Temperament and the Big Five
Running Head: TEMPERAMENT AND THE BIG FIVE
A Hierarchical Model of Temperament and the Big Five
David E. Evans and Mary K. Rothbart
University of Oregon
Key words: Big Five, personality, temperament
1
Temperament and the Big Five
2
Abstract
Temperament constructs and their relation to measures of the Big Five factors of
personality were explored. In Study One, 207 undergraduates completed the Adult
Temperament Questionnaire (ATQ). After minor modifications, a hypothesized four-factor
model fit the temperament data well. The four general scale scores were highly correlated with
four Mini-Marker scales assessing Big Five personality factors. In Study Two, 258
undergraduates completed a revised ATQ, with constructs for affiliativeness and aggressive
negative affect added. An alternative five-factor model (FFM) emerged, with factors labeled
Cognitive Sensitivity, Effortful Control, Extraversion, Affiliativeness, and Negative Affect. This
model converged with the Big Five. In Study Three, the model from Study Two was confirmed
in a community sample of 700 participants, and convergence with the Big Five of personality
was replicated. Temperament may form the biologically-based core of the Big Five domains.
Temperament and the Big Five
3
A Hierarchical Model of Temperament and the Big Five
Although theorists differ in how they define temperament, there is general agreement that
temperamental processes are rooted in biological systems, and that emotion is basic to
temperament (see Goldsmith et al., 1987). Ongoing work in animal neurophysiology, human
brain imaging, and molecular genetics has led to psychobiological models of temperamental
processes that are becoming increasingly compelling and comprehensive (Cloninger, 1998; Gray,
1990; Panksepp, 1998). Based upon heritability research, animal personality, and stability across
development and cultures, McCrae and Costa (McCrae et al., 2000; also see McCrae & Costa,
1996) have recently suggested that each of the Big Five/FFM domains has a temperamental base.
To facilitate the development of temperament models and to investigate relations between
temperament and personality, it is essential to develop psychometrically sound measures of
temperament constructs. The studies included in the current report were motivated by the desire
to explore the hierarchical relations among lower level constructs of temperament, as well as the
relation of temperament to the lexical Big Five model of personality.
One method of developing questionnaire assessments involves a rational approach in
which constructs are explicitly defined, and questionnaire items written to fit the definition of the
construct to which they belong. Researchers have developed rational constructs they consider to
be fundamental to understanding temperament (see Strelau & Zawadzki, 1997). Buss and
Plomin (1984; Buss, 1991), for example, developed a temperament model including scales for
emotionality (fear and anger), activity level, and sociability. Strelau (Strelau & Zawadzki, 1993)
developed a model assessing briskness, perseveration, sensory sensitivity, emotional reactivity,
endurance, and activity. Derryberry and Rothbart (1988) developed the Physiological Reactions
Questionnaire (PRQ) to explore relations among temperament constructs related to arousal,
affect, and attention. Defining temperament as individual differences in reactivity and self-
Temperament and the Big Five
4
regulation as reflected in the emotions, activity, and attention (Rothbart & Derryberry, 1981),
they emphasized definitional specificity. PRQ scales were operationally defined, and items
generated to fit these definitions. Conceptual and empirical analyses were then performed to
remove items that were more strongly related to scales other than the scales to which they were
assigned. Scales were developed to assess multiple components within the affective, arousal,
and attentional domains and factor analysis was performed on data from a sample of
undergraduate students. A negative affect factor with loadings from fear, frustration, sadness,
and discomfort emerged. In a three-factor solution, scales assessing effortful control over the
shifting and focusing of attention loaded negatively on this factor, whereas in a six-factor
solution, one factor included loadings from the negative affect scales and another, loadings from
attentional or effortful control scales. The first factor for both the three- and six-factor solutions
included perceptual sensitivity (a cognitive sensitivity construct) as a core component.
Perceptual sensitivity scales were uncorrelated with the scales that included an effortful
component. In this study, the possibility of further differentiating individual differences in
reactive and effortful aspects of attention was explored.
This research has modified and expanded the original PRQ instrument into the Adult
Temperament Questionnaire (ATQ) to explore the possibility of higher-level (i.e., broader bandwidth) constructs. Using questionnaire measures to assess children, Rothbart, Ahadi, Hershey,
and Fisher (2001) found three broad factors (negative affect, extraversion/surgency, and effortful
control) that reflect higher level constructs of this sort. Higher level, and higher-level constructs
of reactive and effortful attention were also developed, and we were interested in exploring the
earlier findings that effortful attention (i.e., attentional shifting and focusing) and negative affect
scales were both separable and related, depending on the number of factors extracted. Finally,
Temperament and the Big Five
5
the PRQ had included only one extraversion/surgency-related scale, and we wished to further
explore the construct by including theoretically related scales.
Three of the four broad constructs broken down into scale-level constructs for Study One
were consistent with broad factors found in Rothbart’s (Ahadi, Rothbart, & Ye, 1993; Rothbart
et al., 2001) caregiver report Children’s Behavior Questionnaire (CBQ). Preliminary analysis of
the Early Childhood Behavior Questionnaire (ECBQ: Putnam, Ellis, & Rothbart, 2001) resulted
in a similar factor structure. These factors are labeled surgency/extraversion, negative affect, and
effortful control. We included these three general constructs along with cognitive sensitivity in
the initial version of the ATQ. Definitions and sample items for the scales included in these
studies are given in Appendix A. Descriptions of the four general theoretically and empirically
based constructs and their sub-constructs are now presented.
Evidence from research on childhood temperament (Putnam et al., 2001), adult
temperament (Strelau & Zawadzki, 1997), personality (Eysenck & Eysenck, 1985; Tellegen,
1985), neuroscience (Carver & White, 1994; Davidson, 1993; Depue & Collins, 1999;
Derryberry & Tucker, 1992; Gray, 1990), and affective individual differences (Watson & Clark,
1992; Watson & Walker, 1996; Watson, Wiese, Vaidya, & Tellegen, 1999) suggests the
existence of at least two high level temperamental motivational-emotional systems. One of these
systems is associated with potentially aversive stimuli and negative affect, the other with
potentially appetitive stimuli and the experience of positive affect. Our labels for these
constructs are negative affectivity and extraversion/surgency, respectively.
Most models of temperament and personality include variants of negative affect factors
(e.g., negative emotionality, Harkness, Tellegen, & Waller, 1995; Tellegen, 1985; negative
affectivity, Ahadi et al., 1993; Rothbart et al., 2001; Watson et al., 1999; neuroticism, Eysenck &
Eysenck, 1985; McCrae & John, 1992; and emotional stability, Goldberg, 1993). In Derryberry
Temperament and the Big Five
6
and Rothbart’s (1988) PRQ study, the negative affect factor included loadings from scales
assessing fear (unpleasant affect associated with anticipation of pain or distress), frustration
(unpleasant affect associated with task interruption or the blocking of a desired goal), sadness
(unpleasant affect and lowered mood related to disappointment, loss, and exposure to suffering),
and discomfort (unpleasant affect resulting from the sensory qualities of stimulation) (Derryberry
& Rothbart, 1988). These scales were also included in Rothbart’s (Rothbart et al., 2001)
Children’s Behavior Questionnaire (CBQ), and modified versions of these scales were retained
for the ATQ.
As for neuroticism-negative affectivity constructs, most models include variants of a
general extraversion/surgency construct (e.g., positive emotionality, Harkness et al., 1995;
Tellegen, 1985; positive affect, Watson et al., 1999; extraversion, Eysenck & Eysenck, 1985;
McCrae & John, 1992; extraversion/surgency, Goldberg, 1993). In the CBQ caregiver report
measure, a surgency/extraversion factor has also consistently emerged (e.g., Ahadi et al., 1993).
