October 2008 Poll - Discussion
The primary purpose of the final (October 08) Type and Politics Poll was to examine in more detail the relationship between psychological type preferences and political orientations. Prior research in both the psychological type field as well as research employing the Five-Factor Model (FFM) of personality has suggested that the S-N preferences clearly seem to be related to political orientations in terms of "liberal" and "conservative" orientations. Most of these studies, however, have used only a single item to measure these orientations.
Our purpose in the final poll was to extend this research in two ways. First, we created measures based on the assumption that there are at least two related dimensions of liberal and conservative orientations--one concerning social issues and the other concerning economic issues. Second, we created measures to reflect participants' political self-perceptions of themselves as liberal or conservative on social and economic issues as well as measures that reported political attitudes toward economic and social issues. Our approach to liberal-conservative political orientation thus was multidimensional. Given the various theoretical frameworks that suggest people seek cognitive consistency (or avoid cognitive dissonance) in terms of ego identity, however, we should expect these different orientations to be related (as was shown in the correlations of the five measures).
During our analyses we also kept in mind the ongoing debate surrounding whole types. Thus, we analyzed our data in terms of whole types (and subtypes) as well as at the level of the individual preferences.
Before looking at the results of the analyses, it is important to keep in mind the nature of the sample upon which the analyses are based. In particular, most (66%) of the participants reported being APTi members. Thus, it should not be surprising that a large proportion of the sample members reported preferences for Intuition (77%) and were female (59%). ENFPs constituted the modal type (17%) followed by INTJs (11%) and INFPs (11%). Further, as is not unusual, the SPs constituted only a small proportion (14%) of the sample. These results suggest the subtype analyses likely are more reliable than the whole-type analyses given that the number of observations for each subtype are larger than for the whole-type or Jungian-type analyses.
The tendency of the participants to score more liberal on each of the political orientation measures in the type-only analyses (i.e., those without demographics included) also should be considered. As was shown in each of the type-only analyses, participants, on average, provided slightly liberal to more strongly liberal responses on each of the political orientation measures. This skew of the responses suggests we consider certain types and subtype groups as scoring "less liberal" or "more liberal" than other comparison type groupings. This pattern, however, was different in the cluster analyses where, e.g., Republicans provided clearly conservative responses and Democrats provided clearly liberal responses.
A final characteristic of participants' scores involved the general tendency for scores on the social measures to be more liberal, on average, than the scores on the economic measures. Further, the distribution of the economic scores was more uniform across the economic measures whereas the distribution of scores for the social measures clearly was skewed toward liberal responses. Also interesting: in those analyses where the assumption of equal variances across type groupings were met, the analyses involved the economic measures. These results may be specific to this sample due to the timing of the study during which the sample data was collected; i.e., the study was conducted during the week preceding the 2008 U.S. presidential election when worsening economic conditions clearly were salient. Thus, it is important to keep in mind that this study is a one-shot case study and that these economic and social differences might not be obtained under more normal economic conditions. Regardless, however, the results do indicate how the different types and subtypes responded during these kinds of conditions.
Type Dichotomies. We used regression analysis and the General Linear Model "Multivariate" procedure to test the effects of preference dichotomies. We used this approach given recent empirical and theoretical work suggesting that type effects are more likely to result from the individual type dichotomies alone or in additive relationships.
The results from this analyses focusing on the relationship between individual type preferences and political orientations showed a fairly consistent relationship between the S-N and T-F preferences, and political orientations. The regression analyses demonstrated that these preference dichotomies each were significantly related to overall political orientations such that participants with S and T preferences, on average, scored less liberal than did participants with N and F preferences. More specific regression analyses--focusing on the Social and Economic Attitudes measures--suggested that the S-N preferences had a stronger relationship with Social Attitudes whereas the T-F preferences had a stronger relationship with Economic Attitudes.
Although the type preferences only explained a little less than 10% of the variance in general political orientation, these results were theoretically meaningful and consistent with prior research (both in terms of the nature of the relationships as well as the proportion of variance explained). From the theoretical viewpoint, both in terms of political philosophy as well as political psychology, it should come as no surprise that Sensing preferences correlated with political conservatism whereas Intuition preferences correlated with political liberalism. For example, consider the following quotes from Russell Kirk, a prominent conservative political theorist:
- "Conservatives are champions of custom, convention, and continuity because they prefer the devil they know to the devil they don’t know."
