Table 4 Percentage of Rejected Replications with Different Levels of Effects and Measurement Error (n = 2,000)

Table 4 Percentage of Rejected Replications with Different Levels of Effects and Measurement Error (n = 2,000)

Excel | CSV

error sα sP
sI
H 0
0
H
f
Wn H 0.3µP
0
H
f
Wn H 0
0.3µI
H
f
Wn
C 0   5.8 5.6 4.3 4.4 4.6 3.6 5.6 5.6 4.4
C 0.2   12.4 12.2 9.4 11.5 11.6 8.9 11.1 10.6 8.8
C 0.4   57 52.7 47.5 58.1 54 48.6 56.2 52.2 48
C 0.6   99.4 98.7 98.2 99.2 98.5 97.8 99 98.4 97.9
B 0   5.6 5.8 4.6 3.2 3.2 2.6 5.5 5.4 4.5
B 0.2   12 11.6 9 9.8 9.4 7.4 11.8 11.2 9.3
B 0.4   58.4 54.8 49.1 50.2 47 42.3 58.6 53.7 48.4
B 0.6   99.2 98.6 98.1 98.8 97.9 97 99.2 98.6 98.1

Key: C = classical measurement errors; B = Berkson measurement errors.

Note: The null hypothesis was rejected for large values of the test statistic at 5% level (one-tailed test).