This factor includes loadings from activity level (vigor and tempo of motor activity), high
intensity pleasure (enjoyment related to high levels of novelty and intensity), impulsivity (speed
of response initiation), and shyness (negative loading; unpleasant affect and/or behavioral
inhibition in novel and social situations), as core constructs. Sub-constructs for
extraversion/surgency in the ATQ included in Study One were high intensity pleasure, activity
level, sociability, and positive affect.
The original PRQ study included only one scale with surgency/extraversion-related
content (high intensity pleasure). The PRQ high intensity pleasure scale was to a large extent a
sensation-seeking construct, including items related to affective preference for skydiving or
racecar driving. For the ATQ version, we replaced most of the items for this scale in an effort to
remove the influence of fear on responses. One might enjoy the idea of skydiving, for example,
Temperament and the Big Five
7
while also being fearful of jumping from an airplane. A new sociability scale was defined as
enjoyment from interacting and being in the presence of others, distinguished from fearful and
shyness in interactions with others (see Buss, 1991, for a discussion distinguishing sociability
from shyness). In addition to high intensity pleasure and sociability, scales assessing positive
affect (intensity, duration, frequency, rate of onset, and rising intensity of pleasure) and activity
level were also included.
Rothbart, Derryberry, and Posner (1994), in the tradition of Thomas and Chess (1977),
view attentional processes as components of temperament. Effortful attention and cognitive
sensitivity constructs were included in the initial development of the ATQ. Factors labeled
effortful control have emerged from analyses of two caregiver report questionnaires developed
by Rothbart and colleagues: the CBQ (Ahadi et al., 1993; Rothbart et al., 1994; Rothbart et al.,
2001) and the Early Childhood Behavior Questionnaire (ECBQ: Putnam et al., 2001). In
addition, a robust effortful control factor emerged from analysis of a revision of a self-report
measure, the Early Adolescent Temperament Questionnaire (Capaldi & Rothbart, 1992; Ellis &
Rothbart, 2001). The PRQ (Derryberry & Rothbart, 1988) also included scales assessing
effortful control at the adult level.
The initial ATQ did not include behavioral control scales (e.g., inhibitory control).
Hence, the Study One effortful control construct is labeled effortful attention (a narrower
construct). In addition to attentional shifting and focusing scales from the PRQ, attentional
shifting from punishment (ability to control attention in the face of negative emotions and
thoughts) and attentional shifting from reward (ability to control attention in the face of
distracting positive emotions and desires) were added. These scales were intended to explore the
structure of an effortful attention factor, as well as ways in which attentional control may interact
with positive and negative emotionality.
Temperament and the Big Five
8
A second broad construct from the PRQ was related to cognitive sensitivity. It included
constructs of external perceptual sensitivity (awareness of slight, low intensity stimulation
arising from the external environment), internal perceptual sensitivity (awareness of slight, low
intensity stimulation arising from within the body), and cognitive reactivity (general internal
cognitive activity). Our interest was in exploring the separability of reactive and effortful
attentional processes, and breadth of sensitivity was emphasized in the ATQ. The cognitive
reactivity scale was replaced with associative sensitivity, defined as the frequency, diversity, and
remoteness of automatic conscious content not related to standard associations with the
immediate environment.
We were also interested in exploring the extent to which affective awareness could be
empirically differentiated from the motivational-emotional constructs. Thus, an affective
perceptual sensitivity scale (awareness of affect associated with low intensity stimuli) was also
included, and the general construct was labeled cognitive sensitivity. Altogether, cognitive
sensitivity included scales for external perceptual sensitivity, internal perceptual sensitivity,
associative sensitivity, and affective perceptual sensitivity. Table 1 shows the scales based on
sub-constructs associated with each of the four general motivational-emotional and attentional
constructs discussed above, and Appendix A lists scale definitions with sample items for the
scales in both studies. We used factor analysis to investigate a possible hierarchical structure of
these scales.
A second goal of this research was to explore relations betwen temperament and the Big
Five of personality. In recent years, considerable research on personality has supported a fivefactor model. Common labels for the five factors with their corresponding factor numerals are
extraversion (I), agreeableness (II), conscientiousness (III), neuroticism versus emotional
stability (IV), and (V) intellect or imagination, or openness in the Five Factor Model (McCrae,
Temperament and the Big Five
9
1993-1994). The label Big Five is generally reserved for measures derived from analysis of the
trait-descriptive lexicon, whereas McCrae and Costa’s approach is referred to as the Five Factor
Model (FFM). The Big Five and FFM structures are highly convergent (see McCrae & John,
1992), and we refer to the structure as the Big Five/FFM. Appendix B shows the labels for the
Big Five domains along with the 40 Mini-Markers (i.e., eight trait-descriptors as markers for
each of the five domains; see Saucier, 1994) used in these studies.
Previous research has linked measures of temperament with the Big Five/FFM. For
example,using adult subjects, Angleitner and Ostendorf (1994) evaluated the Big Five/FFM in
relation to several temperament questionnaires, including a version of the Strelau Temperament
Inventory (Strelau, Angleitner, Bantelmann, & Ruch, 1990), the EASI-III Temperament Survey
(Buss & Plomin, 1975), the Sensation-Seeking Scale (Zuckerman, Eysenck, & Eysenck, 1978),
and the Dimensions of Temperament Survey (Windle & Lerner, 1986). The authors report fiveand six-factor solutions that are similar to the Big Five/FFM as evidenced by loadings of scores
on the Big Five/FFM measures. The six-factor solution included an additional rhythmicity
factor.
As the authors note, however, including the Big Five/FFM domain was likely to bias the
factor structure in favor of the Big Five/FFM. They therefore included a factor analysis without
the Big Five/FFM measures. A five-factor solution of the temperament scales included a similar
pattern of loadings for these scales as in the six-factor solution, except that content associated
with agreeableness and neuroticism in the other solutions loaded on the same factor. The factor
scores from this solution were then correlated with factor scores from the six-factor solution that
included analysis of the Big Five/FFM measures. These correlations were high, but consistent
with only four factors in the temperament analyses as similar to the Big Five/FFM. Angleitner
and Ostendorf (1994) did not, however, report correlations between measures defined
Temperament and the Big Five 10
exclusively as temperament and measures exclusively defined as Big Five/FFM measures. They
also aggregated multiple measures of temperament, rather than including well-defined
temperament constructs. One of the goals of the current research is to investigate the relations
between specifically defined theoretical constructs and the Big Five.
The studies included in the current report were thus motivated by the desire to explore the
hierarchical relations among lower level constructs of temperament, and to relate temperament
constructs to the lexical Big Five model of personality. To accomplish these goals, we included
older PRQ scales, modified others, and added new temperament scales. Factor analysis was used
to investigate hierarchical relations among temperament variables. Factor scores derived from
this model were then compared to scores from Saucier’s (1994) Mini-Marker measure of the Big
Five.
Method
Participants
A sample of 210 University of Oregon undergraduate psychology students filled out the
first version of the Adult Temperament Questionnaire (ATQ-1) and the Mini-Marker measure of
the Big Five. Subjects received credit toward their psychology courses for participating. Three
subjects who did not complete the questionnaire were excluded. Thus, 207 subjects’ responses
were analyzed. Gender information was not collected for the first study.
Adult Temperament Questionnaire (Version 1)
Each version of the ATQ used a Likert-scale ranging from one to seven, with a response
of one indicating the item does not describe the person, and seven indicating an item to be highly
descriptive of the person. A middle response of four on the scale indicated the subject believed
the item to be neither descriptive nor non-descriptive. Items from the same scale were not placed
Temperament and the Big Five 11
adjacent to each other. The first version of the Adult Temperament Questionnaire (ATQ-1)
included 16 temperament scales, each composed of 11 to 16 items.
