In the two previous diaries I presented some durable patterns in voter turnout, then presented the outlines of a simple theory that accounts for them. Today I turn to some less obvious patterns that have been observed, and sketch how the theory might account for them. Quotation marks indicate that the original text appeared in Achen's paper, for which I provided a link in part one. The rest of the text is mine.
1. "If the costs of voting are small, non-voters will consist of those who do not care very much about the election, or those who care but have little information, or, in smaller numbers, of well-informed, caring citizens who find the candidates equally appealing" (or unappealing).
Such claims have been around for a very long time, and appeared in a scholarly publication in 1944. The important point to note here is that there are plenty of reasons why people might not vote even if the costs of voting drop as close to zero as we can go. Thus, "motor voter" laws, easier registration, shorter lines at polling stations, and weekend voting -- all of which are good ideas and all of which will have some mildly positive effect on turnout -- still do not touch the other barriers to voting.
2. "Those whose turnout rates drop the most from high salience to low salience elections will be disproportionately young, newly enfranchised, or otherwise weak in their partisanship."
This finding points to inexperience and lack of knowledge, as well as weak partisan attachments, to account for why the turnout rate falls as we move from presidential elections, to off-year Federal elections, down to local races for police chief or town clerk. But the declining importance of the race does not affect all voters in the same way. Those whose connection to the electoral political system is the most tenuous are the first to fall away.
3. "Young people who begin voting will vote more often in the future, even though there is no causal relationship between the two events, thus inducing a pseudo habit formation feature to the data which may have mislead scholars" (and others).
Voting is a learning experience as well as an expressive act. One need not appeal to a "habit" of voting to understand why beginning to vote leads to a greater likelihood of voting the next time -- it starts a learning process and also stimulates search for more political information. The difference between "learning" and "habit" as explanations for what is going on with new voters is that the latter term implies a very mechanical process, and also suggests that people do not learn about politics by engaging in it. Other observations suggest that that claim is simply incorrect.
4. "Contact by a trusted source will have a greater effect on the less engaged and less informed. Thus there is no [single] number that is the effect of contacting voters [in a certain way]; everything depends on the population contacted."
Countless arguments about campaign tactics are addressed by this finding. For example, those who are sold on the merits of neighbor-to-neighbor efforts may be so enthusiastic because they tried it in a neighborhood where turnout was low and it seemed to produce dramatic results. They confront skeptics who tried the same approach in a neighborhood of older, well educated, and experienced voters and found that it made little difference in turnout. The two should not be arguing, for they are both correct -- what works well in one kind of neighborhood may be a flop when tried in another.
5. "Party identication (or age) has greater effects on the turnout of poorly informed (less educated) citizens."
Again, the practical implications of this point is that the effects of a given campaign tactic depend critically on the nature of the audience for that tactic. Strong partisan ties in effect substitute for "becoming educated on the issues." Strong partisans don't need to worry about issues; they have a clear conception of which party serves their interests, and the simple cue of party affiliation is enough for them to vote for or against any candidate. That is one reason why business-oriented groups in urban America of the late nineteenth century began pushing the idea of "non-partisan" elections -- it was a way of depriving working class voters of a simple, economical, and -- most importantly -- a valid signal of which candidate to vote for.
6. "Information (or education) has a greater effect on the voting rates of less partisan (or younger) citizens."
The logic here is similar to earlier findings about how the aditional of a single factor has the most effects on people who experience a relative deficit on other factors that lead to voting.
Overall, these findings suggest that there is no one best campaign style. They also suggest that young voters are far more important than most campaigns appreciate. Efforts to reach young voters are an investment whose payoffs will accrue over a long period. "Practical" campaign organizations (or consultants) who are oriented simply to winning the current election, and who consider the value of future won election victories very little if at all, will always argue that the young don't turn out, so resources that could be used on them could be better devoted to older, more informed, and better educated voters. Such a claim is not incorrect, but it neglects the simple fact that how one wages the current campaign affects what is possible in the next one. Those of us who are in it for the long haul often have an intuitive dislike for strategies that dump all the resources into reaching the not-young, the not-poor, the not-poorly educated. Achen's theory and empirical work suggests that our intuitions are well founded. If you have a long time horizon, you fight elections differently, you organize a party differently, and you reach voters differently, than if you do not.