Available on Request

Available on Request

Models with Volume and Proportion Included Separately

Table 3: Tests of the Effects of Negative Advertising on Information about the Candidates

Ability to place pres. candidates ideologicallyb / Certainty of candidates’ ideologyd
Logged volume of exposure to negative ads / .028 (.016)# / -.004 (.012)
Proportion of negative ads to which exposed / .002 (.001) / -.002 (.001)
Total spots in market / -.028 (.046) / .002 (.039) / .033 (.033) / .031 (.028)
Local news viewing / -.008 (.014) / -.004 (.013) / .011 (.011) / .013 (.010)
Newspaper reading / .000 (.011) / .000 (.011) / .006 (.009) / .006 (.009)
Age / -.007 (.002)** / -.007 (.002)** / -.001 (.001) / -.001 (.001)
African-American / .092 (.092) / .086 (.091) / -.027 (.056) / -.025 (.057)
Female / -.209 (.059)** / -.211 (.060)** / -.069 (.044) / -.064 (.044)
Income / .000 (.000)# / .000 (.000)# / .000 (.000) / .000 (.000)
Education / .104 (.020)** / .103 (.020)** / .034 (.016)* / .034 (.016)*
Political information / .273 (.019)** / .273 (.019)** / .094 (.014)** / .094 (.014)**
Strength of partisanship / .494 (.077)** / .496 (.075)** / .161 (.043)** / .160 (.043)**
Mobilized / .171 (.059)** / .172 (.059)** / .021 (.055) / .015 (.056)
House race competitive / .081 (.155) / .076(.115) / .108 (.074) / .098 (.076)
Senate race competitive / .035 (.085) / .019 (.084) / -.007 (.049) / -.003 (.048)
Presidential race competitive / -.002 (.088) / .044 (.090) / .032 (.044) / .040 (.046)
Constant / .583 (.422) / .301 (.376) / 1.246 (.287)** / 1.259 (.254)**
N / 1200 / 1200 / 1046 / 1045
Adjusted R2/ Pseudo R2 / .36 / .36 / .14 / .14

**p < .01 *p < .05 #p < .10 (two-tailed tests) Standard errors are clustered by media market (DMA) and state

Cell entries are all OLS estimates

a exposure measures estimated based on all ads to date of interview b exposure measures estimated based on presidential ads to date of interview

c exposure measures based on all ads d exposure measures based on all presidential ads

Table 3: Tests of the Effects of Negative Advertising on Information about the Candidates (continued)

Bush/Gore likesb / Bush/Gore dislikesb
Logged volume of exposure to negative ads / .081 (.030)** / .123 (.031)**
Proportion of negative ads to which exposed / .006 (.002)** / .008 (.003)*
Total spots in market / -.099 (.082) / -.014 (.083) / .081 (.085) / .218 (.081)**
Local news viewing / .026 (.028) / .038 (.025) / -.019 (.023) / -.000 (.022)
Newspaper reading / -.011 (.020) / .012 (.020) / .003 (.020) / .002 (.021)
Age / .009 (.004)# / .009 (.004)# / -.006 (.003)* / -.006 (.003)*
African-American / -.281 (.207) / -.297 (.210) / -.063 (.153) / -.083 (.156)
Female / .236 (.119)# / .230 (.124)# / -.055 (.104) / -.059 (.109)
Income / .000 (.000) / .000 (.000) / .000 (.000) / .000 (.000)
Education / .172 (.037)** / .169 (.037)** / .188 (.037)** / .183 (.037)**
Political information / .240 (.036)** / .240 (.036)** / .379 (.039)** / .379 (.040)**
Strength of partisanship / .708 (.124)** / .712 (.123)** / .153 (.173) / .162 (.173)
Mobilized / .275 (.113)* / .277 (.112)** / .282 (.114)* / .282 (.112)*
House race competitive / -.191 (.167) / -.206(.161) / -.294 (.151)# / -.330 (.154)*
Senate race competitive / -.065 (.141) / -.113 (.135) / -.245 (.133)# / -.324 (.135)*
Presidential race competitive / .018 (.147) / .151 (.148) / -.309 (.130)* / -.100 (.152)
Constant / .875 (.702) / .059 (.724) / .104 (.813) / -1.214 (.733)
N / 1200 / 1200 / 1200 / 1200
Adjusted R2/ Pseudo R2 / .16 / .16 / .23 / .23

**p < .01 *p < .05 #p < .10 (two-tailed tests) Standard errors are clustered by media market (DMA) and state

Cell entries are all OLS estimates

a exposure measures estimated based on all ads to date of interview b exposure measures estimated based on presidential ads to date of interview

c exposure measures based on all ads d exposure measures based on all presidential ads

