xlm-XLM-2014-1640-BhatiaPersonality

advertisement
Supplementary Materials: Personality, Persuasion and Argumentation
Bayesian argumentation focuses on the effect of content on persuasion (Hahn & Oaksford,
2007). In the social psychology of persuasion, a range of non-content factors have been
shown to be influential, affective and personality variables in particular (for reviews, see,
Eagly & Chaiken, 1993; Johnson, Maio, & Smith-McLallen, 2005). We therefore also
investigated whether the evaluator’s personality influences these different ad hominem
arguments. The variety of contexts and argument forms are suggestive in this respect. For
example, work on message-person congruence (Hirsh, Kang, & Bodenhausen, 2012), would
seem to predict that conscientious individuals, who value order and efficiency (Roberts,
Chernyshenko, Stark & Goldberg, 2005), may reject the use of ad hominem arguments in a
scientific context.
On completing the argument evaluation task, all participants completed the ten item
personality inventory (TIPI) which is readily available on-line and which has, “adequate
levels…[of] convergent and discriminant validity, test–retest reliability, patterns of external
correlates” (Gosling, Rentfrow, & Swann, 2003, p. 523). This instrument provides direct
estimates of the big five personality traits: Extraversion, Agreeableness, Conscientiousness,
Emotional Stability, and Openness to New Experiences.
TABLE 1 ABOUT HERE
To investigate whether the personality of the evaluator affects how persuasive
participants found the ad hominem argument we used the difference between the posterior
and prior ratings as our index of argument force.1 Together with twelve argument
1
We also repeated these analyses on the posterior ratings as an index of argument strength and very similar
results were obtained.
type/context combinations this led to 136 possible correlations. To guard against Type 1
errors, we therefore carried out an exploratory principal components analysis using all
seventeen variables (for the correlation matrix, see Correlations in the Supplemental
Materials available on-line). Six components with eigenvalues greater than 1 were extracted
using the maximum likelihood method and rotated using the Varimax procedure. Summary
statistics for the rotated component solution are presented in Table 1. If the character of the
evaluator influenced how forceful an argument was perceived to be we would expect to find
both argument/context and personality variables loading strongly on one or more of the six
extracted components. Table 1 shows that only the 5th and the 6th components had loadings
from both groups of variables. However, the 5th component loads only -.32 on
conscientiousness and so this is really a single variable abusive ad hominem/scientific context
component. Moreover, for the 6th component the loadings are not high enough for them to be
regarded as reliable (Stevens, 1992, p. 384) and the two argumentation variables loaded more
highly on the 1st component.
The 5th and 6th components make some intuitive sense. The 6th component suggests
that agreeable and emotionally stable individuals do not regard the abusive and tu quoque
arguments as persuasive in a domestic context. Moreover, the similarly bipolar 5th component
seems to show that conscientiousness is inversely related to acceptance of the abusive ad
hominem in a scientific context. While eminently reasonable, further research is required to
establish if such effects are reliable. Provisionally, therefore, we must conclude that these
data show little evidence of any influence of the character of the evaluator on judgements of
the force of the ad hominem argument in these contexts.
It is perhaps unsurprising that we found no effects of personality. Children as young
as five are very discriminating about weighing the testimony they receive from different
sources (Harris, 2012). Moreover, the kinds of testimony provided in these brief ad hominem
arguments was quite stereotypical political, domestic, and scientific debates of which
participants are likely to have had some experience. The fallacies are also likely to be
culturally well embedded. Although not explicitly taught at School (at least not in the UK),
critical thinking is actively encouraged in the School curriculum. “Creativity and Critical
Thinking” is one of seven key cross-curriculum dimensions identified by the UK’s
Qualification and Curriculum Authority. This will involve implicit training in avoiding
fallacies whether explicitly named or not. While people are differentially susceptible to
persuasive messages, argument forms that are culturally and socially well embedded and
important in the development of belief formation mechanisms may be much less susceptible
to the effects of personality variables.
References (not included in main article)
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Orlando, FL, US: Harcourt
Brace Jovanovich College Publishers.
Gosling, S. D., Rentfrow, P. J., & Swann, W. R. (2003). A very brief measure of the Big-Five
personality domains. Journal of Research in Personality, 37, 504-528.
doi:10.1016/S0092-6566(03)00046-1
Hirsh, J. B., Kang, S. K., & Bodenhausen, G. V. (2012). Personalized persuasion: Tailoring
persuasive appeals to recipient’s personality traits. Psychological Science, 23, 578581.
Johnson, B. T., Maio, G. R., & Smith-McLallen, A. (2005). Communication and attitude
change: Causes, processes, and effects. In D. Albarracín, B. T. Johnson, & M. P.
Zanna (Eds.), The handbook of attitudes (pp. 617-669). Mahwah, NJ: Erlbaum.
Roberts, B. W., Chernyshenko, O. S., Stark, S., & Goldberg, L. R. (2005). The structure of
conscientiousness: An empirical investigation based on seven major personality
questionnaires. Personnel Psychology, 58, 103–139.
Stevens, J. (1992). Applied multivariate statistics for the social sciences (2nd ed.). Hillsdale,
NJ, England: Lawrence Erlbaum Associates, Inc.
Table 1
Rotated Component Matrix - Loadings
Component Variables
1st
Abuse/Dom
TuQ/Sci
TuQ/Dom
Circ//Sci
Circ/Pol
2nd
Cont/Sci
Cont/Dom
TuQ/Pol
Circ/Dom
3rd
Openness
4th
Cont/Pol
Abuse/Pol
5th
Abuse/Sci
Conscientious
6th
Agreeable
Emote
Not inc.
Extra
Final Statistics
Component Eigenvalue
1st
1.73
2nd
1.24
3rd
1.21
4th
1.12
th
5
0.97
6th
0.78
1st
2nd
3rd
4th
5th
6th
.62
.52
.46
.62
.34
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
.49
.38
.69
.36
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
.98
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
.58
.68
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
.73
-.32
*
*
*
.36
*
.39
*
*
*
*
*
*
*
*
*
*
*
-.38
-.32
*
% Var
10.18
7.31
7.09
6.60
5.73
4.58
Cum. % Var
10.18
17.50
24.59
31.19
36.91
41.49
* Indicates a component loading less than .30. Abuse = abusive ad hominem,
Circ = circumstantial ad hominem, Cont = control, TuQ = tu quoque ad
hominem, Sci = scientific context, Pol = political context, Dom = domestic
context, Openness = Openness to experience, Conscientious =
Conscientiousness, Agreeable = Agreeableness, Emote = Emotional stability,
Extra = Extraversion, Not inc. = not included in any component, % Var =
percentage variance accounted for, Cum. % Var = cumulative percentage
variance accounted for.
Table 1. Summary statistics for the principal components analysis of the combined
argumentation and personality data.
Download