Presidential Campaigns and Presidential Accountability

Appendix C: Estimated Models for Chapter 5

169 Appendix C Estimated Models for Chapter 5 The regression models used to generate the figures presented in chapter 5 are reported below. Note that party identification was centered at the midpoint, so 0 represents Independents, –3 represents strong Republicans, and 3 represents strong Democrats. Issue priority was also centered at the midpoint, so 0 represents a respondent who gives the issue a “medium” priority, 2 represents a “very high” priority, and –2 represents a respondent who said the issue was not a priority at all. Consequently, the estimates for issue priority represent the effect of prioritizing the issue on the candidate evaluation for an Independent who is situated in a place that has received no candidate advertising on the issue. The estimates on the interactions between candidate emphasis and issue importance, between candidate emphasis and party, and between candidate emphasis , issue importance, and party indicate whether campaign emphasis strengthens the weight given to issue priority in candidate evaluations overall or for supporting partisans. The model can be summarized as (Equation 2): Cand Evalim = β0m + β1m Issue Priorityi + β2m ln(Issue Emphasism ) + β3 Partyi + Xj βj + uim where i indexes individual respondents, m denotes the information environment, designated by a respondent’s media market and time of interview, and Xj includes the individual-level control variables. β1m captures the primary relationship of interest, the relationship between how a citizen prioritizes an issue and her evaluation of the candidate. The weight given to issue priorities, though, is itself modeled as a function of the citizens’ campaign environments. The weight citizens give to a particular issue in producing their evaluations of each candidate is allowed to vary as a function of the attention a candidate has paid to a particular issue within a respondent’s media market, individual partisanship, and the interaction between partisanship and the respondent’s campaign environment (Equation 3): β1m = γ10 + γ11 ln(Issue Emphasism ) + γ12 Partyi + γ13 [ln(Issue Emphasism ) · Partyi ] Substituting Equation 3 into Equation 2 produces the estimated interactive model. The interaction between candidate issue emphasis and individual partisanship is included as well to adhere to the principle of marginality. Table C.1 Tax Priority and Candidate Evaluations Gore Evaluations Bush Evaluations Tax Priority –4.902*** 1.354 (0.845) (0.988) Tax Ads × Tax Priority –0.028 0.493* (0.465) (0.274) Tax Ads × Partisanship × Tax Priority 0.139 0.084 (0.218) (0.131) Tax Ads 0.968* –0.385 (0.544) (0.353) Partisanship × Tax Priority 0.340 0.378 (0.402) (0.470) Tax Ads × Partisanship –0.011 –0.108 (0.269) (0.173) Partisanship 7.099*** –6.497*** (0.557) (0.639) Ideology –0.798** 2.801*** (0.369) (0.370) Economic Perceptions 1.304*** 1.214*** (0.376) (0.376) Education –0.202 –1.554** (0.738) (0.740) Political Awareness 1.669* 1.326 (1.007) (1.014) Female 1.934 –2.476 (1.533) (1.538) Age 0.014 0.004 (0.047) (0.047) Minority 7.026*** –2.046 (2.040) (2.049) Constant 38.751*** 28.419*** (6.257) (6.309) R2 0.393 0.359 N 1023 1019 ***p < .01; **p < .05; *p < .10. Standard errors in parentheses. Table C.2 Education Priority and Candidate Evaluations Gore Bush Evaluations Evaluations Education Priority 2.927*** -2.810** (1.007) (1.253) Education Priority × Education Ads 0.020 0.266 (0.522) (0.302) Education Priority × Education Ads × Partisanship 0.202 0.062 (0.250) (0.141) Education Ads 0.635 -0.430 (0.880) (0.518) Education Priority × Partisanship -1.153** -0.515 (0.458) (0.586) Education Ads × Partisanship -0.030 -0.272 (0.416) (0.239) Partisanship 9.171*** -5.108*** (0.795) (1.001) Ideology -0.858** 2.929*** (0.373) (0.371) Economic Perceptions 1.058*** 1.388*** (0.382) (0.380) Education 0.134 -1.814** (0.739) (0.737) Political Awareness 1.745* 1.257 (1.011) (1.009) Female 0.488 -1.867 (1.550) (1.544) Age 0.047 -0.013 (0.048) (0.047) Minority 6.145*** -0.870 (2.038) (2.043) Constant 32.261*** 31.947*** (6.401) (6.459) R2 0.381 0.353 N 1025 1021 ***p < .01; **p < .05; *p < .10. Standard errors in parentheses. Table C.3 Health Care Priority and Candidate Evaluations Gore Bush Evaluations Evaluations Health Care Priority 2.717*** –1.678 (0.992) (1.103) Health Care Priority × Health Care Ads –0.205 0.251 (0.378) (0.321) Health Care Priority × Health Care Ads × Partisanship 0.013 0.111 (0.175) (0.148) Health Care Ads 1.109* –0.355 (0.577) (0.494) Health Care Priority × Partisanship –0.784* –0.186 (0.436) (0.489) Health Care Ads × Partisanship 0.161 –0.304 (0.265) (0.225) Partisanship 8.469*** –5.704*** (0.704) (0.793) Ideology –0.839** 2.958*** (0.371) (0.372) Economic Perceptions 1.217*** 1.321*** (0.379) (0.380) Education 0.208 –1.895** (0.741) (0.743) Political Awareness 1.396 1.014 (1.019) (1.020) Female 0.551 –2.080 (1.545) (1.547) Age 0.021 0.006 (0.048) (0.048) Minority 5.542*** –0.959 (2.031) (2.048) Constant 32.036*** 29.831*** (6.348) (6.408) R2 0.385 0.347 N 1019 1015 ***p < .01; **p < .05; *p < .10. Standard errors in parentheses. Table C.4 Social Security Priority and Candidate Evaluations Gore Bush Evaluations Evaluations Social Security Priority –0.839 –0.852 (1.009) (1.174) Social Security Priority × Social Security Ads –0.387 0.922*** (0.678) (0.356) Social Security Priority × Social Security Ads × 0.733** 0.004 Partisanship (0.304) (0.170) Social Security Ads 1.416 –0.993* (1.066) (0.574) Social Security Priority × Partisanship –0.693 –0.566 (0.460) (0.527) Social Security Ads × Partisanship –0.650 –0.287 (0.483) (0.276) Partisanship 8.765*** –5.234*** (0.769) (0.877) Ideology –1.040*** 3.016*** (0.375) (0.368) Economic Perceptions 1.296*** 1.261*** (0.384) (0.379) Education 0.062 –1.758** (0.749) (0.743) Political Awareness 1.543 1.253 (1.023) (1.009) Female 1.317 –2.446 (1.573) (1.558) Age 0.049 –0.019 (0.049) (0.048) Minority 5.779*** –1.096 (2.062) (2.047) Constant 35.691*** 29.849*** (6.454) (6.419) R2 0.371 0.356 N 1019 1015 ***p < .01; **p < .05; *p < .10. Standard errors in parentheses. ...