- "The conservative is not opposed to social improvement, although he doubts whether there is any such force as a mystical Progress, with a Roman P, at work in the world."
- "Knowing human nature for a mixture of good and evil, the conservative does not put his trust in mere benevolence. Constitutional restrictions, political checks and balances, adequate enforcement of the laws, the old intricate web of restraints upon will and appetite—these the conservative approves as instruments of freedom and order."
Further, consider this contrast between conservatives and "progressives" provided by conservative David Horowitz in his book Radical Son :
- "...conservatism [is] an attitude about the lessons of the actual past. By contrast, the attention of progressives [is] directed toward an imagined future."
Each of these quotes reflects the Sensing types' preferences for experience, tradition, practicality, what is, and the tried-and-true. Implicit in these quotes also is a concern about "mystical Progress" and an "imagined future" which reflects the Sensing types' view of the Intuitive types' preferences for change, imagination, the future and what could be (possibilities).
Inherent in the conservative's view of change is not so much opposition to change (as noted in the quote above from Kirk), but a cautious approach to change. Certainly some of this caution arises from Edmund Burke's reaction to the dynamics of the French Revolution. (Burke often is referred to as the father of modern conservative thought. And, interestingly, he uses the notion of temperament when considering political change). Conservatives also have the history of various Marixist revolutions and their consequences as models of the results of zealous change in the service of various "social justice" issues: liberté, égalité, fraternité. Again, however, the conservative orientation is not one opposed to change, but one opposed to imprudent change.
At this point it may be worthwhile to note that we are framing our discussion of liberal and conservative politics and their relationship to psychological type preferences in terms of what may be termed traditional conservatism (!) and new liberalism (!). It is a truism with words that their meanings evolve and take on different meanings at different times and in different places. For example, recent political discourse has witnessed the emergence of the terms paleoconservative and neoconservative, and political philosophy makes distinctions between classical liberalism and the new liberalism. Our conceptualization here of conservatism and liberalism thus are referencing the "old conservatism" about which Burke and Kirk write, and the "new liberalism" that is more in keeping with the practiced politics of today's Democratic party.
So, what is a "new liberal" and how might preferences for Intuition incline one to such a political orientation? Our suggestion is that one answer to these questions can be found in the focus that political philosophy places on the notion of autonomy, and the numerous studies showing a positive correlation between continuous scores for Intuition and various psychological scales measuring needs for independence and autonomy (see the 1985 Manual). Given this focus on autonomy, the "new liberal" really is not that new. Rather, the notion of autonomy politically is a part of the evolution of "Modernity" out of the Middle Ages and thus reflects the forces at work during the Renaissance, the Enlightenment , the Protestant Reformation and so on. Indeed, the very names and themes of these historical periods imply intuition at work: inspiration, imagination, originality, creativity. Further, the political connotations of autonomy as a self-reflexivity that reasons free of the influence of tradition--and manifests in such philosophical acts as the hermeneutical act--implies intuitive preferences to the extent that such interpretive reasoning requires one to look at different possibilities where word is treated as symbol (or sign).
This autonomous inclination of the Intuitive's preference to challenge tradition provides one rationale for our finding that Intuitives, on average, scored more liberal on the measure of Social Attitudes than did the Sensing types. That is, our participants with S preferences, on average, were more likely to agree with the following statements than our participants with N preferences:
- The Bible is the actual Word of God and is to be taken literally, word for word.
- This country would have many fewer problems if there were more emphasis on traditional family ties.
- The newer lifestyles are contributing to the breakdown of our society.
As can be seen, each of these statements contains key topics that frequently are associated with Sensing preferences: a literal approach to life, a preference for tradition, and a concern about the "new". Further, given the moral tone of the questions in the Social Attitudes measure, it seems reasonable to ask: "Are the S types more inclined to be social conservatives--and N types social liberals--within the context of the 'culture wars' that have played out in U.S. society and politics over the past 30 years?". Research into the relationship between psychological type and religion provides some support for a "yes" answer to this question. For example, the general theme of results provided from research by Francis is reflected in this quote from one of his works:
A sample of 315 adult churchgoers completed an index of conservative Christian belief together with the Eysenck Personality Questionnaire and the MBTI. Christians who prefer sensing and thinking are more inclined to hold traditional beliefs than Christians who prefer intuition and feeling.