The ATQ-1 contained 214 items, with 77 items taken from the PRQ. Six of the scales
were virtually identical to the PRQ (frustration, discomfort, attentional shifting, attentional focus,
internal perceptual sensitivity, and external perceptual sensitivity), three were substantially
modified (fear, sadness, high intensity pleasure), and seven were new (sociability, positive affect,
activity level, attentional shifting from reward, attentional shifting from punishment, affective
perceptual sensitivity, and associative sensitivity). Appendix A gives scale definitions along
with sample items.
Big Five Mini-Markers
Saucier's (1994) “Mini-Markers” (i.e., 40 trait-descriptors) were used to measure the
lexical Big Five. Eight trait-descriptors were used to represent each of the five domains. The
“Mini-Markers” have acceptable reliability and are derived from and thus highly correlated with
Goldberg’s (1992) set of 100 markers that measure the domains of the Big Five. Saucier and
Goldberg (2003) have developed alternative marker sets that they believe have improved upon
both Goldberg’s 100 marker set and Saucier’s Mini-Marker measures of the Big Five. However,
at the time data was collected, these alternative measures were not available. Appendix B
displays the Mini-Marker items. Using a Likert-scale ranging from one to nine, subjects rated
themselves on trait-descriptive words. The 40 trait-adjectives were presented in alphabetical
order.
Results
Reliability coefficients were computed for each scale followed by exploratory factor
analysis (principal axis) of the scale scores. Oblique rotations (promax) were used to allow
exploration of relations among factors. The resulting factor scores were then related to the
Temperament and the Big Five 12
measures of the Big Five domains. One item from each of three scales was removed to improve
reliability. Final alpha coefficients of reliability ranged from .61 to .84. Alpha coefficients of
reliability, means, and standard deviations of temperament and Big Five scales are reported in
Table 2.
Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were both
used for modeling this data. For all factor analyses reported in this paper, SPSS was used to
perform the EFA, and the structural equation modeling software AMOS (a supplement to the
SPSS program) was used to conduct the CFA. We extracted two-, three-, four-, and five- factor
EFA solutions. Since there were four general constructs, the four-factor solution was of
particular interest. A four-factor maximum likelihood CFA was then performed to see how well
the data fit the model, and to compare the fit of alternative models. Chi-square, Joreskog and
Sorbom’s Goodness of Fit Index (GFI), the Adjusted Goodness of Fit Index(AGFI,), and the
Root Mean Square Error of Approximation (RMSEA) fit indices will be reported. The ChiSquare is rarely informative for these types of models (i.e., virtually all models are rejected), but
is by convention usually reported. Fits of .90 are generally considered good for the GFI (Bollen,
1989) The AGFI is generally lower than the GFI, and rewards simpler models with greater
degrees of freedom. The RMSEA can be thought of as a badness of fit index in that higher
numbers indicate less fit. The RMSEA does not penalize for model complexity, and the fit is
always lower for simpler models. Browne and Cudeck (1993) report that an RMSEA greater
than .10 is generally not considered to be a good model.
Four-Factor Model of Temperament
Four factors were extracted based on the four general constructs explored, and the
eigenvalue greater than one criterion, which suggested a four-factor structure. The pattern
matrix for this four-factor solution is shown in Table 3. This model corresponded strongly to the
Temperament and the Big Five 13
initial four general constructs. Factor I, labeled Effortful Attention, included loadings from
attentional shifting from reward, attentional shifting from distress, attentional focusing, and
attentional shifting. Factor II, labeled Cognitive Sensitivity, included loadings from internal
perceptual sensitivity, affective perceptual sensitivity, external perceptual sensitivity, associative
sensitivity, and a small secondary loading from sadness. Factor III, labeled
Extraversion/Surgency, included loadings from high intensity pleasure, sociability, positive
affect, and activity level. Factor IV, labeled negative emotionality, included loadings from fear,
discomfort, sadness, and frustration.
Correlations among these factors showed extraversion/surgency to be positively
correlated with cognitive sensitivity (r = .56), and negative affect negatively correlated with
effortful attention (r = -.50). There was also a modest negative correlation between negative
affect and extraversion/surgency (r = -.26). Other correlations were close to zero.
The CFA solution is reported in Table 4. The initial solution resulted in frustration having
the lowest loading (.44 on negative affect). This was also true with the EFA in Table 3. All
other loadings were at least .53. The Chi-Square for this model with 98 degrees of freedom was
326.89 (p = .000), suggesting rejection of the model. However, this statistic is not usually
informative for these types of models (i.e., virtually all models would be rejected). The GFI was
.84, AGFI was .77, and RMSEA was .11. These fits are considered less than optimal by some
structural equation modeling standards (GFI > .90 and RMSEA <.10). This model is based on
the assumption that all secondary loadings are zero. After reviewing the modification indices,
and adding sadness, a double-loading scale, as an indicator of cognitive sensitivity, an improved
GFI of .87, AGFI of .81, and an RMSEA of .09 were found (No Table reported for this solution).
This model looks very much like the exploratory factor pattern in Table 3.
Temperament and the Big Five 14
Successive iterations of models based on modification indices resulted in assigning the
following additional parameters: attentional shifting from reward as an indicator of
extraversion/surgency, internal perceptual sensitivity as an indicator of negative affect,
attentional shifting as an indicator of cognitive sensitivity, and activity level as an indicator of
effortful control (Table not reported). All of the model modifications were secondary loadings,
and all were modest with the exception of a .50 loading of sadness onto cognitive sensitivity.
The loadings associated with the initial model are substantially higher than loadings derived
from adapting the model. The fit of this model had an improved GFI of .90, an AGFI of .85, and
an RMSEA of .07, thus indicating the model fits the data reasonably well.
When each general construct’s scales were averaged to form a general construct level
scale, correlating with identically labeled factor scores at .93 or higher (i.e., effortful control =
.99, cognitive sensitivity = .98, positive emotionality = .96, and effortful attention = .93).
Therefore, it is clear that the factors emerging from EFA are highly convergent with our initial
general construct scales.
Comparison of the Four-Factor Models with Alternative Models
The two-factor EFA structure included highest loadings from extraversion/surgency and
cognitive sensitivity scales on the first factor, and highest loadings from effortful attention and
negative affect (negative loadings) scales on the second factor. A CFA model based on
assigning this pattern of loadings resulted in a GFI of .74, an AGFI of .63, and an RMSEA of
.14. Although this model did not fit as well as our initial four-factor model without
modifications (GFI = .84 and RMSEA of .11), the model is interesting in that it shows coherent
links between motivational-emotional and cognitive-attentional constructs.
The three-factor EFA model resulted in the negative affect scales scattering across the
other three factors. A CFA model based on this pattern of loadings had a GFI of .76, AGFI of
Temperament and the Big Five 15
.68, and RMSEA of .13. This model clearly did not fit the data as well as the four-factor model.
The five-factor solution resulted in a frustration and activity scales loading onto a fifth factor,
whereas the other highest loadings were consistent with the initial four-factor structure. The
CFA based on this model had a GFI of .84, an AGFI of .77, and an RMSEA of .10. In
subsequent analyses, however, we have used the four-factor model. Although this model fit
about as well as the hypothesized four-factor model, it was basically the same as the four-factor
model, and the fifth factor was not informative.
