Bartles (1996) Uninformed Votes: Inofrmation Effects in Presidential Elections

This study rejects the theory that cues are an appropriate replacement of actual political knowledge.
Bartels creates a hypothetical “fully informed” electorate. “How might the preferences of this hypothetical “fully informed” electorate differ from the preferences of the actual electorate? One possibility is to assume that increasing information reduces the variability of voters’ choice without altering the central tendency of their underlying preferences. Alternatively, we might suppose that more informed voters are more likely across the board to prefer Republicans (or Democrats), controlling for other relevant factors. Both of these assumptions seems unduly restrictive, however, in that they require information to have essentially similar effects on all voters regardless of their circumstances. In seems much more plausible to suppose that increasing information—by giving voters a better sense of their credibility and likely consequences of some voters systematically more Republican in their preferences but at the same time make others systematically more Democratic” (205).
H: Uninformed voters successfully use cues and information shortcuts to behave as if they were fully informed. Failing that, individual deviations from fully informed voting cancels out in a mass electorate, producing the same aggregate election outcomes as if voters were fully informed.
Results: hypothesis is disconfirmed; at the individual level, the average deviation of actual vote probabilies from hypothetical “fully informed” vote probabilities was about ten percentage points. In the electorate as a whole, these deviations were significantly deluted by aggregation, but by no means eliminated: incumbent presidents did almost five percentage points btter, and democratic candidates did almost two percentage points better, than they would have if voters had in fact been “fully informed.

“Given the variety of demonstrable differences between well-informed and less well-informed citiens in sensitivity to external stimuli, diversity and precision of political perceptions, information-processing strategies, access to shared understandings of politics, and integrative ability, it hardly seems outlandish to entertain the possibility that disparities in political information lead to systematically different vote choices by citizens in otherwise similar political circumstances, despite—or perhaps even, in part because of—the availability of cues and information short-cuts…” (202).

Methods: generate “fully informed” electorate (204).
Measurement; level of political sophistication:
Interviewer assessed: .95, .8, .5, .2, .05. corresponding to very high…very low.

Findings: “relatively uninformed voters are more likely, other things being equal, to support incumbents and democrats” (218). …suggesting that political ignorance has systematic and significant political consequences (220).

See Lau & Redlawsk 1999 for a really good summary of this article