Such findings, when combined with our above discussion, provide further support for the hypothesis that S preferences are related to more conservative political orientations whereas N preferences are related to more liberal political orientations.
Results for the T-F preferences also showed a significant relationship between these type preferences and political orientations. As shown in the results for the regression analyses, participants with T preferences reported less liberal political orientations, on average, that did participants with F preferences. When regressed against the Social and Economic Attitudes measures, the T-F preferences were more significant in predicting Economic Attitudes with T participants again reporting more conservative orientations than F participants.
A look at some of the items from the Economic Attitudes measure provides one perspective for interpreting these results:
- The government should see to it that every person has a job and a good standard of living.
- The government should do more to ensure everyone has an equal chance.
- The government should intervene to reduce income differences.
These items clearly reflect attitudes that support government intervention in the economy that fit well with Democratic policy making that seeks to ensure some sense of égalité regarding economic conditions. Moreover, as evidenced in the recent debates about the economic "stimulus" bill, the Democrat's actions have been oriented to including in the bill such government actions as those which seek to ensure the welfare of those in society who are less fortunate economically in that the bill provided major support for unemployment and food stamp benefits--benefits that clearly reflect a humanitarian orientation. To the extent that F preferences reflect such traits as being empathetic, warm, and concerned with satisfying peoples' needs (as well as a variety of other living critters' needs, in our experience) then the liberal tendency in Democrats' economic policy seems to reflect the expression of the Feeling preference. This rationale, of course, does not eliminate the possibility that some Thinking types also support such economic policy. For the Keynesian or neo-Keynesian Thinking type economist, economic policy that provides such government intervention can be perceived from the more impersonal viewpoint as supporting aggregate demand during a time of economic downturn.
Unfortunately, we have not been able to find any research that employs a psychological type perspective that treats on this topic. Research employing the Five Factor Model of personality, however, does offer some support for our findings. In particular, research in both Germany and Italy, as well as the United States, has shown that research participants who score higher on the measure of Agreeableness (or Friendliness) tend to identify with political parties that support such government intervention as described in the previous paragraph. To the extent that the personality trait of Agreeableness correlates with a preference for Feeling (as is shown in the 1998 Manual), then we have some support for the notion that F types are likely to express liberal economic attitudes more so than T types.
Neither the E-I nor the J-P preferences showed an independent relationship with political orientations. This result may seem somewhat anomalous, especially for the J-P preferences, given that the bivariate correlations between the J-P preferences and most of the political orientation measures were significant. However, as is not unusual, the J-P preferences correlated with the S-N preferences. Thus, the simple relationship of J-P preferences may be mitigated in the multivariate case and represented by the S-N preferences. (The role of E-I and J-P preferences is of interest, however, when we look below at the results of the cluster analysis.)
Type Groupings. We also examined the relationship between various type groupings and the five political orientation measures. The groupings included whole types, Jung's 8 types, mental functions groupings, temperament groupings, the S-N/J-P grouping, the T-F/J-P grouping and the E-I/J-P grouping. Results from the analyses involving these type groupings suggested that various subtype groupings were related to political orientations in theoretically meaningful ways. Both the whole-type and 8-Jungian-types groupings showed statistically significant relationships with the political orientation measures. Given the rather small number of the 4 SP types that were represented in these analyses, however, the reliability of the results should be viewed with caution. Rather, the other subtype groupings had more participants and thus provided more reliable responses.
The mental functions groupings and the temperament groupings provided the most consistent results in that these groupings produced statistically significant effects for each of the five political orientation measures. In particular, the NF types consistently reported more liberal political orientations than did the SJ or ST combination, and also reported more liberal political orientations than did the NT types on most of the political orientation measures.
What is it about the NF mental functions group that predisposes this group to a more liberal orientation than the other groupings? We are aware of no empirical research which supports this finding in either the psychological type or FFM literature. The results from the research by Consulting Psychologists Press, however, do provide some support for this finding. Their results showed that NFs (and INFs in particular) self-identified as Democrat (or Independent) much more so than Republican. To the extent that Democrats self-identify as politically liberal and Republicans self-identify as politically conservative, then the results in our study for the NFs makes some sense. Moreover, the results from Macdaid's1 study showing SJs were overrepresented amongst those who reported they were "extremely" or "very" conservative politically supports the less liberal orientations reported by the SJs in this study.
