Cochran’s Q with Pairwise McNemar for Dichotomous Multiple Responses Data: a Practical Approach

 
 
 
  • Abstract
  • Keywords
  • References
  • PDF
  • Abstract


    When utilizing single-response questions for a survey, researchers often overlook the possibility that an item can have a smorgasbord of viable answers. It results in the loss of information as it forces the respondents to select a best-of-fit option. A multiple-responses question allows the respondent to select any number of answers from a set of preformatted options. The ability to capture a flexible number of responses allows collectively exhaustive concepts to manifest for deductive verification. This paper explores the practical use of Cochran’s Q test and pairwise McNemar test to examine the proportion of responses derived from the results of Multiple Responses Analysis (MRA). This includes Cochran’s Q operation on MRA data table using a simulated data set. Cochran’s Q test detects if there is a difference in the proportion of multiple concepts. In the case of a significant result, it would require a post hoc analysis to pinpoint the exact difference in pairwise proportions. This pairwise difference can be detected by utilizing pairwise McNemar test with Bonferroni Correction. This paper serves as a reference for researchers and practitioners who need to examine the proportion of collectively exhaustive concepts collected from a multiple responses item.

     

     


  • Keywords


    Cochran’s Q, dichotomous multiple responses data, McNemar, Multiple Responses Analysis, proportion

  • References


      [1] Jann B (2005), Tabulation of multiple responses. Stata Journal 5(1), 92–122.

      [2] Plašil M, Vlach P (2006), Analysis of multiple-response data. In 9th International Scientific Conference on Applications of Mathematics and Statistics in Economics. Trutnov.

      [3] Cochran WG (1950), The comparison of percentages in matched samples. Biometrika 37(3), 256–266. http://doi.org/10.1093/biomet/72.2.241

      [4] Tate MW, Brown SM (1970), Note on the Cochran Q test. Journal of the American Statistical Association 65(329), 155–160. http://doi.org/10.1080/01621459.1970. 10481069

      [5] Yarnold PR (2015), UniODA vs Cochran’s Q test: Comparing success of alternatives. Optimal Data Analysis 4(7), 116–117.

      [6] Sheskin DJ (2007), Handbook of Parametric and Nonparametric Statistical Procedures. CRC Press, Boca Raton. http://doi.org/10.1201/9781420036268

      [7] Fagerland MW, Lydersen S, Laake P (2013), The McNemar test for binary matched-pairs data: Mid-p and asymptotic are better than exact conditional. BioMed Central Medical Research Methodology 13(91), 1–8. http://doi.org/10.1186/1471-2288-13-91

      [8] Yarnord PR (2015), UniODA vs. McNemar’s test for correlated proportions: Diagnosis of disease before vs after treatment. Optimal Data Analysis 4, 24–26. http://doi.org/10.1007/BF02295996

      [9] Cohen JF, Chalumeau M, Cohen R, Korevaar DA, Khoshnood B, Bossuyt PMM (2015), Cochran’s Q test was useful to assess heterogeneity in likelihood ratios in studies of diagnostic accuracy. Journal of Clinical Epidemiology 68(3), 299–306. http://doi.org/10.1016/j.jclinepi.2014.09.00

      [10] Santos JR (2000), Getting the most out of multiple response questions. Journal of Extension 38(3), 3–8.

      [11] Lewis-Beck M, Bryman AE, Liao TF (2003), The SAGE encyclopedia of social science research methods. SAGE Publications, Thousand Oaks.


 

View

Download

Article ID: 16662
 
DOI: 10.14419/ijet.v7i3.18.16662




Copyright © 2012-2015 Science Publishing Corporation Inc. All rights reserved.