Table 4: Tests of the Effects of Negative Advertising on Attention to the Campaign, Perceptions of Government, and Voting

Attention
Remember seeing an ada / Attention to the campaigna / Perceptions of closeness of
pres. raceb
Logged volume of exposure to negative ads / .233 (.043)** / .027 (.012)* / .125 (.053)*
Proportion of negative ads to which exposed / .012 (.005)* / .000 (.001) / .006 (.006)
Total spots in market / -.091 (.125) / .218 (.110)* / -.038 (.031) / .007 (.023) / -.133 (.152) / .035 (.135)
Local news viewing / .158 (.036)** / .234 (.036)** / .032 (.008)** / .042 (.007)** / -.011 (.044) / .033 (.038)
Newspaper reading / .026 (.034) / .027 (.033) / .018 (.008)* / .019 (.008)* / -.019 (.034) / -.017 (.034)
Age / -.012 (.005)* / -.013 (.005)* / .004 (.002)* / .004 (.002)* / .008 (.007) / .008 (.007)
African-American / -.051 (.259) / .011 (.250) / .200 (.068)** / .208 (.066)** / -.332 (.300) / -.296 (.288)
Female / -.359 (.162)* / -.252 (.162) / .044 (.032) / .055 (.032)# / -.060 (.180) / -.015 (.175)
Income / .000 (.000) / .000 (.000) / -.000 (.000) / -.000 (.000) / -.000 (.000) / -.000 (.000)
Education / .080 (.061) / .075 (.061) / .047 (.014)** / .046 (.015)** / .242 (.069)** / .242 (.069)**
Political information / -.067 (.040)# / -.063 (.041) / .125 (.012)** / .125 (.012)** / .205 (.072)** / .203 (.073)**
Strength of partisanship / .346 (.254) / .345 (.249) / .283 (.057)** / .285 (.057)** / -.086 (.312) / -.084 (.318)
Mobilized / .664 (.179)** / .653 (.175)** / .151 (.037)** / .155 (.037)** / .351 (.220) / .355 (.220)
House race competitive / -.382 (.375) / -.431 (.355) / .014 (.064) / .004 (.063) / .140 (.304) / .116 (.305)
Senate race competitive / .285 (.209) / .344 (.197)# / .011 (.040) / .023 (.039) / -.429 (.209)* / -.381 (.187)#
Presidential race competitive / .494 (.204)* / .535 (.201)** / -.034 (.036) / -.036 (.036) / .535 (.208)* / .551 (.213)*
Constant / 1.245 (1.231) / 1.374 (.281)** / 1.293 (.287) / .859 (.206)** / 1.493 (1.400) / -.186 (1.198)
N / 1198 / 1198 / 1200 / 1200 / 1180 / 1180
Adjusted R2/ Pseudo R2 / .13 / .11 / .28 / .28 / .08 / .07

**p < .01 *p < .05 #p < .10 (two-tailed tests)Standard errors are clustered by media market (DMA) and state.

Cell entries are OLS (attention to the campaign, external efficacy, perceptions of govt. power) and logit (seeing an ad, closeness of pres. race, voted)

a exposure measures estimated based on all ads to date of interview b exposure measures estimated based on presidential ads to date of interview

c exposure measures based on all ads d exposure measures based on all presidential ads

Table 4: Tests of the Effects of Negative Advertising on Attention to the Campaign, Perceptions of Government, and Voting

