Lambda computed on a bivariate table was .34.Lambdas computed on partial tables were .02 and .05.What step should the researcher take next?
A)disregard Z and select a new control variable
B)incorporate Z into the analysis
C)disregard Y and select a new dependent variable
D)Z should become the new dependent variable

Question

Quizplus · Accepted Answer

Since lambdas computed on partial tables are significantly different from the lambda computed on the bivariate table, it suggests a contingency relationship between the control variable Z and the dependent variable Y. Hence, incorporating Z into the analysis through multivariate analysis (e.g., logistic regression or chi-square analysis) would be the best step.

Lambda Computed on a Bivariate Table Was