Our experience with the mental functions groups and organizational politics also provides some theoretical rationale for the results found in this poll. We have discussed this experience here, as well as here, but will summarize that discussion and apply it to liberal and conservative political orientations in this discussion. In particular, our experience with the mental functions groupings and organizational politics derives from our use of two exercises with various student and work groups. In both exercises the participants are first assigned to mental functions groups. Then they are asked to (1) "Draw something that represents your ideal organization" and then (2) "draw a situation at work that your consider to be political". Our conclusions from this experience is that "good" politics is whatever enacts the ideal organization of each mental functions group. "Bad" politics is that which frustrates the enactment of one's ideal organization, typologically defined. The ideal organization for each mental functions group thus represents various characteristics and values with which each group would like to associate. In this manner, the ideal organization represents a form of ideal management or government.
The general themes of the ideal organization drawings reflect key differences in assumptions about the nature of management (or government) that are key to the differences between political liberals and political conservative. These include:
- Assumptions about human nature: Human imperfectibility (conservative) vs. human perfectibility (liberal).
- The degree of trust in efficacy of government: More suspicious (conservative) vs. more trusting (liberal).
- Orientations to governmental evolution: Prefer more traditional, proven institutions (conservative) vs. prefer more innovative organizations (liberal).
- Assumptions about hierarchy: Hierarchy is the natural order wherein individualism is afforded liberty (conservative) vs. community is the source of the social contract which ensures equality (liberal).
This list is neither exhaustive (more here) nor are the various categories mutually exclusive. The list, however, does provide a start for exploring how the mental functions groups manifest their political orientations through their depiction of their ideal organizations, especially with respect to the ST and NF differences.
The ideal organization of the ST groups is one metaphorically summarized as "the machine". This form of organization or government emphasizes the STs preference for a traditional form of organization that has existed for millennia, from the hierarchy of the Catholic Church to that of Max Weber's conceptualization of bureaucracy as an ideal organizational type. This form of government assumes that employees/citizens not only are imperfectible (expressed somewhat as Theory X in managerial thinking and various doctrines of original sin in theology) but this assumption of imperfectibility is manifest in a respect for the law of variety wherein conservatives recognize people naturally differ in individual abilities (e.g., intelligence) that result in natural differences in classes, orders, material conditions, and thus inequality. That the STs are overrepresented in such career fields as law enforcement and the military correlates well with the managerial notion that mechanistic organization represents a command-and-control form of organization. Conservative bureaucracy's orientation to policies, procedures, rules, specialization of labor, centralization of authority and so on address the ST conservatives' organizational attempts to account for the imperfectibility of employees/citizens, provide for gradual organizational evolution through proven means, and provide for a more trusted centralization of authority via policies, procedures, and rules (as has been created through employment contracts in organizational arrangements and voting in the realm of public politics). Thus, within the ST machine we find the Sensing types' preferences for experience, tradition, and the here-and-now combined with the Thinking types' preferences for the impersonal and skeptical.
The ideal organization of the NF groups is one metaphorically summarized as an organization/government that allows for the full expression of the "human spirit". This form of organization or government emphasizes the NFs preferences involving possibilities (N) for people (F). The NFs thus value organizations that allow for creativity, imagination, and innovation along with an abiding respect, empathy, and concern for life in all its relationships. Their ideal organization is one based on the assumption that people are "perfectible" (at least in terms of human growth and development), are trustworthy (Theory Y in managerial thinking), and are best governed in a manner that allows for an ongoing dialogue that seeks consensus problem solving (equality) in the service of a community's betterment. The tendency of NFs to be overrepresented among such occupational groups as the clergy, conselling psychology, and the fine arts demonstrates the expression of NF energy in the realm of the human spirit. Power is expressed not through Weber's rationalized bureaucracy but, rather, though a social contract that emerges from the ongoing dialogue that takes place within a community of commitment and respect, and incarnates a postmodern version of Rousseau's general will of the people. Rather than the constitutional federalism of the conservative Republican's representative democracy, the NFs' ideal organization finds expression (via Fichte) in Mary Parker Follet's concept of democracy as "a great spiritual force evolving itself from men, utilizing each, completing his incompleteness by weaving together all in the many-membered community life which is the true Theophany." Thus, within the NF ideal organization we find the Intuition types' preferences for creativity and innovation combined with the Feeling types' preferences for the nurturing the whole.