Four Factor ATQ Model in Relation to the Big Five
Table 5 shows correlations between the factor scores derived from the four-factor
temperament solution and the lexical Big Five scales. The factor scores for each general
construct of temperament are related to specific Big Five measures, with these correlations
ranging from .43 to .59. Temperamental extraversion was most highly correlated with Big Five
extraversion (r =.59), and negative affect with neuroticism (r = .53). The cognitive sensitivity
and intellect/openness factors were also positively related (r = .55). Factor scores derived from
the effortful attention factor showed the lowest correlation with a Big Five scale (.43 correlation
with conscientiousness), as well as a correlation with Big Five neuroticism (-.34). Note in the
bottom half of Table 5 that the correlations between general constructor factor scales and the Big
Five scales are virtually the same as the comparison between factor scores and Big Five scales,
thereby indicating that initial development of general constructs converged substantially with the
Big Five independent of conducting factor analysis.
Discussion
The four-factor EFA solution showed substantial convergence with the existence of four
broad constructs, demonstrated by 1) the emergent EFA four-factor solution, 2) the correlation
between the EFA factor scores and the initial construct scales, and 3) the CFA analyses. The
Temperament and the Big Five 16
four-factor structure includes the three factors Rothbart and colleagues (Ahadi et al., 1993;
Putnam et al., 2001; Rothbart et al., 1994; Rothbart et al., 2001) found in toddlers and children
(extraversion/surgency, negative affectivity, and effortful control). In addition, refinement of the
central reactivity scales included in the PRQ resulted in a clear cognitive sensitivity factor. The
modest secondary loading from sadness on cognitive sensitivity in the EFA and the modified
CFA solutions replicated the three- and six-factor structures of the PRQ study, where sadness
loaded on both a negative affect factor and a factor including perceptual sensitivities and
cognitive reactivity scales.
Factor scores from the four-factor model were substantially correlated with four of the
Big Five domains, and this was true when comparing the Big Five to either factor scores or
scales derived from initial construction independent of factor analysis. Ahadi and Rothbart’s
(1994) proposed relations between temperament and Big Five factors were supported, with
extraversion/surgency aligned with extraversion, negative affectivity with neuroticism, and
effortful control with conscientiousness. In addition, the cognitive sensitivity factor scores were
highly correlated with intellect/openness and moderately correlated with extraversion. This
finding linking cognitive sensitivity to intellect/openness suggests that cognitive sensitivity (i.e.,
awareness of peripheral stimuli and remoteness of associations) is related to insight, reflection,
and imagination. The richly textured, abstract trait-descriptors related to intellect/openness (e.g.,
intellectual, deep, creative, imaginative, and philosophical) thus may involve a common
cognitive-attentional substrate.
Correlations among the four factors are also consistent with a two-factor solution, with a
high positive correlation between extraversion and cognitive sensitivity corresponding to the first
factor, and a high negative correlation between negative affect and effortful control
corresponding to the second factor. The two-factor solution and correlations among the four-
Temperament and the Big Five 17
factors supporting this two-factor structure, are also consistent with the pattern of loadings on
Digman’s (1997) higher order two-factor model derived from factor analysis of Big Five-related
scales. Neuroticism (a correlate of negative affect) and conscientiousness (a correlate of effortful
control) loaded onto one factor (along with agreeableness; agreeableness-affiliativeness content
will be addressed in our Study Two), and extraversion (related to temperamental
extraversion/surgency) and intellect/openness (related to cognitive sensitivity) loaded onto the
other factor.
An extensive literature is consistent with the high negative correlation between negative
affect and effortful attention (e.g., Derryberry & Rothbart, 1988; Eysenck & Eysenck, 1985;
Rothbart et al., 2001). For example, studies using the emotional Stroop task have repeatedly
suggested that negative semantic information interferes with executive attentional processing
(Dawkins & Furnham, 1989; MacLeod & Hagan, 1992; Mogg, Bradley, & Williams, 1995;
Myers & McKenna, 1996; Pratto & John, 1991). In addition, the Diagnostic and Statistical
Manual of Mental Disorders: Fourth Edition (American Psychiatric Association, 1994) lists
attentional difficulties as indicators of anxiety and depressive disorders.
Other research is consistent with the high positive correlation between
extraversion/surgency and cognitive sensitivity. Induction of positive affect results in more
inclusive categorization of words and colors (Isen & Daubman, 1984), and positive affect has
also been associated with generating more unusual and diverse word associations (Isen, Johnson,
Mertz, & Robinson, 1985). In general, a more creative and broad ranging cognitive style is
produced when positive affect is induced (Ashby, Isen, & Turken, 1999). Recent factor analyses
of scales comprising the Infant Behavior Questionnaire – Revised (IBQ-R) also resulted in a first
factor with loadings from both perceptual sensitivity and extraversion scales (Gartstein &
Rothbart, 2003).
Temperament and the Big Five 18
Differentiating Aggressive and Non-Aggressive Negative Affect
Because the five-factor EFA, supported by the CFA, suggested an aggressive negative
affect factor, so we were also interested in exploring the differentiation of aggressive and nonaggressive negative affect. At the adult level of self-report, Zuckerman (1997; Zuckerman et al.,
1993) has identified superfactors that discriminate fear-anxiety and anger-aggression related
constructs. However, some models of personality do not make this distinction at a higher level
in the hierarchy. Tellegen and colleagues (e.g., Harkness et al., 1995; Tellegen, 1985; Watson et
al., 1999) have extracted a factor labeled negative affect that includes both aggressive and fearful
components. Costa and McCrae’s (1994) scales comprising neuroticism (e.g., anxiety,
depression, and angry-hostility) are consistent with Tellegen’s (1985) model. Aggression under
the Five Factor Model, however, is related to both neuroticism and the negative pole of
agreeableness (e.g., Costa & McCrae, 1994). Aggression items are also included at the negative
pole of Cloninger’s superfactor labeled cooperation (see Cloninger, Przybeck, Svrakic, &
Wetzel, 1994). In Study One, we included only one anger-aggression related scale (frustration),
so that a common factor for anger-aggression was not possible. The ambiguous status of angeraggression in relation to non-aggressive negative motivated us to empirically explore the
differentiation of these constructs.
Studying Affiliativeness and Aggressiveness as Basic Dimensions of Temperament
In addition to further exploring aggressiveness scales, we were also interested in adding
scales for affiliativeness, thus yielding six general constructs of interest: cognitive sensitivity,
effortful control, extraversion/surgency, non-aggressive negative affect, aggressive negative
affect, and affiliativeness.
Rothbart and Derryberry (2002; Derryberry & Rothbart, 1997) have discussed six general
constructs as core psychobiological dimensions of temperament, and suggest that the six
Temperament and the Big Five 19
temperament dimensions are related to the Big Five. At another level of analysis, Saucier (2001)
has speculated that a lexically derived seven-factor model might be the optimal representation of
at least some indigenous language structures originating outside of northern Europe. Saucier
notes the strong similarity between seven-factor solutions derived from natural language traitdescriptions in Hebrew and Filipino (languages that are not closely related). Saucier’s model
also differentiates anger- and fear-related constructs at the factor level, and includes a general
negative valence factor. Six of these seven factors are thus conceptually similar to the core
psychobiological constructs of temperament described by Rothbart and Derryberry.
The inclusion of affiliativeness as a dimension of temperament deserves special attention.