Perceptions of government / Turnout
External efficacyc / Perceptions of govt. powerb / Votedc
Logged volume of exposure to negative ads / -.013 (.017) / -.023 (.011)* / .128 (.051)*
Proportion of negative ads to which exposed / -.004 (.002)# / -.003 (.002)* / .020 (.008)*
Total spots in market / .024 (.051) / .025 (.050) / .044 (.032) / .030 (.030) / .045 (.141) / .105 (.130)
Local news viewing / .005 (.015) / .004 (.012) / .007 (.010) / -.000 (.008) / -.069 (.047) / -.032 (.040)
Newspaper reading / .001 (.009) / .001 (.009) / -.001 (.006) / -.001 (.006) / .064 (.031)* / .064 (.031)*
Age / -.004 (.002)* / -.004 (.002)* / .001 (.001) / .001 (.001) / .010 (.005)* / .011 (.005)*
African-American / -.006 (.107) / -.002 (.107) / -.177 (.062)** / -.179 (.062)** / .708 (.303)* / .713 (.300)*
Female / -.077 (.054) / -.075 (.055) / -.053 (.040) / -.057 (.041) / .083 (.154) / .102 (.151)
Income / .000 (.000) / .000 (.000)# / .000 (.000) / .000 (.000) / .000 (.000) / .000 (.000)
Education / .165 (.024)** / .164 (.024)** / -.022 (.011)* / -.021 (.021)# / .349 (.056)** / .348 (.057)**
Political information / .097 (.016)** / .097 (.016)** / .038 (.013)** / .038 (.012)** / .364 (.064)** / .364 (.063)**
Strength of partisanship / .194 (.075)* / .196 (.075)* / -.028 (.053) / -.028 (.053) / .909 (.223)** / .914 (.226)**
Mobilized / .261 (.055)** / .263 (.054)** / .099 (.034)** / .099 (.034)** / 1.282 (.189)** / 1.280 (.188)**
House race competitive / .079 (.080) / .076 (.080) / .046 (.049) / .049 (.052) / -.318 (.210) / -.329 (.197)#
Senate race competitive / -.019 (.074) / -.012 (.073) / -.132 (.048)** / -.126 (.046)** / -.233 (.187) / -.255 (.189)
Presidential race competitive / .017 (.075) / .014 (.074) / .024 (.040) / .021 (.039) / .024 (.210) / .044 (.220)
Constant / 1.600 (.477)** / 1.589 (.447)** / 1.928 (.285)** / 2.069 (.265)** / -3.221 (1.361)* / -3.869 (1.259)**
N / 1193 / 1193 / 1157 / 1157 / 1199 / 1199
Adjusted R2/ Pseudo R2 / .22 / .22 / .06 / .06 / .26 / .26

**p < .01 *p < .05 #p < .10 (two-tailed tests)Standard errors are clustered by media market (DMA) and state.

Cell entries are OLS (attention to the campaign, external efficacy, perceptions of govt. power) and logit (seeing an ad, closeness of pres. race, voted)

a exposure measures estimated based on all ads to date of interview b exposure measures estimated based on presidential ads to date of interview

c exposure measures based on all ads d exposure measures based on all presidential ads

1

Three-Stage Least Squares Models

Table 3: Tests of the Effects of Negative Advertising on Information about the Candidates

Ability to place pres. candidates ideologicallyb / Certainty of candidates’ ideologyd / Bush/Gore likesb / Bush/Gore dislikesb
Logged volume of exposure to negative ads / -.127 (.055)* / -.002 (.005) / .098 (.079) / -.257 (.111)*
Proportion of negative ads to which exposed / .046 (.014)** / .002 (.006) / .009 (.021) / .102 (.029)**
Total spots in market / .044 (.066) / .035 (.032) / -.122 (.104) / .116 (.132)
Newspaper reading / -.002 (.011) / .008 (.007) / .014 (.019) / .006 (.020)
African-American / .063 (.102) / -.030 (.063) / -.390 (.191)* / -.123 (.197)
Female / -.202 (.065)** / -.054 (.041) / .271 (.119)* / -.050 (.125)
Political information / .326 (.024)** / .156 (.057)** / .329 (.035)** / .467 (.048)**
Strength of partisanship / .420 (.114)** / -.209 (.059)** / .795 (.162)** / .085 (.232)
Constant / -.193 (.559) / 1.353 (.302)** / 1.869 (.904)** / -.693 (1.113)
Endogenous variable (1) / Volume of exposure / Volume of exposure / Volume of exposure / Volume of exposure
Local news viewing / .188 (.025)** / .287 (.030)** / .192 (.025)** / .189 (.025)**
Age / -.001 (.004) / .011 (.004)* / .000 (.004) / -.000 (.004)
Income / .000 (.000)* / .000 (.000)# / .000 (.000)* / .000 (.000)*
Education / .002 (.041) / -.054 (.048) / -.010 (.041) / -.004 (.041)
Political information / -.033 (.038) / -.005 (.044) / -.027 (.038) / -.030 (.038)
Strength of partisanship / .236 (.169) / .145 (.202) / .230 (.169) / .229 (.169)
Mobilized / .083 (.132) / -.072 (.158) / .048 (.143) / .068 (.133)
House race competitive / -.204 (.191) / -.344 (.226) / -.263 (.193) / -.247 (.192)
Senate race competitive / -.212 (.120)# / -.279 (.141)* / .205 (.121)# / .192 (.120)
Presidential race competitive / 2.604 (.130)** / 3.091 (.151)** / 2.578 (.130)** / 2.588 (.130)**
Constant / .386 (.262) / .992 (.316) / .406 (.263) / .400 (.262)
Endogenous variable (2) / Proportion of exposure / Proportion of exposure / Proportion of exposure / Proportion of exposure
Local news viewing / .618 (.234)** / 1.586 (.256)** / .751 (.289)** / .612 (.227)**
Age / -.113 (.034)** / -.002 (.039) / -.070 (.042)# / -.068 (.033)*
Income / .000 (.000)* / .000 (.000)* / .000 (.000)* / .000 (.000)*
Education / 1.209 (.354)** / .500 (.417) / .203 (.467) / 1.098 (.339)**
Political information / -.668 (.410) / -.310 (.384) / -.142 (.433) / -.610 (.447)
Strength of partisanship / 2.281 (1.914) / .241 (1.756) / 2.123 (1.924) / 1.829 (1.912)
Mobilized / 1.396 (1.160) / -3.856 (1.368)** / -1.001 (1.519) / 1.203 (1.113)
House race competitive / -2. 331 (1.822) / -6. 818 (1.952)** / -5.079 (2.193)** / -4.321 (1.777)*
Senate race competitive / .039 (1.150) / .540 (1.220) / .204 (1.374) / -1.216 (1.118)
Presidential race competitive / 6.801 (1.399)** / 10.785 (1.308)** / 5.651 (1.481)** / 6.091 (1.389)**
Constant / 7.357 (2.520)** / 15.476 (2.734)** / 10.562 (2.998)** / 7.108 (2.458)**
N / 1048 / 1203 / 1203 / 1203

