October 2008 Poll - Discussion

Cluster Analysis

Both the theory of psychological type as well as the notion of political types assume people can be grouped together. These groups--or types--should demonstrate homogeneity within the groups and heterogeneity between the groups.  The statistical techniques associated with cluster analysis (or typological analysis ) are designed to create such groups from a given set of data.  Cluster analysis thus seems to provide an interesting approach to assessing the kinds of groups which result from a combination of type data as well as other factors such as demographic factors and political orientations.

The following charts and discussion present the results from various cluster analyses using type data, demographic data (sex and age), and data concerning political orientations (party identification and political orientation measures).  Cluster analysis is an exploratory statistical method and can be used to extract many number of groups.  Just how many groups should be extracted, however, should be guided by some theoretical rationale. 

For our purposes, we extracted both three-group and four-group solutions. Our theoretical rationale for extracting three groups is based on the hypothesis that political party identification (Republican, Democrat, or Independent) is likely to play a significant role in the structure of the data, particularly given the strong relationship between party affiliation and liberal-conservative political orientations.  Thus, e.g., we might expect three groups reflecting conservative Republicans, moderate Independents, and liberal Democrats.  Our theoretical rationale for extracting four groups gives more weight to the hypothesis that psychological type preferences might structure the data. Thus, e.g., we might expect to find four groups defined primarily by the four preference dichotomies. Or in the case of the mental functions groupings, we might expect to find four groups with each group reflecting one of the mental functions group combinations.

Two analyses were conducted for both the 3-group and 4-group analyses. The following variables were included:

  • First Analysis: E-I, S-N, T-F, J-P, party identification (Republican, Democrat, Independent), sex, age and the five political orientation measures.
  • Second Analysis: E-I, J-P, mental functions grouping (ST, SF, NT, NF) party identification (Republican, Democrat, Independent), sex, age and the five political orientation measures.

The results for each analysis are presented in two charts, one for the categorical variables (e.g., E-I, sex, Mental Functions) showing proportions, and one for the continuous variables (age, political orientation measures) showing means. 

Note:  Only those variables that made statistically significant contributions to the groupings are shown. When looking at the charts, the variables with the most influence have scores closer to the circumference of the circle.  The variables with less influence have scores closer to the center of the circle.

 

Three Clusters
 
Type Dichotomies (Results are 

A very significant influence on cluster 1 was party identification in that 100% of those who self-identified with the Republican party fell into this cluster (followed by Independents at 32% and Democrats at 2%). The T-F preferences and sex also contributed to cluster one in that this cluster contained the smallest proportion of Feeling types (13%) and the largest proportion of Thinking types (37%), as well as the smallest proportion of females (15%) and the largest proportion of males (38%).

The most significant influence on cluster 2 was the E-I preference. Approximately 76% of the Introverts in our sample fell into this cluster whereas no Extraverts were in the cluster.  Party identification also contributed to this cluster in that 51% of Democrats are in this cluster followed by 32% of the Independents and 0% of the Republicans.

The most significant influence on cluster 3 was the E-I preference. Approximately 74% of the Extraverts in our sample fell into cluster three.  Party identification also contributed to this cluster in that 47% of Democrats are in this cluster followed by 36% of the Independents and 0% of the Republicans. Finally, the T-F preference contributed slightly with the largest proportion of Feeling types (46%) and the smallest proportion of Thinking types (28%) in this cluster.

   

 

Note: Scores range from 1 (Extremely Liberal) to 9 (Extremely Conservative).  A score of 5 would correspond with a "Moderate" political orientation.

Results for cluster 1 (Republicans) not surprisingly showed the participants in this cluster reported, on average, conservative political orientations.  The results for the self-perceptions (General) and the economic measures were the strongest.  

Results for cluster 2 and cluster 3 (Democrats) were identical. Participants in these clusters reported, on average, liberal political orientations.  The social measures were more liberal than the economic measures.

 
Mental Functions (Results are here. )

The most significant influences on cluster 1 were sex and party identification.  A majority (80%) of the females who participated fell into this cluster (with 0% males).  Most (64%) of the participants who identified as Democrat also fell into cluster 1, with about 37% of the Independents (and 0% Republicans). Mental functions groupings also contributed to cluster 1.  A majority (62%) of the participants who reported NF preferences as well as SF preferences (48%) fell into cluster 1.  However, about 1/3 of ST and NT participants also fell into this cluster.

The most significant influences on cluster 2 were sex and party identification. A majority (61%) of the males who participated fell into this cluster (with 0% females).  Approximately 34% of the participants who identified as Democrat also fell into cluster 1, with about 19% of the Independents (and 0% Republicans).

