 |
Campaign and Media Attention to an Issue Causes Learning-Based Effects, Not Priming
| |
| | Unformatted Document Text:
19
matter much to citizens’ voting decisions. Unfortunately, determining the extent to which these two effects lie behind these results presents a more difficult econometric problem because it involves unraveling the direction of causation—never easy with public opinion variables, even with panel data. Nevertheless, I present three imperfect tests of these rival explanations.
One approach to causation with panel data is to use cross-lagged designs. In the British example, if
issue-driven effects lie behind these results, then attitudes about European integration in 1994 should better explain vote choice in 1997 than vote choice in 1994 (“had there been an election”). In contrast, if these results arise from people adopting their party’s position, then vote choice in 1994 should better explain attitudes about European integration in 1997 than in 1994. Using this cross-lagged design, Table 6 presents differences of means for each case among the learners. Increases on the left-hand side reflect issue-driven learning effects and increases on the right-hand side reflect vote-driven learning effects. Across all of these issues, the table suggests that these learners are not changing their vote choice to reflect their issue attitudes, but changing their issue attitudes to match their vote choice. For instance, compared with learners who opposed integration with Europe in 1994, learners who favored it became only slightly more likely to support Labour between 1994 and 1997: they are no more supportive in 1994 and only two-percentage points more supportive in 1997. In contrast, compared to learners who supported the Conservatives in 1994, learners who supported Labour in 1994 became much more favorable towards integration between 1994 and 1997: they are only two percentage points more favorable in 1994, but 26 percentage points more favorable in 1997. Essentially the same results hold in multivariate analyses.
Although I use pre-treatment measures of candidate choice in Table 6, other variables, such as
partisan identification, may be more important in driving opinion change among the learners. In three of the four cases, candidate choice measures work sufficiently well. In the case of Reagan in 1980, however, respondents do not appear to change their attitudes about defense spending from January to September to become substantially more consistent with their feelings for Reagan in January. Although this may reflect a greater stability in attitudes about defense spending, it could also arise because the public knew little about Reagan as a political candidate in January of 1980 and so had poorly formed opinions about him. I therefore use an index of partisan identification and attitudes about Reagan in January 1980, which performs better than either does individually.
Although the evidence in Table 6 supports vote-driven explanations for each case, these findings
could arise because this method may be biased in its favor. Although analyses using lagged instead of current values are common, they generally yield consistent estimates only when synchronous effects are absent (Finkel 1995, 32). With public opinion data, of course, synchronous effects can always exist. In this case, attitudes about integration may be less stable overtime compared to vote choice, and thus prior attitudes may serve as a poor substitute for post attitudes while prior vote choice may serve as a better substitute for post
|
| |
| |
|
|
19
matter much to citizens’ voting decisions. Unfortunately, determining the extent to which these two effects
lie behind these results presents a more difficult econometric problem because it involves unraveling the
direction of causation—never easy with public opinion variables, even with panel data. Nevertheless, I
present three imperfect tests of these rival explanations.
One approach to causation with panel data is to use cross-lagged designs. In the British example, if
issue-driven effects lie behind these results, then attitudes about European integration in 1994 should better
explain vote choice in 1997 than vote choice in 1994 (“had there been an election”). In contrast, if these
results arise from people adopting their party’s position, then vote choice in 1994 should better explain
attitudes about European integration in 1997 than in 1994. Using this cross-lagged design, Table 6 presents
differences of means for each case among the learners. Increases on the left-hand side reflect issue-driven
learning effects and increases on the right-hand side reflect vote-driven learning effects. Across all of these
issues, the table suggests that these learners are not changing their vote choice to reflect their issue attitudes,
but changing their issue attitudes to match their vote choice. For instance, compared with learners who
opposed integration with Europe in 1994, learners who favored it became only slightly more likely to support
Labour between 1994 and 1997: they are no more supportive in 1994 and only two-percentage points more
supportive in 1997. In contrast, compared to learners who supported the Conservatives in 1994, learners who
supported Labour in 1994 became much more favorable towards integration between 1994 and 1997: they
are only two percentage points more favorable in 1994, but 26 percentage points more favorable in 1997.
Essentially the same results hold in multivariate analyses.
Although I use pre-treatment measures of candidate choice in Table 6, other variables, such as
partisan identification, may be more important in driving opinion change among the learners. In three of the
four cases, candidate choice measures work sufficiently well. In the case of Reagan in 1980, however,
respondents do not appear to change their attitudes about defense spending from January to September to
become substantially more consistent with their feelings for Reagan in January. Although this may reflect a
greater stability in attitudes about defense spending, it could also arise because the public knew little about
Reagan as a political candidate in January of 1980 and so had poorly formed opinions about him. I therefore
use an index of partisan identification and attitudes about Reagan in January 1980, which performs better
than either does individually.
Although the evidence in Table 6 supports vote-driven explanations for each case, these findings
could arise because this method may be biased in its favor. Although analyses using lagged instead of current
values are common, they generally yield consistent estimates only when synchronous effects are absent
(Finkel 1995, 32). With public opinion data, of course, synchronous effects can always exist. In this case,
attitudes about integration may be less stable overtime compared to vote choice, and thus prior attitudes may
serve as a poor substitute for post attitudes while prior vote choice may serve as a better substitute for post
|
|
 Convention | | All Academic Convention can solve the abstract management needs for any association's annual meeting. | | Submission - Custom fields, multiple submission types, tracks, audio visual, multiple upload formats, automatic conversion to pdf. | | Review - Peer Review, Bulk reviewer assignment, bulk emails, ranking, z-score statistics, and multiple worksheets! | | Reports - Many standard and custom reports generated while you wait. Print programs with participant indexes, event grids, and more! | | Scheduling - Flexible and convenient grid scheduling within rooms and buildings. Conflict checking and advanced filtering. | | Communication - Bulk email tools to help your administrators send reminders and responses. Use form letters, a message center, and much more! | | Management - Search tools, duplicate people management, editing tools, submission transfers, many tools to manage a variety of conference management headaches! | | Click here for more information. |
|
|
|
| |
|
|
|