**p < .01 *p < .05 #p < .10 (two-tailed tests) Standard errors are clustered by media market (DMA) and state

Table 4: Tests of the Effects of Negative Advertising on Attention to the Campaign

Remember seeing an ada / Attention to the campaigna / Perceptions of closeness of
pres. raceb
Logged volume of exposure to negative ads / .106 (.018)** / .100 (.028)** / .031 (.015)*
Proportion of negative ads to which exposed / -.003 (.004) / -.001 (.006) / -.010 (.004)**
Total spots in market / -.042 (.028) / -.100 (.043)* / -.009 (.023)
Newspaper reading / .003 (.005) / .030 (.007)** / .000 (.004)
African-American / -.023 (.040) / .157 (.062)* / -.053 (.033)
Female / -.061 (.026)* / .055 (.040) / .001 (.021)
Political information / .004 (.008) / .154 (.012)** / .028 (.006)**
Strength of partisanship / .047 (.036) / .331 (.055)** / .006 (.030)
Constant / .807 (.226)** / 1.950 (.348)** / .978 (.184)**
Endogenous variable (1) / Volume of exposure / Volume of exposure / Volume of exposure
Local news viewing / .393 (.027)** / .393 (.027)** / .389 (.027)**
Age / -.008 (.004)* / -.005 (.004) / -.008 (.004)#
Income / .000 (.000) / .000 (.000) / .000 (.000)
Education / .020 (.043) / -.003 (.043) / -.045 (.044)
Political information / -.069 (.041)# / -.075 (.041)# / -.055 (.041)
Strength of partisanship / .195 (.181) / .164 (.181) / .193 (.183)
Mobilized / .374 (.140)** / .344 (.141)* / .241 (.144)#
House race competitive / .102 (.202) / .164 (.204) / .154 (.207)
Senate race competitive / 1.716 (.128)** / 1.721 (.129)** / 1.732 (.131)**
Presidential race competitive / .754 (.139)** / .704 (.139)** / .723 (.141)**
Constant / 1.995 (.279)** / 1.840 (.280)** / 2.177 (.285)**
Endogenous variable (2) / Proportion of exposure / Proportion of exposure / Proportion of exposure
Local news viewing / 1.020 (.215)** / 1.018 (.215)** / .916 (.213)**
Age / -.024 (.031) / -.013 (.031) / -.041 (.029)
Income / .000 (.000)* / .000 (.000)# / .000 (.000)*
Education / -.518 (.349) / -.372 (.347) / -.904 (.326)**
Political information / -.287 (.323) / -.373 (.323) / -.008 (.317)
Strength of partisanship / .566 (1.437) / .397 (1.434) / .482 (1.428)
Mobilized / 2.320 (1.133)* / 2.611 (1.126)* / 1.133 (1.056)
House race competitive / -.118 (1.635) / .002 (1.625) / -.666 (1.517)
Senate race competitive / 7.806 (1.025)** / 7.682 (1.021)** / 7.880 (.989)
Presidential race competitive / -3.566 (1.105)** / -3.756 (1.103)** / -3.373 (1.080)**
Constant / 16.554 (2.236)** / 15.587 (2.227)** / 19.739 (2.140)**
N / 1201 / 1203 / 1203

