If the event happened to them, they had to rate items with a five-point Likert scale to assess the impact of the events on them (1 = no impact; 5 = extremely severe impact). Otherwise, they only chose the option ‘not applicable’ and we would assign zero to this option. ...
If the parallel assumption holds prior to the treatment, then the two groups can be compared using difference-in-differences estimation. Fig. 2 shows school attendance for children from married and cohabiting parent families. The rate of enrollment is similar for both groups before treatment; the ...