The results from the analyses of the effects of the mental functions groupings on political orientations showed the NFs scored more liberal than the STs on 4 of the 5 measures of political orientations (Social Attitudes was the exception). In some of these analyses, however, the NFs scored as significantly more liberal than the SF and NT groups (as well as the ST group) while these groups (SF and NT) did not differ significantly from the ST groups. Based on our experience with the ideal organization exercise, we speculate that the SF and NT groupings produce ideal organizations that, while somewhat more egalitarian (or "decentralized") than the ST ideal organization, the NF orientation to egalitarianism is much more pronounced than that of the other groupings (especially when the NF group is predominantly NFP).
For example, the ideal organization for the SF groups is one metaphorically summarized as "small family" or "team". Within this organization there clearly is an emphasis on equality of the team members and respect for the uniqueness of each team member. Invariably, however, the SF groups recognize an overarching goal or authority that governs the group (and thus implies some degree of hierarchy). A couple of examples might help. In one situation, the SF group drew a key holder much as one might find in one's kitchen. The structure holding all the keys can be seen as the hierarchically organizing principle of the group. Each group member, however, drew his or her own unique key with his or her name on the key--representing the uniqueness of each member. In another situation, the SF group drew an arrow. In this case, the single head of the arrow represented the group's unity of purpose and direction. However, each member of the group also drew in their own unique portion of the tail feathers that would guide the arrow--again representing the uniqueness of each member.
We think these metaphorical portraits reflect validly some of the liberal and conservative differences between the mental functions groups. As with any use of metaphor, however, some aspects of a phenomenon are emphasized while others are de-emphasized. For example, we believe the ST characterization presented above clearly emphasizes many aspects of conservatism. On the other hand, characterizing conservatism as mechanistic in the sense of the ST ideal organization produces something of a conundrum in that the mechanistic form of industrial organization was something that early conservatives seem to have regarded with disdain. As discussed in Kirk's book The Conservative Mind , early conservatives viewed the rise of the industrial organization as emblematic of a growing secularism and a decline of a traditional, more agrarian and religious society. To the extent that various writers of the Romantic movement were involved in this conservative critique of industrialization (as Kirk notes), then we have the interesting hypothesis of the NF conservative, in that the Romantic movement emphasized intuition and "feeling" (N and F) over the evolving empiricism and rationality (S and T). Such relationships deserve further investigation, as well as remind us that politics indeed may make strange bedfellows, even typologically.
Cluster Analysis. We also performed a cluster analysis using participant responses to the October 08 poll. Cluster analysis is a statistical technique that seeks to form homogeneous clusters that differ from other clusters. Cluster analysis sometimes is refered to as "typological analysis" and thus seems quite appropriate for deriving various clusters of "political types" from our data. We performed two cluster analyses. The first extracted three clusters and the second extracted four clusters.
Unlike the above analyses (dichotomies and type groupings), the cluster analyses included participants' self identification as Republican, Democrat, or Independent. Thus, the variables included in the cluster analyses included type preferences, party identification, sex, age, and the five political orientation measures. In both the 3-cluster and 4-cluster analyses, we explored the role of type as dichotomies only (E-I, S-N, T-F, and J-P) and as mental functions groupings (ST, SF, NF, and NT) with the E-I and J-P dichotomies also included.
Results from the analyses produced both common themes and a some interesting differences. One consistent theme involved participants who identified with the Republican Party. This cluster invariably included more than 98% of the Republican participants, significantly more males than females, and the smallest proportion of F and NF types. The cluster also included a moderate proportion of T and ST type participants (with the NT types modestly represented in one analysis). As might be expected, this cluster reported conservative scores on the political orientation measures with economic conservatism being stronger than social conservatism. The proportion of Independents in this cluster ranged from about 27% to 44% depending upon the analyses. Fewer than 2% of Democrats fell into this cluster in each of the analyses.
A second theme was the tendency of the remaining clusters to be defined largely by Democrats followed by Independents. Further, the remaining clusters reported strong liberal orientations with social orientations generally more significant than economic orientations. The remaining clusters, however, differed in terms of the influence of the following variables: E-I, T-F, J-P, mental functions, and sex.
