- 1,2 Okeh UM*, 3 Nwuzor O and 4 Okoh S.
- 1 Department of Industrial Mathematics and Health Statistics, David Umahi Federal University of Health Sciences (DUFUHS) Uburu, Nigeria. 2 International Institute for Nuclear Medicine and Allied Health Research, David Umahi Federal University of Health Sciences Uburu Ebonyi State, Nigeria. 3 Department of Industrial Mathematics and Applied Statistics,Ebonyi State University Abakaliki, Nigeria. 4 Department of Statistics, Enugu State Polytechnic Iwolo,Enugu State, Nigeria.
Background
The Cochran’s Q test is a statistical procedure for assessing the conformity of proportions for various samples in dichotomous outcomes. This process extends from the traditional McNemar test for paired data.
Method
In other to use Cochran’s Q test, paired and independent samples are assumed while the outcome variables are dichotomous also. This test calculates the Q test statistic, which calculatesunder the null hypothesis the difference between the observed and expected frequencies. The test statistic is then compared to a Chi-squared distribution to determine the p-value. In order to control for Type I Error, the study employed extended McNemar test from a study carried by Okeh and Obiora-Ilouno. These drug preparation samples were paired and a comparison study was carried out using the extended McNemar tests.
Results
Cochran’s Q test was applied on the drug preparation data and result concluded the acceptance of Ho of equality of drug samples. We conducted a comparison of the paired samples to determine the contributions of various pairs of the sample data using extended McNemar tests. Results showed that patients did not improve equally for some of the pairs of drugs preparations. This is a deviation from the result obtained when Cochran‘s Q test was applied on the data.
Conclusion
The use of extended McNemar test for the pairwise comparison of the data clearly identified sample pairs which their contributions would have made the study to suggest the rejection of Ho of equally of drug preparations, thus controlling for Type I Error. This is particularly a significant role usually played by McNemar test.
