October 2008 Poll - Discussion
A Proposed Model
The results from the various Type & Politics Polls indicated psychological type preferences were related to various political orientations and attitudes. Though the results suggested that type plays only a small role in terms of the amount of statistical variation accounted for in these various political orientations, the results were statistically and theoretically significant. Moreover, further research that explores the role of type preferences as exogenous variables within a path-analytic framework may show a more significant effect for type within a larger network of variables. For example, results from our studies showed that type preferences were related to various political orientations and attitudes and that type preferences were related to political party membership. These effects were obtained when the relationship of type to political orientations and party membership were examined separately. When examined at the same time, the role of type was mitigated and the liberal-conservative orientations remained significantly related to party membership. A more realistic perspective may be that type preferences influence the development of various traits and values which then influence political orientations which then influence party identification and voting behavior (as is suggested by various FFM studies).
This kind of "path analytic" model is suggested in the collage presented below. When viewed left to right, the model proposes at least five topics to consider.
First, the model proposes that both a "nature" and "nurture" perspective be employed. Sufficient research to date exists to propose that genetics influence brain development and thus personality. For example, research has shown that variations in gray matter and white matter in the brain are genetically determined. Other research shows that frontal lobe volume is more influenced by heritability than is hippocampal volume. With respect to the heritability of personality, Bourchard's work shows the heritability of type preferences as similarly has been found for the FFM traits. More specifically, research reported in the 1998 Manual suggests relationships between various neurological activity and various type preferences. Conservatism also has been demonstrated to be heritable. Although this research focuses on the heritability perspective of brain structure and functioning as well as personality, rather than the nurture perspective, the research remains open to the influence of environment or nurture. The theory of type development clearly suggests that family and other social situations likely influence at least the expression of type preferences.
Second, the model proposes that type preferences (in interaction with environmental factors) influence the development of certain values and attitudes (including political attitudes). Research in the FFM area clearly has shown a relationship between personality, values, and political orientations, with values more strongly related to political orientation than was personality. Translated into psychological type terms, this research suggests that more conservative voters are likely to be more E, S, T, and J, and to value security, power, achievement, conformity, and tradition. More liberal voters are likely to be more I, N, F, and P, and to value universalism, benevolence, and self-direction. This translation or interpretation of the relationships between the FFM personality traits and values seems supported by recent research1 looking at the relationship between type and values (where values are conceptualized and measured within Schwartz's theory ). Although most types in this recent study tended to rank order Schwartz's values generally in the same order, there were some significant type differences reflecting the FFM results. For example, ESTJs ranked higher in terms of achievement, conformity, tradition, and power than did INTPs. That ESTJs express a strong orientation to these values is consistent with ESTJs identifying as Republican. Moreover, the values are consistent with those espoused by conservatives. That INTPs would rank lower on these values also seems consistent with INTPs identifying as Independents, for whom such values as conformity and tradition are somewhat constraining.
Third, the model proposes that the relationship between type preferences, personality traits, values and political attitudes should be explored both at a general level as well as a more specific level that focuses on political values and attitudes. Research to date suggests small--but theoretically relevant--relationships between type preferences and various values generally, as well as various political orientations. Future research should explore path analytic models which specify type preferences as influencing general values which then are viewed as influencing specific political attitudes. Specific political attitudes may included such attitudes as the social and economic attitudes measured in this study as well as political attitudes that are derived from other theoretical perspectives on personality and politics. For example, in a previous poll we examined the relationship between type preferences and one scale from a measure of Social Dominance Orientation (SDO). Our results showed that those participants with T preferences, on average, scored higher on the scale than did those participants with F preferences. Given that the SDO scale measures attitudes toward equality/inequality in social groups, our results suggested that T types were more likely to be comfortable with inequality whereas F types were less likely to be comfortable with inequality in social groups. These results seem quite consistent with our discussions on the preceding page about the relationships between hierarchy vs. equality and conservative vs. liberal political orientations. (And, in fact, we found in our previous poll that conservatives--as well as the T types--scored higher on the SDO scale than did liberals and the F types). In addition to the SDO scales there are a number of other measures of personality and attitudes that seem relevant for study. These scales include those measuring authoritarianism and Machiavellianism, as well as the personality variables surveyed in studies of conservatism and liberalism. (See here for a list of summary studies and criticisms.)
Fourth, the model proposes that the relationships between political attitudes, political values, political self-concept, party identification and voting behavior be explored within the context of various social and psychological theories that suggest one's political identity and behavior likely are interrelated. Theories of cognitive dissonance, heterophily-homophily , and social identity theory clearly suggest that people seek some consistency in their identity and that this process is influenced by the groups, including political groups, with which they identify. Within the realm of political psychology, this viewpoint is represented in the work on political congruency. To some extent, this aspect of the model also addresses the proposal that the relationship between type preferences and political factors keep in mind both nature and nurture. For example, the notion of social identity theory implies the existence of groups with which individuals might identify, and the notion a target group is one example of the environment and how the environment might influence a person's political identity.
Finally, the model presented here suggests a path analytic approach in terms of methodology. There are at least six reasons to consider such an approach. First, path analytic models such as structural equations models allow researchers to examine multiple relationships between variables in more detail than is done with the usual type table analysis or with basic ANOVA. Second, these methods allow a more detailed look at sources of error in a model. Third, these methods generally provide a better estimate of the "true" relationships among variables than do more traditional methods, particularly when correlations are estimated. (That is, where relationships exist--and the measurement model meets standards--correlations generally are not as attenuated as those with traditional correlational methods.) Fourth, these methods allow a researcher to study relationships between observed variables and latent variables, as well as relationships involving nominal, ordinal, and continuous data. Fifth, these methods are appropriate for studying interactions amongst variables as is done with traditional methods (and which is of interest to type researchers). Finally, there are a variety of statistical programs available to implement these analyses. An overview of this approach (with one statistical package) can be viewed here.
This recommendation that structural equations modeling may be useful, of course, does not rule out other approaches. The cluster analyses approach used in this study provided quite interesting results. Taxometric methods also seem relevant.
1Raymond Moody & Robert Randall. (n.d.) Cultural Values and Type: Explaining Values. Personal Communication.
- << Prev