Jennings & Stoker (2006) "Another and Longer look at the impact of Higher Education on Political Involvement and Attitudes."

"Another and Longer look at the impact of Higher Education on Political Involvement and Attitudes."

-where do the attitudinal and engagement regularities (for the connection between education and involvement) come from? i.e.... what is the actual effect of education?

-they explore four models/explanations.
1.) The Civic Education Explanation: the straight up "education leads to more involvement" explanation. Individuals acquire political skills and knowledge through education, which lowers the opportunity costs of getting involved.

-explains tolerance/social issues: education exposes individuals to diversity, other people, etc. = direct effect

2.) The Cognitive Proficiency Explanation: higher education results in higher levels of information seeking, processing, and organization.

-explains tolerance/social issues: they are more complex, require more complex cognitive functioning

3.) The Social Allocation/Sorting/Credentialing Hypothesis Explanation: higher education leads to access to higher status and status related social networks that involve political influence.

-explains tolerance/social issues: individuals with higher education have access to social networks that are more involved with civil liberties/rights

4.) (seldom mentioned) Pre-Collegiate Socialization Explanation/Argument: individuals are pre-selected by socialization factors... those that are college-bound ALREADY differ from those that aren't. Higher education REINFORCES previous socialization.

-explains tolerance/social issues: pre-adult liberal stances foster the selection for higher education


-apply to political tolerance and social issues (integrated above)


-rule out cognitive explanation: have no measure, and social allocation, have no post post data


-use their longitudinal study/data to assess the first hypothesis, and the fourth as a counter-hypothesis (research design = cross-sectional analysis of that massive longitudinal survey data he collected)

-POLITICAL INVOLVEMENT
"Thus whether we stratify the high school seniors according to eventual educational status or according to contemporary (1965) indicators of educational achievement and aspirations the implication is the same: supposed gains in political involvement as a result of higher education are at the very least partly illusory."

-civic education = weak
-pre-collegiate socialization = strong explanation for these results

-POLITICAL ATTITUDES
-(fewer and less reliable measures)
"Although limited in number, the results based on issue positions lend support to the civic education (and perhaps cognitive sophistication) accounts of how higher education is associated with more liberal stances on civil liberties and civil rights issues."

-THE MATCHING ANALYSIS METHOD
-ends up supporting the notion that education DOES make a difference directly to involvement

-addresses problem of confound of those who had already intended to go to college vs. those who didn't but still ended up going (are they different... is that skewing the data? etc)


-employed the Genetic Matching procedure "which relies upon the genetic optimization algorithm for achieving covariate balance"

-need to identify a treatment and control (they had three groups, so repeated the procedure)

-uh... i re-read this a couple times and still don't think i fully understand it, so am going to ask him to explain it better tomorrow i think. Sorry!


-results showed that "some college" and then more education makes a difference in terms of attitude and involvement, and tolerance.

conclusion about matching method:
"Matching methods such as those that we have employed, applied to panel data where what is prior and what is subsequent to educational achievement can be reliably distinguished, appear to hold the most promise for making progress on this front. But matching methods are vulnerable too. The loss of cases that accompanies the attempt to produce matched groups may yield low statistical power and results that fail to replicate. It also raises questions about external validity that we have set aside here: just what populations are or are not being represented in the effectively matched samples? Perhaps even more importantly, the question of whether the relevant causal variables have been matched upon is always open"

-however, the conclusion about pre-collegiate socialization also has a lot of strengths