Rothbart and colleagues (Derryberry & Rothbart, 1997; Rothbart, 1994; Rothbart, Ahadi, &
Evans, 2000; Rothbart & Bates, 1998) have proposed affiliativeness as a dimension of
temperament, although to date, it has been explored only in revisions of the Infant Behavior
Questionnaire (Gartstein & Rothbart, 2003) and the Early Adolescent Temperament
Questionnaire (Ellis & Rothbart, 2001). A few decades ago, behavioral scientists did not have
evidence for affiliative brain systems (Panksepp, 1998). Social bonds were thought to emerge
from reinforcement contingencies related to organisms receiving food, warmth, and protection
from attachment figures. Based upon animal research findings, however, Panksepp (1998) has
posited brain systems related to affiliativeness. He notes that an understanding of specialized
neural substrates for social attachment and separation distress is beginning to emerge, and that
these two systems can be viewed as evolutionarily and neurochemically related. In addition to
Panksepp’s psychobiological modeling of affiliation, Cloninger (1998; Svrakic, Svrakic, &
Cloninger, 1996) includes a reward dependency construct as a dimension of temperament based
upon the need for close relationships with others.
Temperament and the Big Five 20
Study Two
In Study Two, scales for six general constructs were included (see Appendix A for scale
definitions and sample items). These constructs included non-aggressive negative affect (fear,
sadness, and discomfort), anger-aggressive negative affect (frustration, social anger, and
aggression control), affiliativeness (emotional empathy, social closeness, and empathic guilt),
extraversion/surgency (sociability, positive affect, and high intensity pleasure), cognitive
sensitivity (affective perceptual sensitivity, general perceptual sensitivity, and associative
sensitivity), and effortful control (effortful attention, activation control, and inhibitory control).
Table 6 displays the scales associated with each of the six general constructs, and Appendix A
lists scale definitions with sample items for the scales in both studies.
Aggressive and Non-Aggressive Negative Affect
In addition to the frustration scale from the original PRQ, the aggressive negative affect
construct included scales assessing social anger (hostility felt toward other people) and
aggression control (capacity to inhibit the behavioral expression of anger). In Study One,
negative affect was uncorrelated with agreeableness. Ahadi and Rothbart (1994) nevertheless
speculated that negative affect is negatively correlated with agreeableness. A possible reason for
the absence of this finding in Study One was the lack of content assessing social anger. Thus a
social anger scale was included in the present study. The components of non-aggressive
negative affect included fear, sadness, and discomfort.
Affiliativeness
Affiliative scales assessing emotional empathy (affective response congruent with the
feelings of others), empathic guilt (distress in response to self negatively affecting other people),
and social closeness (feelings of warmth, closeness, interest and involvement with others) were
developed. The goal in designing these scales was for each scale to be substantially
Temperament and the Big Five 21
differentiated from the others while also sharing a higher-level definition (positive concern for
others). Our goal was to explore whether these measures empirically clustered together
suggesting a more general affiliativeness dimension.
Extraversion, Effortful Control, and Cognitive sensitivity
Extraversion included three scales that were also used in Study One: sociability, positive
affect, and high intensity pleasure. The effortful control construct was differentiated into
effortful attention and two behavioral control scales used in the PRQ: inhibitory control (ability
to inhibit inappropriate behavior) and activation control (capacity to perform an action when
there is a strong tendency to avoid it). The four effortful attention scales in Study One
(attentional shifting, focusing, shifting from reward, and shifting from punishment) were highly
intercorrelated, so their content was collapsed into a single effortful attention scale in Study Two.
We believed that including the PRQ behavioral control scales with effortful attention would
extend the breadth of the construct.
Cognitive sensitivity included scales assessing general perceptual sensitivity, affective
perceptual sensitivity, and associative sensitivity. These scales were taken from Study One,
except that external and internal perceptual sensitivity were collapsed into a single general
perceptual sensitivity scale (these scales were conceptually similar and highly correlated).
Method
A total of 258 undergraduate psychology students (150 women and 108 men) filled out
the revised version of the Adult Temperament Questionnaire (ATQ-2), followed by completion
of the Big Five Mini-Markers. English was the first language for 95% of the subjects and 91.5%
identified the United States as their country of origin. The ATQ-2 included the changes from
ATQ-1 described above, with a total of 253 items. Administration of the questionnaires was
Temperament and the Big Five 22
identical to that in Study One. Subjects received participation credit toward their undergraduate
psychology courses.
Results
Table 7 reports the means, standard deviations, and reliabilities for the temperament
scales. A total of 3 items from the 253 original items in ATQ-2 were removed to improve
reliability. After removal of these 3 items, the reliability as assessed by coefficient alpha for 13
of 18 of the temperament scales reached a level of .80 or higher, and only one scale was lower
than .70 (inhibitory control at .66).
The initial EFA and CFA approach to analyzing this data set was the same as in Study
One. First an EFA was performed on all of the scales. A six-factor solution was extracted to
examine its correspondence to the initially developed six constructs, but some of the factors
deviated from the six constructs. Aggressive negative affect, cognitive sensitivity, and effortful
control factors were consistent with initial construction, but the other scales were distributed
across the other three factors in a less interpretable manner. Exploring the hypothesized sixfactor solution at the CFA level of analysis resulted in a solution that was not mathematically
acceptable (i.e., not positive definite).
The EFA of the five-factor solution was consistent with the initial construction, except
that the aggressive and non-aggressive negative affect scales loaded on the same general
negative affect factor. The pattern matrix for the five-factor solution is reported in Table 8. The
first factor (negative affect) included loadings from scales assessing facets of both aggressive
negative affect (frustration, aggression control, social anger) and non-aggressive negative affect
(fear, discomfort, and sadness). Effortful attention and inhibitory control also loaded modestly
and negatively on the first factor. Factor II included loadings from scales of cognitive sensitivity
(affective perceptual sensitivity, general perceptual sensitivity, and associative sensitivity) along
Temperament and the Big Five 23
with small positive secondary loadings from sadness and high intensity pleasure. Factor III
included highest loadings from scales derived from the extraversion construct (sociability, high
intensity pleasure, and positive affect) along with moderate loadings from social closeness and
inhibitory control (negative loading), and modest secondary loadings from discomfort (negative
loading) and emotional empathy. Factor IV included loadings from affiliativeness scales
(emotional empathy, empathic guilt, and social closeness) along with a moderate secondary
loading from aggression control and even smaller loadings from social anger (negative loading),
fear, and positive affect. Factor V, the effortful control factor, included highest loadings from
activation control and effortful attention along with a moderate loading from inhibitory control.
Although inhibitory control is a sub-construct of effortful control, it loaded relatively equally
onto effortful control, extraversion, and negative affect.
The negative correlation between the negative affect and effortful control factors was
replicated (r = -.50). The positive correlation between cognitive sensitivity and extraversion
factors was also replicated, but the correlation was only .26 in this study. Both cognitive
sensitivity and extraversion were modestly correlated with the affiliativeness factor (r = .28 and r
= .20, respectively), and other correlations were close to zero.
The five-factor CFA solution is reported in Table 9. The Chi-Square with 125 degrees of
freedom was 588.40 (p = .000). The GFI was .78, an AGFI of .70 and an RMSEA of .12. This
model did not fit as did the four-factor model from Study One. However, the EFA five-factor
model is still similar to initial construction, except that aggressive and nonaggressive negative
emotionalities were collapsed into a single scale, and there were a number of secondary loadings.
However, it was the best fitting model. As noted, the six-factor construct-driven CFA model was
not admissible. This was also true of the two-, three-, and four- factor CFA models derived form
EFA analyses. Using the modification indices to improve the fit of the model in Study One
Temperament and the Big Five 24
required very few modifications, and each of these modifications involved modest secondary
loadings, with the exception of one moderate secondary loading. However, in the present study
it would require virtually reconstructing the EFA model by freeing the majority of parameters.