**p < .01 *p < .05 #p < .10 (two-tailed tests) Standard errors are clustered by media market (DMA) and state

Table 4: Tests of the Effects of Negative Advertising on Perceptions of Government, and Voting

External efficacyc / Perceptions of govt. powerb / Votedc
Logged volume of exposure to negative ads / .050 (.092) / .098 (.053)# / -.028 (.037)
Proportion of negative ads to which exposed / -.035 (.030) / -.033 (.017)# / .009 (.012)
Total spots in market / .070 (.048) / .018 (.029) / .012 (.019)
Newspaper reading / .004 (.010) / .002 (.006) / .014 (.004)**
African-American / -.053 (.096) / -.205 (.058)** / .077 (.040)#
Female / -.057 (.059) / -.060 (.036)# / .033 (.025)
Political information / .185 (.019)** / .038 (.011)** / .077 (.007)**
Strength of partisanship / .236 (.088)** / .019 (.053) / .179 (.035)**
Constant / 2.354 (.347)** / 2.481 (.210)** / .211 (.142)
Endogenous variable (1) / Volume of exposure / Volume of exposure / Volume of exposure
Local news viewing / .503 (.023)** / .513 (.023)** / .515 (.023)**
Age / .006 (.003)# / .004 (.003) / .005 (.003)
Income / .000 (.000) / .000 (.000) / .000 (.000)
Education / -.127 (.036)** / -.057 (.038) / -.065 (.037)#
Political information / -.019 (.034) / -.048 (.035) / -.054 (.034)#
Strength of partisanship / .317 (.153)* / .272 (.158)# / .306 (.152)*
Mobilized / .052 (.118) / .099 (.122) / .174 (.120)
House race competitive / .024 (.170) / -.004 (.177) / .040 (.174)
Senate race competitive / .946 (.108)** / .995 (.111)** / .979 (.109)**
Presidential race competitive / .559 (.115)** / .548 (.118)** / .560 (.117)**
Constant / 3.304 (.234)** / 3.118 (.242)** / 3.031 (.238)**
Endogenous variable (2) / Proportion of exposure / Proportion of exposure / Proportion of exposure
Local news viewing / 1.396 (.143)** / 1.511 (.145)** / 1.498 (.143)**
Age / .044 (.020)* / .018 (.020) / .028 (.021)
Income / .000 (.000)* / .000 (.000) / .000 (.000)*
Education / -.910 (.221)** / -.223 (.227) / -.280 (.232)
Political information / .181 (.213) / -.127 (.216) / -.173 (.215)
Strength of partisanship / 1.621 (.960)# / 1.520 (.981) / 1.656 (.955)#
Mobilized / -.242 (.715) / .139 (.735) / 1.126 (.753)
House race competitive / -.557 (1.036) / -.706 (1.073) / -.433 (1.086)
Senate race competitive / 4.772 (.675)** / 5.412 (.687)** / 4.925 (.682)**
Presidential race competitive / 1.112 (.705) / 1.054 (.719) / 1.107 (.734)
Constant / 19.915 (1.439)** / 17.901 (1.478)** / 16.983 (1.487)**
N / 1196 / 1160 / 1202

**p < .01 *p < .05 #p < .10 (two-tailed tests) Standard errors are clustered by media market (DMA) and state

Table 5: The Direct and Indirect Effects of Negative Advertising on Turnout

Logged exposure to negative advertising / .058 (.113)
Proportion of negative ads to which exposed / .002 (.016)
Total spots in market / .164 (.145)
Local news viewing / -.103 (.067)
Newspaper reading / .051 (.040)
Age / .017 (.006)**
African-American / .819 (.357)*
Female / .353 (.176)*
Income / .000 (.000)
Education / .210 (.078)**
Political information / .243 (.085)**
Strength of partisanship / .426 (.324)
Mobilized / 1.222 (.233)**
House race competitive / -.505 (.249)*
Senate race competitive / -.314 (.221)
Presidential race competitive / .049 (.263)
Remember seeing an ad / .179 (.194)
Perceptions of closeness of race / .576 (.348)#
Ability to place pres. candidates ideologically / .186 (.117)
External efficacy / .314 (.118)**
Perceptions of government power / -.019 (.181)
Certainty of candidates’ ideology / .063 (.151)
Attention to the campaign / .581 (.129)**
Bush/Gore likes / -.046 (.064)
Bush/Gore dislikes / .073 (.068)
Constant / -6.319 (1.541)**
N / 992
Pseduo R2 / .30

**p < .01 *p < .05 #p < .10 (two-tailed tests)

Standard errors are clustered by media market (DMA) and state.

1