Skip to main content

The Role of Partisanship in Aggregate Opinion

Abstract

Despite the centrality of party identification in U.S. politics, the effects of partisanship on public opinion remain elusive. In this article, we use monthly economic opinion data disaggregated by partisanship to evaluate the role of party identification on economic perceptions. Using both static and time-varying error correction models, we find strong evidence of partisan bias in the public’s assessment of the state of the economy, and importantly, this bias changes over time. This evidence of the changing influence of partisanship helps reconcile some of the different findings of individual and aggregate level opinion studies. We also examine how the time-varying influence of partisanship affects aggregate public opinion. Specifically, we show that the increased influence of partisanship has led aggregate economic perceptions to respond more slowly to objective economic information.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5

Notes

  1. 1.

    Even in his counterargument to claims of rising polarization, Fiorina (2006) concedes that on political “evaluations,” like presidential job approval, polarization has increased.

  2. 2.

    As we note below, not all time series opinion studies ignore partisanship. See, for example, Page and Shapiro (1992), Erikson et al. (2002), and Gerber and Huber (2010).

  3. 3.

    Several other notable theories of partisanship (e.g., Achen 1992; Downs 1957; Fiorina 1981) propose that causality runs the other direction, where party identification is a function of voters perceptions of past (or expected) party behavior. Thus, voters adopt and solidify their partisanship based on the perceived benefit of the party’s actions. These models imply that partisanship will become stronger as more information enters what Fiorina (1981) calls voters’ “running tally” of past political evaluations. Whether or not this strengthening partisanship influences individuals’ reception of and responses to new information is an open question within this literature.

  4. 4.

    In contrast to Bartels, Bullock (2009) recently showed that Bayesian learning need not produce opinion or belief convergence. However, given our interest is in whether partisans update their economic evaluations differently during different periods, we are comfortable with the claim that in a relative sense, parallel opinion change signals more similarities across partisan groups than divergent change.

  5. 5.

    This headlines comes from the New York Times, 23 Sep 1998.

  6. 6.

    It is important to remember that Bayesian updating does not necessarily produce convergence (Bullock 2009), but as previously mentioned we believe it is safe to view periods of convergence as evidence of less partisan bias than periods of divergence or even parallelism. This is especially likely for opinion updating in response to economic information, which is readily available and consistently relevant for decisions and behavior.

  7. 7.

    The Gallup question asks, “How would you rate economic conditions in this country today–as excellent, good, only fair, or poor?” CBS asks, “How would you rate the condition of the national economy these days? Is it very good, fairly good, fairly bad, or very bad?” The ABC surveys ask, “Do you think the nation’s economy is getting better, worse, or staying about the same?” For the Gallup series, the percent rating the economy as excellent or good was tabulated, and for the CBS series, the percent rating the economy as very good or fairly good was calculated. For ABC, the percent saying the economy was getting better is tabulated. In the questions used to identify partisanship, respondents were asked whether “you usually consider yourself a Republican, Democrat or Independent.”

  8. 8.

    We combine the series using a technique for merging time series described in Shumway and Stoffer (2006). This procedure is similar to Stimson’s (1999) Dyad Ratios Algorithm but provides a smoother estimate of the underlying opinion for the early years of the analysis. This was necessary since there were fewer polls to disaggregate for the subgroup analysis during the 1980s than in subsequent years. The smoother estimate provided by the Shumway and Stoffer (2006) approach guards against sampling error affecting the time-varying analysis that we employ. Both algorithms scale the series to a common metric and then use a factor analytic approach to extract the common variance of the series.

  9. 9.

    The analysis begins in 1985 because monthly economic opinion data are not consistently available prior to this year.

  10. 10.

    Dickey–Fuller tests indicate that we cannot reject the null hypothesis of a unit root in the partisan series and provide further evidence that the ECM is the appropriate technique for analyzing these data. Thus, our approach is similar to De Boef and Kellstedt (2004) who find that the consumer sentiment time series is not stationary, leading them to employ ECM for their analysis. Other techniques for dealing with non-stationary data, like ARIMA models, make it difficult to isolate long-term, short-term, and equilibrium relationships. Vector Error Correction Models provide another strategy for dealing with problems of stationarity, but their utility is primarily in modeling the equilibrium relationship, not the short-term and long-term effects.

  11. 11.