Although the alternative five-factor model emerging from EFA did not fit the data
particularly well with the CFA, the factor scores from this five-factor EFA model did
demonstrate strong one-to-one correlations with the five constructs, after collapsing aggressive
and nonaggressive negative emotionality into the same construct, (affiliativeness, r = .90;
positive emotionality, r = .93; negative emotionality, r = .97; effortful control, r = .97, and
cognitive sensitivity, r = .99), demonstrating strong convergence between initial construction and
the factors emerging from EFA.
Correlations between Temperament and the Big Five
Table 10 shows the correlations between the EFA temperament factor scores and the Big
Five scales. The correlations indicate robust levels of one-to-one correspondence. The five
correlations suggesting convergence with the Big Five range from .64 to .74. Only one
additional correlation exceeded .35, the negative correlation between effortful control and
neuroticism (r = -.41). The negative affect factor scores were highly correlated with Big Five
neuroticism (r = .74), cognitive sensitivity factor scores with Big Five intellect/openness (r =
.65), temperamental extraversion factor scores with Big Five extraversion (r = .67), and the
affiliativeness factor scores with Big Five agreeableness (r = .69). The effortful control factor
scores were highly correlated with Big Five conscientiousness (r = .64), while also having a
substantial negative correlation (r = -.41) with Big Five neuroticism. Note that the temperament
scale scores derived from initial construction independent of factor analysis show the same
pattern and strength of correlation with the Big Five (correlations along the main diagonal
ranging from .60 to .70).
Temperament and the Big Five 25
Discussion
Study Two demonstrated the emergence of five broad factors of temperament composed
of negative affect, cognitive sensitivity, extraversion/surgency, affiliativeness, and effortful
control. These factors emerged from the EFA as an alternative to the hypothesized six-factor
solution, and the CFA showed modest support for this model. Although the fit indices were not
as high as Study One, it was the only model that fit the data. In addition, the factor scores from
the EFA factor model were very highly convergent with initial construction, except that
aggressive and non-aggressive negative affect constructs were collapsed into a single negative
affect construct. Contrary to expectations, aggressive and non-aggressive negative affect scales
loaded together on the same factor.
Unlike Study One, however, there were several instances of secondary loadings.
Although several of the scales in Study Two differed from those used in Study One, the
conceptual nature of the four factors from Study One was replicated, and the affiliativeness
construct introduced in Study One resulted in a fifth factor labeled affiliativeness. The factors
showed substantial levels of convergence with the Big Five, and although the CFA indices did
not suggest exceptional fit, scales derived from initial construction of factor breadth scales were
highly correlated with the EFA factor scores (i.e., r  .90).
High correlations were observed between temperament and Big Five, suggesting
substantial convergence between five-factor models. We suggest that temperamental processes
might be at the core of Big Five domains that were derived without theoretical assumptions.
Both the EFA factor scores and general scale scores derived independent of factor analysis
showed strong one-to-one correlations with the Big Five.
The high negative correlation between negative affect and effortful control was
replicated, and the positive correlation between extraversion/surgency and cognitive sensitivity
Temperament and the Big Five 26
to a lesser degree. However, both the negative association between effortful control and negative
affect and the positive association between cognitive sensitivity and extraversion/surgency were
supported by the two-factor EFA structure, as was the case in Study One. As noted in Study
One, these results may be related to experimental findings differentially linking positive and
negative emotions with aspects of cognitive-attentional processing. The introduction of
affiliativeness content also resulted in an affiliativeness factor showing a modest positive
correlation with both cognitive sensitivity and extraversion. In this study, inhibitory control
loaded relatively equally across effortful control, extraversion (negatively), and negative affect
(negatively) factors. Social closeness also loaded relatively equally on both affiliativeness and
extraversion factors, and there were also several other modest secondary loadings. We now
consider these findings.
Inhibitory Control
All of the inhibitory control items explicitly refer to the capacity to inhibit behavior
involving positive affect and approach. Therefore, although the scale fits the definition of
effortful control, its negative loading on an extraversion/surgency factor (defined as positive
affect and approach; see Derryberry and Rothbart, 1997) is, after the fact, not surprising. People
who have stronger approach tendencies are more likely to find it difficult to inhibit approach
tendencies. In addition, the negative loading of inhibitory control on a negative affect factor is
consistent with other findings showing negative affect and effortful control to be negatively
related (Ahadi et al., 1993; Derryberry & Rothbart, 1988; Ellis & Rothbart, 2001).
Affiliativeness
Scales assessing social closeness, emotional empathy, and empathic guilt clustered
together forming an affiliativeness factor in Study Two. However, social closeness loaded
equally on both affiliativeness and extraversion factors. This finding may be related to one of
Temperament and the Big Five 27
our initial motivations for exploring affiliativeness as a dimension of temperament, namely that
affiliation is often viewed as a facet of extraversion (e.g., Depue & Collins, 1999). Sociability
(enjoyment from interacting and being in the presence of others) and social closeness (interest,
warmth, and involvement with others) were highly correlated (r = .58). However, sociability
loaded exclusively on the extraversion factor, whereas social closeness loaded equally on both
affiliativeness and extraversion. This pattern of results suggests that having close relations with
others is linked to both extraversion/surgency and affiliativeness, whereas the experience of
sociability (enjoyment when among people) appears to be more purely a component of
extraversion/surgency, perhaps through mechanisms of reward reactivity. In a study across 39
countries, Lucas, Diener, Grob, Suh, and Shao (2000) found support for reward sensitivity rather
than sociability as basic to defining extraversion.
Although emotional empathy loaded highest on the affiliativeness factor, this scale also
loaded modestly on the extraversion factor, even though the majority of emotional empathy
items were concerned with negative affect. The scale did not load on the negative affect factor.
Emotional empathy was also empirically differentiated from the conceptually similar affective
perceptual sensitivity scale. Emotional empathy is defined as experiencing emotions congruent
in valence with others, whereas affective perceptual sensitivity items involve awareness of affect,
including many items that involve awareness of incongruency between affect and social
presentation. The distinction between emotional empathy and affective perceptual sensitivity (or
“cognitive empathy”) is supported by emotional empathy’s loading on a motivational-emotional
factor (affiliativeness) and affective perceptual sensitivity’s loading on a cognitive-attentional
factor (cognitive sensitivity).
The affiliativeness factor also included an interesting pattern of loadings from negative
affect scales. Aggression control, which loaded highly on negative affect (negative loading)
Temperament and the Big Five 28
loaded moderately on affiliativeness, suggesting that the capacity to inhibit aggression increases
with higher levels of affiliativeness. Social anger, loading highest on the negative affect factor,
also had a modest negative loading on affiliativeness. Finally, the modest positive loading of
fear on affiliativeness suggests differential associations between fearful and aggressive negative
affects in relation to affiliativeness.
As in Study One, the highest loading of discomfort was on negative affect, but in this
study there was also a noteworthy negative loading from discomfort on extraversion. High
intensity pleasure and discomfort are quite similar in their definitions, except that the former
refers to positive affect in response to intense and complex stimuli, whereas the latter refers to
negative emotions in response to sensory stimulation. Eysenck and Zuckerman (see Zuckerman,
1997) were both influenced by constructs involving optimal level of stimulation in their early
theorizing related to extraversion. In this study, there is thus modest evidence linking
extraversion to optimal level of stimulation. Also, as in both the PRQ study and Study One,
cognitive sensitivity had a modest loading from sadness.
Rationale for Study Three
Our analyses showed support for a five-factor model at both the EFA and CFA levels of
analyses, with aggressive and non-aggressive negative affect sub-constructs collapsing into a
single negative affect factor, suggesting that the factor structure found in Study One is
representative. However, since this model was generated post hoc, we explored confirmation of
this model in another sample.
