanced. One security has 51000 data points (the minimum), and the mas was about 400000. My data set covers a 2 year period. The initial year, and the subsequent year. Im studying the min by min changes in the years. A technological change was implemented in the second year.

[Jared Greathouse]

Okay this makes more sense then. Well then yeah, you can do DD straightforward this way. My only other question for you is, is the minute by minute REALLY necessary? Like why not do the average weekly outcome or the monthly level, maybe? I'm not saying minute by minute data is bad, I'm just curious if it's necessary for your purposes? I think it would make it easier to interpret, too, if you aggregated your time periods

[Imtiaz Bhayat]

A minute by minute analysis facilitates an understanding of the market micro structure. Thats the way the equities market react in particular instances. For example, my phd is assessing the changes in the market 1 year prior compared to 1 year after, when a much faster trading technology was available for participants. Given that 50-70% of trading is done by algorithms, a faster platform makes a big difference. Later today, I'll try and run the Did again, would you mind if I reach out again if I have questions?
I really appreciate you help!