In the analyses where type preferences were included as the four separate dichotomies, two of the remaining clusters were distinguished by E or I preferences. The E cluster contained a large proportion of E participants along with a moderate proportion of participants who identified as Democrat or Independent. The E cluster also contained a moderate representation of F types (and the fewest T types) or a larger proportion of females than males (depending upon the analysis). The I cluster contained a large proportion of of the I participants and a moderate proportion of Democrats and Independents. In the four-cluster solution, the I cluster also contained a moderate proportion of females (and few males). A fourth cluster was not defined by the E-I dichotomy; rather this extra cluster contained about 1/3 of the male participants and about 1/4 of the F types (with only 1% of the T types). Party membership was not a significant influence on this group. And, in contrast to the other non-Repubican clusters, this cluster tended to express stronger economic liberalism.
In the analyses where type preferences were included as mental functions (along with E-I and J-P dichotomies), the remaining clusters largely were defined by sex and the J-P preferences, along with the mental functions grouping. The three-custer solution produced one cluster defined by a large proportion (80%) of female participants (with 0% males) and a large proportion (62%) of the NF participants. A majority (64%) of Democrat participants also were in this cluster along with about 37% of Independents. The second cluster was a male cluster (61% of male participants and 0% female) along with about 34% of the Democrat participants and 18% Independents. In the four-cluster solution, the female-Democrat-NF group split off another cluster, distinguished largely by the J-P preferences. The female-Democrat-NF cluster became a female-Democrat-NF-P group (with slightly smaller proportions of Democrat and NF participants--but with no J participants). The new cluster was a female-Democrat-J cluster. The cluster contained no males, no P types, and no Republican participants.
The foregoing discussion of the results from the cluster analyses raise at least three questions for further discussion:
- Why did only one "Republican" cluster emerge whereas two or three "Democrat/Independent" clusters emerged?
- Why did the E-I and J-P preference dichotomies influence the cluster formations when no significant effect was found for these preferences in the other analyses?
- Why did participant sex influence the formation of so many clusters?
As an answer to the first question we must quote Will Rogers: "I belong to no organized political party -- I am a Democrat". This quote reflects an ongoing stereotype about Democrats, and thus Republicans, in that Republicans generally have been seen--at least over the past decade or so--as the talking-points-memo party. This characterization suggests Republicans are more unified and organized in contrast to Rogers' sarcasm suggesting that Democrats are not so organized. The correlations between participants who identify with the Republican party and their type preferences as E, S, T, and J also suggests that Republicans may have a more unified and organized political self-identity than do Democrats. Political dynamics over the past few years leading up to the 2008 U.S. presidential election would, however, seem to discredit this perspective somewhat. In particular, the Republican party clearly has become more divided as latent schisms between social conservatives, economic conservatives, neo-conservatives, and so on have become more apparent as a result of the 2008 election process. Rather, another answer to this question may come from the limitations in the distribution of the sample data. That is, given the relatively smaller proportion of participants in the poll who identified as Republican, the sample size may not be large enough to provide an opportunity for differences in this cluster to emerge (and form other Republican clusters).
Given these limitations, however, the results for the "Republican" cluster accord with the perspective of Republicans as more likely to be conservative politically and male than are Democrats. That T types and ST types are more frequently represented in the Republican cluster while F and NF types are only marginally represented seems reasonable given the liberal-conservative orientations for these types discussed above.
Our second question involves the E-I and J-P preferences. These preferences showed no significant relationships in the analyses focusing solely on the relationships between type and political orientations. Yet, these preferences played a significant role in the formation of some clusters. Why the different effects? One key to answering this question is to keep in mind that the cluster analyses included party identification whereas the other analyses did not. Further, research using the FFM suggests that personality and party identification are related such that Es and Js are likely to identify with more conservative or rightist political parties whereas Is and Ps are more likely to identify with more liberal or leftist parties. Unfortunately, these FFM results do not seem to apply in our situation given that the E-I and J-P effects are influential only in discriminating non-Republican clusters. Further, the influence is to discriminate clusters which generally contain females, Democrats, F types, and NF types. Like other results in this study, these results may be due to the distribution of the participants in the study. However, the results from the 1998 Consulting Psychologists Press study do provide an interesting perspective. In that study, INFs were the least likely of all the 16 types to self-identify as Republican and the INFJs were the most likely of all the whole types to self-identify with Democrats. ENFs did not show as clear a discrimination between party leanings with both the ENFPs and the ENFJs self-identifying almost uniformly across all three party identifications: Republican, Democrat, and Independent. When combined with the results from the cluster analyses, the implication is that (at least in the NF case) the E preference is associated with identification with a variety of parties whereas the I preference is associated with a disinclination to identify as Republican. Further, the P preference is associated with a greater willingness to identify as Independent whereas the J preference is negatively associated with Independence.
