The data from this study, however, indicate The New York Times did make some small -- but statistically significant -- changes in its coverage of women following its 1978 settlement of the Women's Caucus lawsuit. In fact, the changes appear to have begun around 1974, shortly after the Women's Caucus filed its lawsuit and the Times began building its defense. Independent t- tests show the Times significantly increased its female bylines, F(1, 1224) = -4.56, p < .001, and pictures of women, F(1, 603) = -2.82, p < .005.
But the Times did not significantly increase the number of female-authored op-ed pieces or letters to the editor that it printed. Throughout the sample years, an average of 12.2 percent of the op-ed pieces in the Times were written by women and an average of 17.4 percent of the published letters to the editor were by women, with no discernable trend in either area. Nor was there any increase in women's coverage in the one area where women undeniably have parity with men -- death. Over the 28 years of the study, obituaries about women dropped, not only in sheer numbers but also in proportion to men's obituaries. As late as 1994, men continued to die at five times the rate of women -- judging by The New York Times obituaries. This disparity can only be attributed to a news logic bias that does not recognize the importance of how many women spend their lives.
Living women, however, are showing up with increasing regularity in The New York Times. Female references doubled between the time the Women's Caucus lawsuit was filed and 1994, F(1, 1450) = -3.88, p < .001.
Part of the increase in female references is attributable to the increasing prominence of women in politics, business and other facets of life long dominated by men. Multiple regression analysis yields a .15 correlation between female references and the progression of years in the sample, N = 1225, p < .001. The R2 statistic indicates .02 of the variance in female references is due to the advancement of women from 1966 to 1994.
T-tests show a significant drop in the average number of male references made by male writers, F(1, 698) = 2.06, p < .05. The mean of male references by male writers fell from 25.7 per section front in 1966-78 to 23.4 in the latter period. That was partially offset by an increase in male references by female writers, from 3.5 per section front to 4.9, F(1, 778) = -2.73, p < .007. At the same time, male and female writers alike significantly increased their references to women, from means of .8 to 1.5 for female writers, F(1, 777) = -2.48, p < .01, and from 2.7 to 3.5 for male writers F(1, 778) = -2.42, p < .01. Support for the first and fourth hypotheses of this study can better be seen in percentage terms. Male writers referred to women in 9.4% of the names in their stories in 1966-78, and that rose to 13.1% in 1982-94. Female writers, meanwhile, referred to women 18.3% of the time before the lawsuit settlement and 23.2% of the time since then.
Looking only at Page One of each edition, male and female writers significantly increased references to women over time, while overall male references on Page One fell 21%, R = .33, R2 = .11, N = 224, p < .001. Linear regression also showed significant increases in female references on New York Times metro section fronts, R = .19, R2 = .03, N = 109, p < .05, initial arts pages, R = .18, R2 = .03, N = 224, p < .006, and business fronts, R = .13, R2 = .02, N = 224, p < .05.
The advance of time was not the only variable that contributed to an increase in female references in The New York Times. Increased diversity in the newsroom also played a small but significant role. Linear regression showed that female references throughout the paper rose as female bylines increased, while references to women fell when the number of male bylines was higher. A multiple regression of male and female bylines raised the correlation with female references to .24 (from .15 for female bylines alone), while the R2 rose to .06 (from .02). Adding the time variable to the model improved the correlation to .26 percent and the R2 to .07. In other words, the model predicts that female references in the media can be brought closer to parity with male references by adding female writers, reducing the number of male writers and reflecting the increasing prominence of women in society.
A question for further study is whether the model would be enhanced by assuming a time lag between an increase in female bylines and the concomitant increase in female references. There is a moderate correlation when Page One female references are shifted two periods (eight years) behind female bylines. Conversely, it appears even more likely that female bylines lag behind female references; i.e., more women reporters are hired after women become more prominent in the news. The correlation improves when Page One female bylines are shifted one period (four years) behind female references.
Neither of these time-shifted models, however, quite achieve statistical significance, R = .11, R2 = .02, N = 196, p < .12), partly because data at both ends must be thrown out in the new correlation analysis.
Back to Times women index page.