Statistical Test for Comparing Counts/Frequencies in Two Different Time Period?

[Paul Park]

Hi, I am trying to see if there's a statistical test to compare two counts or frequencies from two different time periods.

For example, the total number of surgical operations was 100 in March 2020 and 50 in April 2020.

The absolute difference is 50 and the percent change is (100-50)/100=0.5.

For simple descriptive statistics purposes, is it sufficient to report 50% change from March 2020 to April 2020, or is there a way to find whether that change is statistically significant with a p-value with a hypothesis of predicting a decrease in the total surgical operations from March 2020 to April 2020.

Thank you in advance.

[Joro Kolev]

I do not think you can have such a test based on 2 observations (you have two observations, 100 and 50).

[Mike Lacy]

I'd conceptualize this differently, regarding the number of operations as counts that arise from a Poisson process. If Paul knew the size of the populations at risk for each month, and therefore the amount of person-time from which the operations arose, then the built-in -iri- command would offer tests and CIs for a rate difference or rate ratio.