McNemar's Test: Sample Size Parameters - Quick Tab

Select the Quick tab of the McNemar's Test: Sample Size Parameters dialog box to access options to establish the basic parameters for analyzing sample size for McNemar's test on two dependent proportions.

Fixed Parameters

The entries in the boxes under Fixed Parameters establish the fixed, or baseline, parameters for subsequent sample size calculations and graphs. Parameters that are not varied explicitly as the independent (X-axis) variables in a graphical analysis will be set equal to these values.

Delta

In the Delta box, enter the difference in the population proportion of times the event of interest occurred during the first measurement, and the population proportion of times it occurred during the second measurement, i.e.,

δ = π₁ - π₂

Eta

In the Eta box, enter the total proportion of times, in the population, different events occur on the two measurement occasions. For example, if 10% of the time the event occurs on the first occasion and not on the second, and 15% of the time the event occurs on the second occasion but not on the first, then Eta is equal to .25. Note that this value is often considered a "nuisance parameter" in this situation.

Alpha

In the Alpha box, enter the type I error rate used in determining the critical value for the t-statistic.

Power Goal

In the Power Goal box, enter the minimum acceptable power, for which a minimum sample size is calculated. If the search for an acceptable sample size is successful, the actual power of the statistical test will be greater than or equal to this value.

Type of Hypothesis

The choice of option buttons under Type of Hypothesis determines the type of null hypothesis tested. Note: McNemar's test is derived using an estimate for the sampling variability of the difference between proportions that assumes the two population proportions are equal. Under the 2-Tailed (Delta = 0) option, this derivation yields a consistent estimator of this variance. In the 1-tailed variants, this estimator is no longer consistent.  McNemar's test is usually performed as a 2-tailed test of the hypothesis that Delta = 0.

2-tailed (Delta = 0)

Select the 2-tailed (Delta = 0) option button to test null and alternative hypotheses of the form

H₀: δ  = 0   H₁: δ ¹ 0

1-tailed (Delta <= 0)

Select the 1-tailed (Delta <= 0) option button to test null and alternative hypotheses of the form

H₀: δ  ≤ 0  H₁: d > 0

1-tailed (Delta >= 0)

Select the 1-tailed (Delta >= 0) option button to test null and alternative hypotheses of the form

H₀: δ  ³ 0  H₁: d < 0