McNemar test

Mc·Ne·mar test

(mak'ne-mahr),
a form of chi-square test for matched paired data.

Mc·Ne·mar test

(mik'nĕ-mahr test)
A form of chi-square (Χ2) test for matched paired data.
References in periodicals archive ?
Using the McNemar test, a statistically significant change occurred from pretest to posttest (sign test = 28.9, p < .001), with an increase in abstinence as the preferred response and a decrease in the preference for condoms.
Simulation results indicate that the McNemar change test could be robust even if the correction for continuity was not made, and the McNemar test was not as conservative as the Binomial test.
The McNemar test for two or more independent samples (Sheskin, 1997) was used for the inferential comparison estimates between responses in parts five and six.
When the 33 incidents in which drinking by the opponent was unknown were excluded from the analysis, a significant difference between self and opponent drinking was found, with 58.0% of respondents reporting that their opponent had consumed alcohol prior to the aggressive incident compared with 45.4% who reported that they themselves had been drinking prior to the incident (both drinking: 42.9%; neither drinking: 39.5%; only respondent drinking: 2.5%; only opponent drinking: 15.1%; p=.001, McNemar test).
In those cases where a two by two table was inappropriate, a modification of the McNemar test was used.
The McNemar test for significance of changes was used to test for significance of differences in Type I and Type II error rates.
The results of this combined bankruptcy model and that based on accrual accounting data alone were compared using the McNemar test of changes.
Of the 55 persons who responded to the question about whether they had had sex during the previous 30 days, 19 (35%) at time 1 stated that they had not, compared with 29 (53%) at time 2 (p<0.05, McNemar test matching client's responses at time 1 and time 2) (Figure 1).
In group 1, the prevalence at baseline was 38% (41/108) and at follow-up 29% (32/108) (McNemar test, p=0.09).
McNemar test and paired t test were applied to identify statistical significance.
of articles evaluated 200 300 Articles using no statistics, % 30 6 Descriptive methods only, % 12 5 Articles using statistical tests, % 58 90 1 Student t test 19 59 2 Regression/ANOVA 14 42 3 [chi square] test 15 9 4 Mann-Whitney test (rank sum) 9 22 5 Fisher exact test 14 6 Survival analysis 13 14 6 Kaplan-Meier/log-rank 13 13 7 Cox regression 9 8 Multiple comparison adjustment 1 29 8 Tukey 1 11 9 Bonferroni 1 8 10 Newman-Keuls 0 4 11 [kappa] Statistic 4 1 12 ROC analysis 5 2 13 Logistic regression 4 1 14 Spearman rank correlation 3 5 15 Kruskal-Wallis test 4 6 16 Pearson correlation statistic 4 3 17 Normality test 2 2 18 McNemar test 1 0 Other tests 4 5 Avg tests per article (g) 1.3 2.1 Avg tests in articles using tests (h) 2.2 2.3 Journal Category Item Trans (d) Cyto (e) No.
McNemar test with 5% significance level was used to determine the independence of the results and disagreement frequencies found between extraction kits in each round, comparing them individually.