Our final question is: why did participant sex influence the formation of so many clusters? As discussed above, the distribution of males and females varied significantly across some clusters. In particular, the Republican cluster consistently contained a larger proportion of male participants than female participants. In the mental functions analyses there was a clear tendency for a male-Democrat cluster to emerge separate from a largely female-Democrat-NF cluster. Further, the J-P influence was evident only in female clusters, with 0% males in these two clusters. One possible answer to this question, again, is that the results are indigenous to the participant pool. Another more theoretical speculation is provided by those authors and commentators who have stereotyped the political parties in terms of the sexes. For example, political commentator Chris Matthews has referred to the Republican Party as the "Daddy" party and the Democratic party as the "Mommy" party. Similarly, University of California, Berkeley linguistics professor George Lakoff has referred to liberal and conservative political orientations in terms of family dynamics where liberals are framed in terms of the "Nurturing Parent" metaphor and conservatives are framed in terms of the "Strict Father" metaphor.
These speculations reflect an ongoing debate in the social sciences as well as the psychological type community about the "gendering" of various social and psychological constructs. For example, the effect of sex-role stereotyping in the realm of occupations has long been studied . Moreover, the type community has long focused on sex differences in responses to the T-F index (and the fact that 76% of females scored as F in the norming of the Form M appears to do little to assuage these concerns). Research on the role of sex in terms of party membership, attitudes toward various political issues, and voting behavior, however, suggest that there are differences along these lines both within parties by sex as well as between parties by sex. Thus, a study with more conservative and Republican participants is needed along with studies that discriminate between sex (male and female) and gender (masculinity and femininity) as they relate to type and political orientations. The general regression results reported above further support this need for research in that males and T types, more so than females and F types, reported more conservative political orientations.
As we have noted in various parts of our discussion, there has been an ongoing debate in the type community about what constitutes a psychological type and how best to statistically analyze type data. On the one hand, there are those2 who have argued for something of a "subgroup" approach where each whole type is taken as the unit of analysis (especially when using correlations as a foundation for analysis). When applied to this study, such an approach would seem to require, e.g., that our factor analysis of the political orientations measures be conducted on each of the 16 types separately, with subsequent analyses similarly constrained. To employ such an analysis would require a sample size of at least 10 responses for each of our 20 political orientation items for each of the 16 types, or 10x20x16=3600 participants. Clearly, our sample did not allow for such an examination.
On the other hand, there are those3 who have argued that the whole-type perspective--along with the associated notion of type dynamics--has not met the test of empirical examination. Rather, the suggestion amongst these researchers is that type effects are most likely to occur at the level of the individual types alone or in additive combinations. Our analyses, to some extent, have been able to test this view and the results here generally are supportive of this viewpoint. For example, our regression analysis indicated that main effects--especially for the S-N and T-F preferences--were related to our political orientation measures in theoretically meaningful ways. Further, our tests for relationships between interactions of type variables and political orientation measures met with no success. On the other hand, our analyses of the subgroups suggested the type preferences do combine to produce effects on political orientations.
Within the context of these two consideration, we thus offer the results here as a limited examination of the relationship between psychological type preferences and political orientations. And, hopefully the results provide suggestions for future research. We cannot say, however, that the results provide any rigorous insight into the ongoing debate discussed above.
1Macdaid, G. P. (1999, February). Facts from the CAPT databank. Paper presented at the Second Biennial Clinical Conference of the Center for the Applications of Psychological Type, Gainesville, FL.
2Mitchell, W. D. (2000). Cautions regarding aggregated data analysis in type research. Journal of Psychological Type, 53, 19-30.
3Reynierse, J. H. (2009, January). The case against type dynamics. Journal of Psychological Type, 69(1), 1-21.