The most significant influence on cluster 3 was party identification: 100% of participants who identified as Republican fell into this cluster along with approximately 44% of the Independents (and 2% of the Democrats).  Sex and mental functions groupings also contributed to cluster 3.  More male participants (39%) than female participants (20%) fell into this cluster. The smallest proportion of NFs (12%) fell into this cluster while the largest proportion of STs (46%) fell into cluster 3.

   

 

Note: Scores range from 1 (Extremely Liberal) to 9 (Extremely Conservative).  A score of 5 would correspond with a "Moderate" political orientation.

Results for cluster 3 (Republicans) again showed the participants in this cluster reported, on average, conservative political orientations.  The results for the self-perceptions (General) and the economic measures were the strongest.  

Results for cluster 1 and cluster 2 (Democrats) were identical. Participants in these clusters reported, on average, liberal political orientations.  The social measures were more liberal than the economic measures.
 
 

 

 

 

Four Clusters
 
Type Dichotomies (Results are here. )

A very significant influence on cluster 1 is party identification in that 98% of participants who self-identified with the Republican party were in this cluster (followed by Independents at 31% and Democrats at 1%).  The T-F preferences and sex also contributed to cluster one in that this cluster contained the smallest proportion of Feeling types (12%) and the largest proportion of Thinking types (37%), as well as the second smallest proportion of females (15%) and the largest proportion of males (37%). The S-N preferences also made a contribution to the cluster with the largest proportion of S participants (37%) and the second smallest proportion of N participants (20%) falling into this cluster.

The most significant influences on cluster 2 were the E-I preferences followed by party identification and sex.  The majority (66%) of Introverted participants fell into this cluster (with 0% of the Extraverts).  The largest proportion of Democrat participants (44%) also fell into this cluster.  The cluster also contained the largest proportion of female participants (45%) and the second smallest proportion of male participants (17%).

The most significant influences on cluster 3 were the E-I preferences followed by party identification and sex.  The majority (60%) of Extraverted participants fell into this cluster (with 0% of the Introverts).  The second largest proportion of Democrat participants (38%) also fell into this cluster.  The cluster also contained the second largest proportion of female participants (40%) and the smallest proportion of male participants (13%).

The most significant influences on cluster 4 were sex and the T-F preferences.  This cluster contained the second largest proportion of male participants  (33%) and no female participants.  The cluster also was defined by Feeling preferences (26% of F participants) with only one participant with T preferences.

   

 

Note: Scores range from 1 (Extremely Liberal) to 9 (Extremely Conservative).  A score of 5 would correspond with a "Moderate" political orientation.

Results for cluster 1 (Republicans) showed the participants in this cluster reported, on average, conservative political orientations.  The results for the self-perceptions (General) and the economic measures were the strongest.  

Results for cluster 2 and cluster 3 (Democrats) were identical. Participants in these clusters reported, on average, liberal political orientations.  The social measures were more liberal than the economic measures.

Results for cluster 4 showed this cluster reported, on average, liberal orientations.  The economic measures were more liberal than the social measures (and the Attitudes Social scores were not a significant influence).

 
Mental Functions (Results are here. )

The most significant influences on cluster 1 were the J-P preferences and sex. Approximately 51% of the P participants fell into this cluster with 0% of the J participants. The largest proportion (45%) of the females who participated fell into this cluster (with 0% males). Party identification also influenced this cluster with 34% of the participants who identified as Democrat in the cluster along with about 27% of the Independents (and 0% Republicans). Mental functions groupings also contributed to cluster 1.  The largest proportion (38%) of the participants who reported NF preferences  fell into cluster 1 along with the smallest proportion of STs (15%) and SFs (15%).  About 20% NT participants also fell into this cluster.

The most significant influences on cluster 2 were the J-P preferences and sex. Approximately 46% of the J participants fell into this cluster with 0% of the P participants. The second largest proportion (38%) of the females who participated fell into this cluster (with 0% males). Party identification also influenced this cluster with 34% of the participants who identified as Democrat in the cluster along with about 15% of the Independents (and 0% Republicans).

The most significant influences on cluster 3 were sex and party identification. A majority (61%) of the males who participated fell into this cluster (with 0% females). Party identification also influenced this cluster with 34% of the participants who identified as Democrat in the cluster along with about 19% of the Independents (and 0% Republicans).

The most significant influence on cluster 4 was party identification with 100% of Republican participants and 40% of Independent participants falling into this cluster (and 1% Democrat). The Mental Functions groups and sex also influenced the cluster.  The cluster contained the smallest proportion (11%) of NF participants and the largest proportion 44% of ST participants (along with 30% of the SFs and 37% of the NTs).  The second largest proportion (39%) of males also fell into this cluster along with 17% of female participants.

 

Print Email

 
 
 
 
 © Politicaltypes.com 2017. All Rights Reserved.

Share This