THE Z-TEST for binomial data indicates whether the rates of behavior of two groups are significantly different. Specifically, the Z-test for binomial data allows us to compare two discrete random variables. Where n and m are the sample sizes for two independent groups (such as Democratic and Republican House members) and x and y are the number of successes (members cosponsoring any pro-women's rights bills, for example) experienced by each, respectively, we can reject the null hypothesis (H0: px = p) if
where α is the level of significance (Larsen and Marx 1986, 378-82).
The New York Times index is employed to derive a measure of the level of public attention to women's rights as reflected in media coverage, employing a methodology loosely adopted from Baumgartner and Jones (1993). To track changes in the level of press coverage of varying issues over an extended period of time, I code all items listed under the heading of “women” in the annual New York Times index. Baumgartner and Jones replicate the process with the index to the Reader's Guide and report strikingly similar findings across the two sources. I am thus confident in limiting data collection to the New York Times.
The New York Times index provides a brief abstract of each item listed, in addition to basic descriptive information (data and placement). Using that abstract, I determine whether each story involved women's rights. Consistent with the definition of women's rights employed in this research, a story was coded as relevant to the women's rights debate if it involved questions of women's social, economic, or political role; referenced a