1. Pardon the Interruption: Time-Series Analysis in Program Evaluation. Consider Wagner et al. (2002), which estimated the effect of a three-drug reimbursement limit in New Hampshire's Medicaid Program. Figure 4 shows the time series of the mean numbers of dispensed prescriptions per patient per month in a cohort of 860 NH Medicaid enrollees who received an average three or more drugs per month in the baseline year. In September 1981, the NH Medicaid program restricted the number of prescriptions reimbursed to a maximum of three drugs per patient per month as a measure to decrease state Medicaid program expenditures.

Figure 4

Was the three-drug reimbursement cap effective in reducing dispensed prescriptions in New Hampshire? Using Table 1 below, is the effect statistically significant? How do you know? [10 points]

Table 1

Variable	Definition	Regression Coefficient	Standard Error	t-statistic	p-value
X1	a continuous variable indicating time in months at time t from the start of the observation period	0.003481	0.006791	0.51	0.6128
X2	an indicator for time t occurring before intervention (X2 =0) or after the intervention (X2=1)	-2.5931	0.1572	-16.49	<0.001
X3	a continuous variable counting the number of months after the intervention at time t, coded 0 before the cap and time-20 after the cap	0.0263	0.0193	1.36	0.1849
Constant		5.1389	0.0748	68.69	<0.0001

Source: Wagner, et al. (2002)