    Note that in the tables that follow, the lagged independent variables are used to calculate the long-run multiplier (LRM). Both the lagged dependent variable and the LRM tell us about the long-run effect, but since the LRM provides more information, we only report the LRM.

  12. 12.

    As noted by De Boef and Keele (2008) and many standard econometric texts, there are a variety of transformations that can be made to an Autoregressive Distributed Lag (ADL) to produce an error correction model. De Boef and Keele use the form consistent with one-step estimation. The form above makes it easier to see how the effects of independent variables are filtered through the error correction factor and this is even more important for the time-varying ECM described below.

  13. 13.

    As De Boef and Keele (2008) note, because the LRM is calculated from the formula for the ratio of the lagged independent variables and the speed of adjustment parameter, (using the notation from Eq. 1, \(({\gamma}/({\alpha-1}))\)), the standard error is calculated using the formula for the variance of a ratio. Because the LRM is the byproduct of a ratio, it is possible for the LRM to be significant even when the coefficient for the X t−1’s are not. This can occur when (α − 1) is small or the covariance between the coefficients is high. See De Boef and Keele (2008, p. 192).

  14. 14.

    For a more general application of time-varying analysis of aggregate opinion, see McAvoy (2006).

  15. 15.

    Although our theoretical priors focus on the rate of responsiveness, it is possible to allow the γs to vary with time as well. This suggests that partisanship also changes the weight that the public attaches to unemployment, not just the speed with which it returns to equilibrium. For the analysis of economic evaluations, the γs were not time-varying and as a result we leave the discussion of a full time-varying model for future research and a different context. In their estimates of time-varying ECMs, Barassi et al. (2005) also just allow α t  − 1 to vary with time.

  16. 16.

    The relationship between economic evaluations and presidential approval is well established (see, e.g., Sigelman and Knight 1985). Even during periods of divided government, the public holds the president responsible for the economy (Norpoth 2001). See Wlezien et al. (1997) for an application to vote choice.

  17. 17.

    Although we differ from Lebo and Cassino in that we expect the influence of partisanship to vary over time, when partisanship does influence opinion updating, we expect our results to corroborate their findings.

  18. 18.

    Our analysis does not disaggregate beyond partisan groups, because party is the concept of theoretical interest. This decision is also consistent with past research, which shows that partisan differences in overtime evaluations of the president typically trump differences across educational groups (McAvoy and Enns 2010, see also, Enns and Kellstedt 2008).

  19. 19.

    The upper bound for the vertical axis in the figure is set at 5 to emphasize the systematic patterns across administrations as opposed to highlighting Republican’s atypical pattern (i.e., increasing positive economic ratings when unemployment rises and decreasing positive ratings when it falls) of updating under Bush.

  20. 20.

    As noted earlier, these results support recent research by Lebo and Cassino (2007) on presidential approval. In the next section, however, our analysis of how these effects influence aggregate opinion shows how the influence of out-party updating varies over time and that Independents tend to move in concert with the out party—not the in party.

  21. 21.

    As noted above, the error correction rate ranges from −1 to 0 and estimates near 0 indicate a slow return to equilibrium and ones near −1 indicate an immediate return to equilibrium.

  22. 22.

    An alternative approach to studying the heterogeneity in partisan is to use models of fractional integration (Box-Steffensmeier and Smith 1996; Lebo et al. 2000). However, tests of these the partisan subgroup data and the aggregate series suggest that they do not exhibit the long-memoried characteristics of fractionally integrated series. This maybe the result of merging several time series together, which is likely to diminish any long-memoried features of the series. As noted earlier, the partisan economic evaluation series do have a unit root and, thus, are consistent with models of “permanent” memory employed here.

  23. 23.

    The time-varying parameter model is estimated using the Kalman filter and maximum likelihood estimation of the hyper-parameters, as described in Beck (1983, 1990). Estimation was done using the R package, “dlm.” (Petris and Petrone 2009). The hyper-parameters to estimate are the error variances of the time-varying coefficients, and in this setup the key parameter is the error variance of α t  − 1, which is \(\sigma _{{\alpha {}_{t} - 1}}^{2}\). We can conduct a hypothesis test to see if the error variance is significantly different from zero. If the error variance is not significantly different from zero, the parameter reverts to a constant effect coefficient as in OLS. In this case, rejecting the null hypothesis means that the error correction varies over time, and failing to reject the hypothesis means it is constant. The time-varying model was estimated allowing short and long-term effects to vary overtime, but only the error correction rate and the constant were significant over time. The time-varying error correction rate is consistent with the model described in Eq. 2.

