In model B, we estimate the effects of the average number of new cases in the past 14 days instead of using two separate lag variables.