Hill, Lo, Vavreck and Zaller (2013) showed that the persuasive impact of televised campaign advertising decays rapidly, but that all duration estimates hinge on the parametric form assumed for the decay curve: competing standard forms fitted to the same data imply advertising half-lives ranging from about 2 days to about 11 days and 80%-extinction times from 3 to 28 days. They stated that no study had ever systematically compared decay models for persuasion effects, and they left open which functional form is correct and what cognitive mechanism generates the decay. The problem — a coherent characterization of the temporal decay profile of mass-communication persuasion and its mechanisms — remains open.

Background

Political mass communication—campaign advertising in particular—is believed to move voters' preferences, and a long empirical tradition asks not only how large such persuasion effects are but how long they last. A precursor randomized field experiment (Gerber, Gimpel, Green and Shaw 2011, on televised advertising in a 2006 gubernatorial race) measured both size and duration and found short-lived effects. The pivotal study is Hill, Lo, Vavreck and Zaller's "How Quickly We Forget" (2013), which connected two literatures that had never been joined. The first is the psychology of forgetting: Rubin and Wenzel (1996) fitted over a century of memory-retention experiments and identified a small set of best-performing parametric decay functions—logarithmic, power, hyperbolic-power, and exponential-power (a form political scientists know as a Weibull survival function, i.e. a curve giving the probability that an effect survives at least a given number of days)—to which Hill et al. added the widely used simple exponential. The second is dual-process theory of political cognition: under "memory-based" processing, a citizen evaluates candidates at the moment of judgment using whatever happens to be recalled from short-term memory, so effects last only as long as the memory trace; under "online" processing, the citizen updates a running evaluation at the moment of exposure and stores it, so effects are durable. Hill et al. analyzed daily-advertising exposure (ad viewings from media-market tracking data) and rolling cross-section surveys (daily national samples with fresh randomization each day) plus county-level vote returns for the 2000 U.S. presidential election, and survey data on 2006 subnational (House, Senate, gubernatorial) races. Each day's net advertising was weighted by a decay function, summed, and passed through an ordered probit model (a regression model for ordered categorical outcomes, here the strength of vote preference) of vote preference. Their findings: the bulk of persuasive impact decays within days; in the presidential race roughly 5–10% of impact from ads aired 42 days earlier survives; subnational effects decay faster and more completely. But the estimates were dramatically form-dependent—an exponential form implied an 11-day half-life while the power form implied 2 days—and the authors explicitly noted that "no one, to our knowledge, has made a comparable study of models for studying the decay of persuasion effects," that the dual-process account was a plausible interpretation rather than a tested mechanism requiring additional field research to confirm, and that weak statistical power left open the possibility of durable subnational effects, while arguing that their model and results are broadly applicable within political communication. Half-life here means the time for half of an exposure's initial persuasive impact to disappear.

Problem Statement

Identify the law of temporal decay for persuasion effects of mass political communication—the open problem posed by Hill, Lo, Vavreck and Zaller (2013), who asked "So which function to use?" after showing that every duration estimate depends dramatically on the assumed functional form. The open target comprises two questions recognized by the source: (1) Decay-function identification: determine the functional form of the decay of persuasive impact with time since exposure (discriminating among the logarithmic, power, exponential, Weibull/exponential-power, and hyperbolic-power families, or establishing a better-fitting alternative), so that half-life and long-lag survival estimates no longer depend on an arbitrary choice of form. (2) Mechanism: determine whether the decay profile is generated by the mix of memory-based and online (effortful) processing the dual-process account hypothesizes, with individual processing mode linked to measured persistence of persuasion. Hill et al. also found decay to be faster and more complete in subnational than in presidential races and argued that their results are broadly applicable within political communication. Cross-setting variation in decay parameters — such as the one-sided message environments in which Jäger (2020) documented durable effects, or changing media modalities — is optional context only: it is not a required component of the target, and only (1) and (2) must be resolved.

References

  1. Seth J. Hill, James Lo, Lynn Vavreck and John Zaller (2013), "How Quickly We Forget: The Duration of Persuasion Effects From Mass Communication", Political Communication 30(4):521–547. DOI: 10.1080/10584609.2013.828143, https://doi.org/10.1080/10584609.2013.828143 (source of the open question)
  2. David C. Rubin and Amy E. Wenzel (1996), "One hundred years of forgetting: A quantitative description of retention", Psychological Review 103(4):734–760. DOI: 10.1037/0033-295X.103.4.734, https://doi.org/10.1037/0033-295X.103.4.734 (source of the decay-function taxonomy used by Hill et al.)
  3. Alan S. Gerber, James G. Gimpel, Donald P. Green and Daron R. Shaw (2011), "How Large and Long-lasting Are the Persuasive Effects of Televised Campaign Ads? Results from a Randomized Field Experiment", American Political Science Review 105(1):135–150. DOI: 10.1017/S000305541000047X, https://doi.org/10.1017/S000305541000047X (randomized-field-experiment precursor on size and duration)
  4. Larry M. Bartels (2014), "Remembering to Forget: A Note on the Duration of Campaign Advertising Effects", Political Communication 31(4):532–544. DOI: 10.1080/10584609.2014.956198, https://doi.org/10.1080/10584609.2014.956198 (published note directly engaging Hill et al.)
  5. Joshua L. Kalla and David E. Broockman (2018), "The Minimal Persuasive Effects of Campaign Contact in General Elections: Evidence from 49 Field Experiments", American Political Science Review 112(1):148–166. DOI: 10.1017/S0003055417000363, https://doi.org/10.1017/S0003055417000363
  6. Kai Jäger (2020; first published online November 2019), "When Do Campaign Effects Persist for Years? Evidence from a Natural Experiment", American Journal of Political Science 64(4):836–851. DOI: 10.1111/ajps.12488, https://doi.org/10.1111/ajps.12488
  7. John Sides, Lynn Vavreck and Christopher Warshaw (2022; first published online November 2021), "The Effect of Television Advertising in United States Elections", American Political Science Review 116(2):702–718. DOI: 10.1017/S000305542100112X, https://doi.org/10.1017/S000305542100112X
  8. Matthew Baum, Adam J. Berinsky, Justin de Benedictis-Kessner, Joohye Jeong, Zachary Markovich and Teppei Yamamoto (2026), "Dynamic persuasion: decay and accumulation of partisan media persuasion", Political Science Research and Methods, pp. 1–17. DOI: 10.1017/psrm.2026.10095, https://doi.org/10.1017/psrm.2026.10095