  24. 24.

    When the null hypothesis cannot be rejected, this means that there is no error variance α t  − 1 and that the error correction rate is fixed and can be estimated with a standard OLS approach. But, in this case, since the null hypothesis can be rejected, there is time-variation in the error correction rates.

  25. 25.

    It is important to note, however, that the error correction rate itself has a significant impact on economic evaluations, but that impact does not change significantly over time.

  26. 26.

    Gerber and Green (1999) provide an important defense of this practice.

  27. 27.

    The time series plotted in Fig. 1 suggest parallelism during the Clinton administration, but the ECM provides a more systematic analysis and shows parallel updating among the parties at the beginning of the Clinton administration and divergent updating after 1995.

  28. 28.

    Such a finding would be consistent with the work of De Boef and Kellstedt (2004) in which they find that political judgments and events play a role in consumer sentiment.

References

  1. Achen, C. (1992). Social psychology, demographic variables, and linear regression: Breaking the iron triangle in voting research. Political Behavior, 14, 195–211.

    Article  Google Scholar 

  2. Barassi, M. R., Caporale, G. M., & Hall, S. G. (2005). Interest rate linkages: A Kalman filter approach to detecting structural change. Economic Modelling, 22, 253–284.

    Article  Google Scholar 

  3. Bartels, L. (2002). Beyond the running tally: Partisan bias in political perceptions. Political Behavior, 24, 117–150.

    Article  Google Scholar 

  4. Beck, N. (1983). Time-varying regression models. American Journal of Political Science, 27(3), 557–600.

    Article  Google Scholar 

  5. Beck, N. (1990). Estimating dynamic models using Kalman filtering. Political Analysis, 1, 121–156.

    Article  Google Scholar 

  6. Berelson, B. R., Lazarsfeld, P. F., & McPhee, W. N. (1954). Voting: A study of opinion formation in a presidential campaign. Chicago: University of Chicago Press.

    Google Scholar 

  7. Binder, S. A. (1999). The dynamics of legislative gridlock, 1947–96. American Political Science Review, 93(3), 519–534.

    Article  Google Scholar 

  8. Binder, S. A. (2003). Stalemate: Causes and consequences of legislative gridlock. Washington, DC: Brookings Institution Press.

    Google Scholar 

  9. Box-Steffensmeier, J. M., & Smith, R. M. (1996). The dynamics of aggregate partisanship. American Political Science Review, 90, 567–580.

    Article  Google Scholar 

  10. Bullock, J. G. (2009). Partisan bias and the Bayesian ideal in the study of public opinion. Journal of Politics, 71(3), 1109–1124.

    Article  Google Scholar 

  11. Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. New York: Wiley.

    Google Scholar 

  12. Carroll, C. D. (2003). Macroeconomic expectations of households and professional forecasters. Quarterly Journal Of Economics, 118(1), 269–298.

    Article  Google Scholar 

  13. De Boef, S., & Keele, L. (2008). Taking time seriously. American Journal of Political Science, 52(1), 184–200.

    Article  Google Scholar 

  14. De Boef, S., & Kellstedt, P. M. (2004). The political (and economic) origins of consumer confidence. American Journal of Political Science, 48, 633–649.

    Article  Google Scholar 

  15. Downs, A. (1957). An economic theory of democracy. New York: Harper and Row.

    Google Scholar 

  16. Enns, P. K., & Kellstedt, P. M. (2008). Policy mood and political sophistication: Why everybody moves mood. British Journal Of Political Science, 38(Part 3), 433–454.

    Google Scholar 

  17. Eppright, D. R., Arguea, N. M., & Huth, W. L. (1998). Aggregate consumer expectation indexes as indicators of future consumer expenditures. Journal of Economic Psychology, 19, 215–235.

    Article  Google Scholar 

  18. Erikson, R. S., MacKuen, M. B., & Stimson, J. A. (2002). The macro polity. New York: Cambridge University Press.

    Google Scholar 

  19. Finkel, S. E. (1993). Reexamining the “minimal effects” model in recent presidential campaigns. Journal of Politics, 55, 1–21.