In addition, although the model was supported by EFA and CFA at moderate levels, some
of these scales loading on more than one factor lessened overall fit. These deviations had
interesting post hoc explanations discussed above. With respect to model refinement and further
exploration, it also makes sense to delete these multiple loading scales from the model, so that
Temperament and the Big Five 29
each scale is more exclusively representative of a single general construct. Therefore, we
explored the model further after deleting these multiple loading scales. Finally, our data in
Studies 1 and 2 was limited to college students of approximately the same age, cohort, and level
of education. Study Three data was collected from a larger community sample that was
substantially older and more diversified.
Study Three
The primary goal for Study Three was to explore replication of Study Two findings in a
large community sample. A special version of the ATQ was constructed for participants in the
Eugene-Springfield Community Sample. The goals for constructing this measure included: 1)
using primarily the same constructs as Study Two; 2) constructing a short form; 3) eliminating
scales that loaded substantially on more than one factor; and 4) developing a measure that could
be used to further explore relations between temperament and the five-factor model. The above
goals were also considered with respect to developing a measure that could be administered in a
short period of time as a part of the biannual collection of data in the Eugene-Springfield
Community Sample (see Goldberg, 2003). This 100-item version of the ATQ is a subset of
items and scales from Study Two.
The structure of the items was also changed to fit the typical format used by Goldberg
(2003) with the Eugene-Springfield Community sample. Pronouns and other unnecessary words
were omitted from items. For example, the fear item, “I become easily frightened.” was changed
to “Become easily frightened.” However, the meanings of the items were not changed, and it is
not likely that the changes in item structure would have influenced responding significantly.
Table 11 displays the five general construct scales and their corresponding scales.
Empirical Refinement of the Model
Temperament and the Big Five 30
In Study One, the EFA and the CFA supported the initially constructed four-factor
model. In Study Two, the EFA and the CFA showed substantive convergence with initial
construction, but with notably more deviations from the model as well. One way to explore
improving the fit of at the confirmatory level of analysis is to remove double loading scales, and
in the case of the six-factor EFA, model loadings that were more purely deviant. If such a model
fit the data better, this would be an improved measure showing coherency within each domain as
well as differentiation from other factors/constructs.
The following scales were removed: Inhibitory control loaded on three factors, and social
closeness equally on two factors. Discomfort had a secondary negative loading on
extraversion/surgency, and aggression control and social anger (negative loading) both had
secondary negative loadings on affiliativeness. Moreover, the exploration of a separable
aggressive and non-aggressive negative emotion constructs was no longer of central concern
based on the five factor findings. We now present confirmation of this model in the community
sample.
Table 12 shows the ATQ general constructs and associated sub-constructs assessed in this
study. No new scales were included, but aggression control, social anger, discomfort, inhibitory
control, and social closeness were omitted, permitting development of a much shorter ATQ form.
Costa & McCrae’s NEO-PI-R (1992b; McCrae & Costa, 1996; McCrae et al., 1996) was
also included. The NEO-PI-R includes is a 240 item questionnaire measure of the Big Five/FFM
domains, and each domain of the NEO-PI-R domains includes six facet scales that each include
eight items. Costa & McCrae’s (1992b) NEO-PI-R measure of the Big Five/FFM domains had
been completed by this community sample about eight years prior, in 1994.
Temperament and the Big Five 31
Method
Participants and Measures
A total of 700 participants from a Eugene-Springfield, Oregon community sample
completed this special 100-item version of the ATQ. The majority of these subjects have also
completed a large number of questionnaires during the past decade, including questionnaires
associated with the prominent models of personality discussed in the introduction to this chapter.
The Eugene-Springfield Community Sample is managed by Goldberg (2003), and originally
included 1,062 participants (700 of whom completed this version of the ATQ). Participants were
recruited by mail solicitation in 1993 from lists of local homeowners. Data was collected
through the mail. Age and gender data was known for 693 of the 700 people. Participants
included 296 men, 397 women, and seven of unknown gender, and ranged in age from 26 to 91
years with a median of 57 and a mean of 58.7 years. Only 30 participants were younger than 40
years of age.
Special Short Form of ATQ
A 100-item ATQ questionnaire was adapted from the version of the ATQ used in Study
Two, as described previously. Scales included 6 to 8 items each. To be consistent with other
questionnaires completed by the community sample, this version of the ATQ used a 5-point
Likert-scale instead of the previously used 7-point scale. This version of the ATQ was
completed in 2002, and included two additional scales that were not of interest in this study.
Results
Table 12 reports the means, standard deviations, alpha reliability coefficients, and
number of items for each of the scales. Table 13 reports the EFA of the 13 scales. Note that all
the loadings from scales load highest onto factors consistent with the initial construction, with
only one noteworthy secondary loading. In addition to loading highest onto
Temperament and the Big Five 32
extraversion/surgency, positive affect also loaded negatively on negative affect. The CFA
loadings are reported in Table 14. The Chi-Square for this model with 55 degrees of freedom
was 359.86 (p = .000). The GFI was .93, and the AGFI was .88, substantially higher than in
Studies One and Two, even after modifying the model. In addition the RMSEA was in the
acceptable range at .09. The EFA and CFA results replicate the four-factor structure from Study
One, as well as the five-factor structure from Study Two. This replication was from a larger and
more diversified community sample, suggesting that replication generalizes beyond the limited
samples in Studies One and Two.
Table 15 reports the correlations between temperament factor and scale scores in relation
to the general domain or factor breadth scales of the NEO-PI. Note that factor scores derived
from the EFA show one to one correspondence, with correlations of .59 or higher, except for the
affiliativeness and agreeableness correlation, substantially lower at .49. The correlations
between temperament scale scores and the NEO-PI-R scales were similar, demonstrating that
initial construction maps substantially onto the Big Five/FFM.
General Discussion
In Study One, four-factor EFA and CFA models resulted in substantial convergence
between initial construction of constructs and empirical structure. Factor scores derived from a
four-factor solution were substantially correlated with four of the five Big Five domains. In
Study Two, the five-factor solution included conceptual replication of the four-factor solution
from Study One, with added affiliativeness content resulting in a fifth factor. Factor scores
derived from this model showed high levels of one-to-one correspondence with the Big Five
scales, with correlations ranging from .64 to .74. General construct measures and Big Five
scales were also related, with correlations ranging from .60 to .70. In Study One, correlations
between Big Five scales and the temperament factor scores were not as high as in Study Two.
Temperament and the Big Five 33
Part of this difference across studies appears to be the result of substantially higher reliabilities in
Study Two.
In Study Three, we explored empirical refinement of the five-factor model by removing
multiple loading scales. This model was confirmed in a large-scale and more diversified
community sample. Again, the temperament factor and general scale scores converged
substantially with the Big Five/FFM.
Convergence between Temperament Model and the Big Five
How might analysis of thousands of trait descriptive words such as “kind,” “intellectual,”
and “organized,” lead to a five-factor model of personality that is also linked to the organization
of temperament? Digman and Shmelyov (1996) suggested that temperament could be subsumed
under the Big Five. An alternative approach, however, is the possibility that the Big Five
structure is shaped by early temperament. According to this view, since language is not
sufficiently developed in infancy and early childhood, temperament in its earliest stages would
not be influenced by explicit words a person used to describe the self and others. Later in
development, explicit self-views (e.g., “I am not fearful.” or “I’m tough.”) will develop,
influenced by and possibly influencing temperament.
