PSYCH 2B03 – (7) Factor Analytic Trait Theories.docx
53 views6 pages
11 Jul 2014
School
Department
Course
Professor
PSYCH 2B03 – Personality – November 7th
FACTOR ANALYTIC TRAIT THEORIES:
Reason for its dominance has nothing to do with it being more accurate. It allows for quantification,
putting numbers on trait dimensions. Primarily when we quantify them we can do further analytic
work. Doesn’t mean its any better or the right theory. Just different ways of understanding personality.
 We tend to describe ourselves in terms of traits. They refer to positions along some trait
dimension. There are many synonyms. What were the fundamental traits and factors in
personality?
 AG are trait dimension, authenticity, genorousity, honesty, etc. The individual cells are the
correlations between the ratings. Shows a pattern of intercorrelations. This shows only positive,
but usually some are negative.

 Positive groups of traits for a high number tend to be highly correlated. The reason they maybe
highly positive correlated is bc they measure the same underlying “factor”. Don’t know what
that factor is or why it exists, but there is an underlying something that determines the ratings
on several different traits. Only explanation is a pattern of correlation that contributes to both
traits.
 Factor analysis provides us with a mathematical model and how to account for these patterns of
intercorrelations by hypothesizing that there is a small number of underlying factors. Maybe a
section of correlation of traits, could be correlated bc of a similar idea that contributes to both.
 They used a technique called factor analysis. About 100 trait descriptions, give it to a 1000
people, and ask them to describe themselves along a point scale. Ratings given from 110 of
how much it relates to you. Ratings of some of these traits are consistently correlated with each
other. Positive correlations and negative correlations.
 A group picked out by math, shows that this group of correlations bc there is some underlying
factor responsible for the correlation. Our math model will produce a factor mathematically in
an equation.
The first factor: F1: is a combo of each original correlation value for the overall factor.
Factor loading is the degree to which each of the individual traits (C, D ,E) is determined or shaped by
underlying factor. How much does the underlying factor contributes to the factor. If 0.8C + 0.6D, then
it shows that C contributes more to the factor than D. The original correlation of each trait that
contributes to the overall factor. Name our first factor. It does not tell you what this is from the factor
analysis. Supposed that C, a trait with a high correlation in
this factor, is that
C = likes to party
D = seeks out sensations
E = likes to take risks
F – learns more slowly
G = leadership … You can call that extraversion.
The name we give to it depends on what items are affected in the factor. Brings together all the things
we associate with extraversion. Naming based on the items we see are interrelated.
Factor 2: also has factor loading. The extent to which each of these various traits is related to the
overall factor. Sometimes both factor 1 and factor 2 share similar factors. Like both influence C.
 Factors are nonorthogonal. Both factors can contribute to the same sum of trait dimensions.
Factors themselves are intercorrelated. Like small intercorrelation between neuroticism and
stability.
 All factor analysis does is provide a mathematical models for how one might account for
intercorrelated measures by hypothesizing the existence some small number of underlying
factors.
 These factors are hypothetical constructs. By math does not make them more real. No more real
than superego, id or anything else. Only trait theorists use this as it correlates traits. Rogers,
Jung or Bandura/Mischel could have used it. But they didn’t use this.
 Trying to understand the factor analytic structure of trait dimensions.
What Factor Analysis Does:
•It does NOT find ‘real’ things, it describes hypothetical constructs.
•It does NOT identify factors – it doesn’t say it’s a set of intercorrelated traits, it doesn’t tell you
what the underlying factors are. They could be biological or anything, doesn’t tell you anything.
•NOT necessarily traits – they just use it on trait data
•Results depend on measures – like personality tests instead of trait dimensions.
•It depends on how you set it up!!! Take the same data, run factor analysis on it, depends on how
you restrict it. No one definitive factor analysis solution. Number of ways to do it.
•Results depend on parameters – orthogonal vs. nonorthogonal factors. You can set it up to get
orthogonal factors, only identify a model in which they are unrelated factors, like Eysenek. Or
show me any possible way of understanding the correlation, which is the nonorthogonal, you
end up with more factors. You have removed the restriction, factors could be correlated.
What makes a Factor ‘Basic’:
Reliable – stable over time and observers. You must get similar ratings over time, your relative
position on these factors should be fairly constant. Bandura wouldn’t make this assumption, as
they can be changing. Costa and McCrae see personality as unchanging.
Used by both theorists and laypersons – Everyone can know something about personality, they
should all be consistent with how they describe personality factors.
Appear across cultures – this factor should appear everywhere. It’s about basic human nature.
Must have some biological basis – it’s present at all ages. Primarily biological.
Early Factor Analytic Theories:
•Raymond Cattell was the first to do factor analytic theory to trait theories.
•He developed the 16PF, personality factor test. 16 factors and you run out of names. He created
greek names: Affectia, Premsia, Surgency, Tensidia.
•Methodology is influencial.
•Eysenck’s PEN use orthogonal factors: Extroversion, Neuroticism, Psychotism. Unrelated to
each other.