Descriptive statistics for dummy variable

[Jordi Imbrechts]

Dear all

As a master student I'm currently working on my thesis. The data consists of all big Belgian firms (with some filters used in the database) for period 2011-2020 (34 187 firms). The purpose of my thesis is to research motivations for a firm to dismiss more than 10% of the workforce (a layoff). After cleaning my data, correcting for outliers, correlation graph, ... I'm currently working on the descriptive statistics and I'm currently stuck with dummy variables.

My mentor told me that for dummies it is important to look at the t-tests, although I'm not quite sure what she exactly means. All my independent variables are dummies:
1) is there a decrease in ROA? compared to t-1. (=1)
2) Is there a decrease in productivity? compared to t-1. (=1)
3) Is the firm a stand alone firm? (=1)
4) Does the firm has a global owner (variable: "GUO" = 1) or does the firm has a domestic owner (variable: "DUO" =1)?

My dependent variable is also a dummy: wether or not >= 10% of the workforce has been let go.

The descriptive statistics for my control variables and dummies were quite easy till now: #observations, mean, median, min, max ... & also how many layoffs that happend each year & per sector.
Although I'm not quite sure what my mentor meant with "t-tests" for dummies. I hope anyone might be able to help.

Thanks in advance
Jordi


(Stata 16 on Windows 11)

[Carlo Lazzaro]

Jordi:
1) you seem to have a panel dataset;
2) you "corrected" the original sample for outliers (something that is really questionable, as the so called outliers, provided they are not the consequence of a mistaken data entry, may well be perfectly legal given the data generating process you're investigating);
3) -ttest- is for comparing the means of continuous variables. For categorical variable, see -table related entries and helpfiles.