Concordance: Black Lawmaking in the U.S. Congress from Carter to Obama

Notes

171 Notes Chapter 1 1. The Poole-Rosenthal or NOMINATE measures have won almost universal acceptance in congressional studies. Lee (2009) is a critic of the NOMINATE methodology, contending that it measures not ideology but “team party cooperation .” The NOMINATE methodology includes the universe of all bills, but Lee points out that some bills are nonideological, reflecting a spoils system rather than intrinsic policy preferences. She advocates classifying bills by ideological content, which she did for about 40 percent of all roll-call votes from 1981 to 2004. Her data are limited to the U.S. Senate. I agree that additional coding work needs to be done to support my broad claims. Yet even if NOMINATE scores measure only “team support,” it seems clear that Black Democratic House members have moved closer to their team. This shift, I contend, is rooted in their political incorporation. Furthermore, my analysis in chapter 7 introduces party unity measures, which are not identical to the NOMINATE scores but also predict, as expected, more liberal scores. 2. In appendix B, I report the average scores of the second dimension for Black House Democrats from 1977 to 2010. The gap for average second dimension scores between CBC members and the House Democrats is very small relative to the gap for the first dimension. After the 1980s, these second dimension scores no longer meaningfully predict the voting behavior of members; therefore, I chose not to analyze the second dimension scores of Black Democrats in the House. Chapter 2 1. Quoted in Trescott 1977a. 2. See the October 5, 2011, publication on TED for these data, http://www. bls.gov/opub/ted/2011/ted_20111005_data.htm. 172 • Notes to Pages 85–135 Chapter 5 1. “Can Bush Mend His Party’s Rift with Black America?” New York Times, December 17, 2000, section 4, 17. Chapter 7 1. In the case of autocorrelation, all the key or statistically significant findings are most likely wrong (see King 2001).To test for autocorrelation, dummy variables for all the years as “fixed effects” were constructed and included to see whether results emerging from the multivariate analysis were different from the first model. ...