    Article  Google Scholar 

  20. Fiorina, M. (2006). Culture war?: The myth of a polarized America. New York: Pearson Longman.

    Google Scholar 

  21. Fiorina, M. P. (1981). Retrospective voting in American national elections. New Haven, CT: Yale University Press.

    Google Scholar 

  22. Gaines, B. J., Kuklinski, J. H., Quirk, P. J., Peyton, B., & Verkuilen, J. (2007). Same facts, different interpretations: Partisan motivation and opinion on Iraq. Journal of Politics, 69, 957–974.

    Article  Google Scholar 

  23. Gelman, A., Park, D., Shor, B., & Bafumi, J. (2008). Red state, blue state, rich state, poor state: Why Americans vote the way they do. Princeton, NJ: Princeton University Press.

    Google Scholar 

  24. Gerber, A., & Green, D. (1999). Misperceptions about perceptual bias. Annual Review of Political Science, 2, 189–210.

    Article  Google Scholar 

  25. Gerber, A. S., & Huber, G. A. (2010). Partisanship, political control, and economic assessments. American Journal of Political Science, 54(1), 153–173.

    Article  Google Scholar 

  26. Goren, P. (2002). Character weakness, partisan bias, and presidential evaluation. American Journal of Political Science, 46, 627–641.

    Article  Google Scholar 

  27. Goren, P. (2007). Character weakness, partisan bias, and presidential evaluation: Modifications and extensions. Political Behavior, 29, 305–325.

    Article  Google Scholar 

  28. Green, D., Palmquist, B., & Schickler, E. (2002). Partisan hearts and minds: Political parties and the social identities of voters. New Haven, CT.: Yale University Press.

    Google Scholar 

  29. Hibbs, D. A., Rivers, D., & Vasilatos, N. (1982). The dynamics of political support for American presidents among occupational and partisan groups. American Journal of Political Science, 26, 312–323.

    Article  Google Scholar 

  30. Huth, W. L., Eppright, D. R., & Taube, P. M. (1994). The indexes of consumer sentiment and confidence: Leading or misleading guides to future buyer behavior. Journal of Business Research, 29, 199–206.

    Article  Google Scholar 

  31. Iyengar, S., & Hahn, K. S. (2009). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59(1), 19–39.

    Article  Google Scholar 

  32. Jacobson, G. C. (2006). Divider, not a uniter: George W. Bush and the American people. New York: Pearson Longman.

    Google Scholar 

  33. Jacobson, G. C. (2009). The president, the war, and voting behavior in the 2006 House elections. In J. J. Mondak, & D.-G. Mitchell (Eds.), Fault lines, Chap. 7 (pp. 128–147). New York: Routledge.

    Google Scholar 

  34. Jacoby, W. G. (1988). The impact of party identification on issue attitudes. American Journal of Political Science, 32, 643–661.

    Article  Google Scholar 

  35. Juster, F. T., & Wachtel, P. (1972). Inflation and the consumer. Brookings Papers on Economic Activities, 1, 71–121.

    Article  Google Scholar 

  36. Kruglanski, A. W., & Webster D. M. (1996). Motivated closing of the mind: ”Seizing” and ”freezing”. Psychological Review, 103(2), 263–283.

    Article  Google Scholar 

  37. Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480–498.

    Article  Google Scholar 

  38. Lebo, M. J., & Cassino, D. (2007). The aggregated consequences of motivated reasoning and the dynamics of partisan presidential approval. Political Psychology, 28(6), 719–746.

    Article  Google Scholar 

  39. Lebo, M. J., Walker, R. W., & Clarke, H. D. (2000). You must remember this: Dealing with long memory in political analysis. Electoral Studies, 19, 31–48.

    Article  Google Scholar 

  40. Levendusky, M. (2009). The partisan sort. Chicago: University of Chicago Press.

    Google Scholar 

  41. MacKuen, M. B., Erikson, R. S., & Stimson, J. A. (1992). Peasants or bankers: The American electorate and the U.S. economy. American Political Science Review, 86, 597–611.

    Article  Google Scholar 

  42. Mankiw, N. G., & Reis, R. (2002). Sticky information versus sticky prices: A proposal to replace the New Keynesian Phillips curve. Quarterly Journal Of Economics, 117(4), 1295–1328.

