October 2008 Poll - Discussion
Political Orientations: Self-Perceptions and Attitudes Toward Political Issues
Previous versions of the Type & Politics Poll have examined liberal-conservative political orientations using the one-item question adapted from the ANES surveys . A primary reason for using this item has been that the item has a long and productive history in research on political orientations. Moreover, most people tend to consider political orientations along a liberal-conservative continuum (although, ironically, popular discourse often takes a typological approach as evidenced by political pundits who refer to politicians as "liberals," "conservatives," and "moderates".)
More recently, political psychologists and others have begun to frame their research in terms of two continuums : economic and social (although there are variations on this framework). From this perspective, people may be seen, e.g., as socially liberal yet economically conservative.
A primary purpose of the final Type & Politics Poll was to investigate the relationships between psychological type and these more refined political orientations. Further, the final poll sought to study the relationships between type and political orientations in terms of self-perceptions of of one's own political orientation as well as one's attitudes toward political issues.
Participants in the poll indicated their perceptions of themselves as liberal or conservative using three one-item questions. The first item was the item adapted from the ANES surveys and provided information about participants' perceptions of their liberal-conservative orientations in general. The next two items provided information about participants' perceptions of their liberal-conservative orientations economically and socially.
Participants also provided responses to 20 questions indicating their attitudes toward specific political issues. The 20 questions were selected to represent attitudes toward both economic and social issues. A key question, of course, concerns the extent to which these questions provided meaningful indicators of social and economic political attitudes.
The chart below shows results from a statistical (factor) analysis of responses to the 20 attitude questions as well as responses to the three political self-perception items. Given the assumption that there is an economic dimension as well as a social dimension to political orientations, the analysis was based on a two-factor model (i.e., two factors were extracted from the data. See results here .).
As can be seen in the chart, the two-dimensional assumption seems valid in that questions designed to indicate economic attitudes cluster together (the "blue" clusters) and questions designed to indicate social attitudes cluster together (the "yellow" clusters). Further, the clusters are closer to their respective axes and are located at extreme ends (and not in the center at the intersection of the axes) thus demonstrating the questions are "good" indicators of a particular set of attitudes.
The horizontal axis (Factor 2) indicates items measuring attitudes toward economic issues and the vertical axis (Factor 1) indicates items measuring attitudes toward social issues. For example, item po15 (where "po" simply stands for "political opinion") reads "The government should intervene to reduce income differences" and item po5 reads "The government should see to it that every person has a job and a good standard of living." These items clearly indicate attitudes toward the role of government intervention in economic matters to ensure income and opportunity equality. On the other hand, item po19 reads "Government should play a stronger role in regulating personal morality" and item po17 reads " We should be more tolerant of people who choose to live according to their own moral standards, even if they are very different from our own." These items clearly indicate attitudes toward the role of government intervention in social matters as well as general social attitudes towards morality (and thus politics). This interpretation seems further supported by the "plot" or location of participants' responses to the self-perception items. That is, the items measuring participants' self-perceptions of themselves as liberal or conservative economically (libcon_econ) and socially (libcon_social) are closer to their respective axes.
Based on these considerations, the "best" items were selected to form summated scales to measure respondents' political attitudes toward social and economic issues. Further statistical (reliability) analysis showed both the economic scale (results here ) and the social scale (results here ) have good reliability. The wording of each item can be found here .
A look at the correlations between the 3 self-perception items and the 2 attitude scales shows some interesting results. First, as might be expected, the correlations between the 3 self-perception items tend to be fairly strong--self-perceptions of one's self as economically liberal-conservative and self-perceptions of one's self as socially liberal-conservative correlate strongly with the general self-perception item (r = . 77 and r = .85, respectively). The correlation between the economic and social self-perception items also is strong (r = .53), although somewhat lower than with the general self-perception item. This lower correlation thus provides some support for the assumption that two related dimensions are involved in one's perceived political orientation. Second, the correlations amongst the summated attitude scales and the self-perception items similarly are strong. Again, however, the correlations are not so strong as to suggest these scores are measuring the same constructs.
The "blue" line drawn diagonally in the chart provides further insight into participants' responses. Note that this line passes through both the intersection of the two axes as well as through the plot of the general liberal-conservative self-perception item (labeled libcon_research on the chart). Items that are plotted close to this line also should indicate attitudes toward political issues that might be considered along another liberal-conservative dimension (other than social or economic). For example, item po14 reads "The government should solve international problems by using diplomacy rather than military force." People who agree with this statement probably would be considered liberal whereas people who disagree probably would be considered less liberal (or more conservative and maybe even neo-conservative). The item, however, seems to reflect a third dimension of liberal-conservative attitudes, in this case attitudes dealing with foreign policy (vs. economic or social issues). Item po7 also is located close to the line and reads "The government should increase support for the U.S. military." Again, this result suggests a foreign policy dimension to the extent that the military acts politically as an agent of foreign policy (e.g., Clausewitz's notion that "war is the continuation of Politik by other means."). Future research thus might consider adding a third "foreign policy" dimension to the economic and social dimensions considered here.
The plot of item po4 may be of interest to some readers. The item reads " I prefer a Libertarian government." As can be seen in the chart, the location of the item indicates respondents to this item interpreted the term libertarian uniquely as an economic issue (given the close proximity of the plot to the horizontal axis). It should be noted, however, that a large proportion of the participants in the poll responded "Don't Know" or "Haven't Thought About It Much" to this item.
So, in summary, what do we mean by "political orientations"? For this study (as used in the following pages), the notion of political orientations consists of five measures:
- Liberal-Conservative Self-Perception: General (1 item)
- Liberal-Conservative Self-Perception: Economic (1 item)
- Liberal-Conservative Self-Perception: Social (1 item)
- Liberal-Conservative Attitudes: Economic (1 summated scale)
- Liberal-Conservative Attitudes: Social (1 summated scale)
Again, the items used in the summated scales are here .