    Article  Google Scholar 

  43. Mankiw, N. G., & Reis, R. (2006). Pervasive stickiness. American Economic Review, 96(2), 164–169.

    Article  Google Scholar 

  44. Markus, G. B., & Converse, P. E. (1979). A dynamic simultaneous equation model of electoral choice. American Political Science Review, 73, 1055–1070.

    Article  Google Scholar 

  45. McAvoy, G. E. (2006). Stability and change: The time varying impact of economic and foreign policy evaluations on presidential approval. Political Research Quarterly, 59, 71–83.

    Article  Google Scholar 

  46. McAvoy, G. E., & Enns, P. K. (2010). Using approval of the president’s handling of the economy to understand who polarizes and why. Presidential Studies Quarterly, 40(3), 545-558.

    Article  Google Scholar 

  47. McCarty, N., Poole, K. T., & Rosenthal, H. (2006). Polarized America: The dance of ideology and unequal riches. Boston: MIT Press.

    Google Scholar 

  48. Mueller, E. (1966). The impact of unemployment on consumer confidence. Public Opinion Quarterly, 30, 19–32.

    Article  Google Scholar 

  49. Mueller, J. (1973). War, presidents and public opinion. New York: Wiley.

    Google Scholar 

  50. Norpoth, H. (2001). Divided government and economic voting. Journal of Politics, 63(2), 414–435.

    Google Scholar 

  51. Page, B. I., & Shapiro, R. Y. (1992). The rational public: Fifty years of trends in Americans’ policy preferences. Chicago: University of Chicago Press.

    Google Scholar 

  52. Petris, G., & Petrone, S. (2009). Dynamic linear models With R. New York: Springer.

    Google Scholar 

  53. Prior, M. (2007). Post-broadcast democracy. Cambridge: Cambridge University Press.

    Google Scholar 

  54. Shumway, R. H., & Stoffer, D. S. (2006). Time series analysis and its applications: With R examples. New York: Springer.

    Google Scholar 

  55. Sigelman, L., & Knight, K. (1985). Public opinion and presidential responsibility for the economy: Understanding personalization. Political Behavior, 7, 167–191.

    Article  Google Scholar 

  56. Stimson, James A. (1999). Public opinion in America: Moods, cycles, and swings (2nd ed.). Boulder, CO: Westview Press.

    Google Scholar 

  57. Stimson, J. A., MacKuen, M. B., & Erikson, R. S. (1995). Dynamic representation. American Political Science Review, 89, 543–565.

    Article  Google Scholar 

  58. Stokes, D. E. (1966). Some dynamic elements of contests for the presidency. The American Political Science Review, 60, 19–28.

    Article  Google Scholar 

  59. Taber, C. S., Cann, D., & Kucsova, S. (2009). The motivated processing of political arguments. Political Behavior, 31(2), 137–155.

    Article  Google Scholar 

  60. Taber, C. S., & Lodge, M. (2006). Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50(3), 755–769.

    Article  Google Scholar 

  61. Wlezien, C. (1995). The public as thermostat: Dynamics of preferences for spending. American Journal of Political Science, 39, 981–1000.

    Article  Google Scholar 

  62. Wlezien, C., Franklin, M., & Twiggs, D. (1997). Economic perceptions and vote choice: Disentangling the endogeneity. Political Behavior, 999, 7–17.

    Article  Google Scholar 

  63. Zaller, J. R. (1992). The nature and origins of mass opinion. New York: Cambridge University Press.

    Book  Google Scholar 

  64. Zellner, A. (1962). An efficient method of estimating seemingly unrelated regressions and tests for aggregation bias. Journal of the American Statistical Association, 57, 348–368.

    Article  Google Scholar 

Download references

Acknowledgments

We would like to thank Gary Jacobson for suggesting the CBS and Gallup questions about the public’s evaluation of the economy and for providing some of the data used to construct these series. We would also like to thank Paul Goren for helpful comments on an earlier version.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Gregory E. McAvoy.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Enns, P.K., McAvoy, G.E. The Role of Partisanship in Aggregate Opinion. Polit Behav 34, 627–651 (2012). https://doi.org/10.1007/s11109-011-9176-7

Download citation

Keywords

  • Party identification
  • Partisan bias
  • Time series analysis
  • Public opinion
  • Economic perceptions