McCrae et al. (2000) make an argument for subsuming the Big Five/Five Factor Model
under temperament. They note that behavioral genetics, animal personality, and stability across
development and cultures all support this view. They also suggest that one aspect differentiating
temperament and personality is that temperament researchers tend to emphasize basic processes
such as attention and affect, whereas Big Five and Five Factor Model researchers emphasize
prognostic outcomes, as in using conscientiousness to predict job performance. Personality
researchers also often stress the effects individual differences have on others, especially in the
agreeableness construct (Hogan, 1995).
Temperament and the Big Five 34
In agreement with McCrae et al.’s (2000) analysis, we propose that basic temperament
processes may form the substrate of global personality traits. Thus, effortful control can be seen
as an attentional substrate for conscientiousness and cognitive sensitivity as a substrate for
intellect/openness. Similarly, distress proneness may be central to neuroticism, and a reward and
incentive system to extraversion. Of course, our data are correlational and based upon the
assumption that the temperament scales assess temperamental processes. Additional research is
needed to support or refute this view.
Another reason for giving temperament priority in interpretation is a developmental one.
Many of the temperament constructs explored in this paper are conceptually related to constructs
of temperament assessed early in life (Rothbart & Bates, 1998). If constitutionally based
individual differences provide the core for the developing personality, we would expect them to
be reflected in broad personality descriptions. An important contribution of further research will
be to build a stronger bridge linking child temperament to adult personality. Diener (2000) notes
that developmental researchers of temperament and adult personality have been relatively
isolated from each other. This isolation is attributed to the difficulty of obtaining data linking
child and adult personality. However, the temperament constructs we have found to be
associated with the Big Five in the self-report were developed on the basis of those applicable to
the child. Future longitudinal research will be required to more thoroughly examine these links.
The temperament model developed here supports the validity of the Big Five in new
ways. Several prominent personality theorists have questioned the validity of the Big Five
(Block, 1995; Eysenck, 1991; Tellegen, 1993; Zuckerman, 1992), and Block’s argument
deserves special attention. He notes that the Big Five is derived from factor analysis, and that no
statistical technique should be given the power to decide which concepts will be used for
personality assessment, especially when factor solutions are in many respects arbitrary. He
Temperament and the Big Five 35
suggests that movement of the Big Five bandwagon be halted in favor of identifying its contents.
This research addresses Block’s content concern. To the degree that temperament scales reflect
constitutionally based processes, (and are a priori and rationally constructed rather than
“identified” by factor analysis) their relation with the Big Five helps specify its content.
Block (1995) also questions the centrality of the Big Five for its lack of dynamism,
making the important point that motivational aspects of personality organize thinking and
behavior with respect to goals. Although this research does not empirically examine the intraindividual structure of personality, our model suggests that basic attentional and affectivemotivational processes are related to the Big Five. This understanding allows for greater
modeling of personality dynamics than is possible using only abstract labels like extraversion,
agreeableness, conscientiousness, neuroticism, and intellect/openness. For example, effortful
control includes the capacity to inhibit pre-potent positive (extraversion/surgency) and negative
(negative affect) responses in favor of subdominant response tendencies; cognitive sensitivity
involves noticing peripheral stimuli with emotional relevance (see Derryberry & Rothbart,
1997). Hence, reactive and effortful facets of attentional processing can be mapped dynamically
in relation to the suppression and activation of positive and negative emotionality.
In this paper, we have provided empirical evidence that scales assessing lower-level
constructs demonstrate a structure reflecting higher-level constructs. Further, these general
constructs show empirical convergence with global traits (i.e., the Big Five) developed from a
very different lexical approach. It is hoped that this work will contribute to future research
exploring developmental links between temperament and personality.
Temperament and the Big Five 36
Table 1
Study One General Constructs and Associated Scales
General Constructs
Effortful Control
Extraversion/Surgency
Negative Affect
Cognitive sensitivity
Associated Scales
Attentional Focusing, Attentional Shifting, Attentional Shifting
from Punishment, Attentional Shifting from Reward
Activity Level, High Intensity Pleasure, Positive Affect,
Sociability
Discomfort, Fear, Frustration, Sadness
Affective Perceptual Sensitivity, Associative Sensitivity, External
Perceptual Sensitivity, Internal Perceptual Sensitivity
Temperament and the Big Five 37
Table 2
Means, Standard Deviations, and Alpha Reliability Coefficients of Adult Temperament
Questionnaire (Version 1) and Big Five Scales for Study One
Mean
SD
Alpha
Associative Sensitivity
5.15
.69
.78
Affective Perceptual Sensitivity
5.29
.67
.80
External Perceptual Sensitivity
5.14
.76
.81
Internal Perceptual Sensitivity
4.83
.75
.75
Activity Level
4.71
.68
.65
High Intensity Pleasure
4.95
.70
.74
Positive Affect
4.99
.67
.76
Sociability
5.08
.81
.83
Discomfort
4.16
.67
.61
Fear
3.68
.81
.68
Frustration
4.62
.72
.75
Sadness
4.79
.77
.72
Attentional Focusing
3.80
.92
.85
Attentional Shifting
4.30
.73
.76
Attentional Shifting from Punishment
3.95
.85
.82
Attentional Shifting from Reward
3.64
.71
.73
Scales
ATQ Scales
Temperament and the Big Five 38
Table 2 (continued)
Big Five Scales
Extraversion
5.73
1.42
.87
Agreeableness
7.12
1.01
.81
Conscientiousness
6.23
1.26
.82
Neuroticism
4.91
1.31
.80
Intellect/Openness
6.73
1.02
.78
Note. SD = Standard Deviation
Temperament and the Big Five 39
Table 3
Pattern Matrix for Four-Factor Solution of Adult Temperament Scales for Study One
Factor Loadings
Adult Temperament Scales
EA
RA
E
NA
Attentional Shifting from Reward
.88
-.11
-.12
.24
Attentional Focusing
.80
-.03
.06
-.07
Attentional Shifting from Punishment
.79
.07
.04
-.19
Attentional Shifting
.69
.18
.11
-.12
Internal Perceptual Sensitivity
.02
.81
-.17
.12
Affective Perceptual Sensitivity
-.03
.81
.04
-.04
External Perceptual Sensitivity
.03
.78
-.02
-.09
Associative Sensitivity
-.07
.62
.11
-.03
High Intensity Pleasure
-.09
-.01
.85
-.03
Sociability
-.01
-.05
.77
.10
Positive Affect
.05
-.03
.61
.02
Activity Level
.15
.01
.52
-.10
Fear
-.04
-.07
-.06
.68
Discomfort
.13
-.05
-.09
.66
Sadness
.07
.32
.14
.62
Frustration
-.11
-.02
.12
.37
Note. Abbreviations: EA = Effortful Attention, RA = Cognitive Sensitivity, E = Extraversion,
NA = Negative Affect. Loadings .40 or greater listed in bold print.
Temperament and the Big Five 40
Table 4
CFA Matrix for Initial Four-Factor Solution of Adult Temperament Scales for Study One
Factor Loadings
Adult Temperament Scales
EA
RA
E
NA
Attentional Shifting from Distress
.90
.00
.00
.00
Attentional Focusing
.82
.00
.00
.00
Attentional Shifting
.78
.00
.00
.00
Attentional Shifting from Reward
.68
.00
.00
.00
Affective Perceptual Sensitivity
.00
.82
.00
.00
External Perceptual Sensitivity
.00
.77
.00
.00
Internal Perceptual Sensitivity
.00
.71
.00
.00
Associative Sensitivity
.00
.69
.00
.00
High Intensity Pleasure
.00
.00
.81
.00
Sociability
.00
.00
.71
.00
Positive Affect
.00
.00
.62
.00
Activity Level
.00
.00
.58
.00
Fear
.00
.00
.00
.67
Discomfort
.00
.00
.00
.64
Sadness
.00
.00
.00
.53
Frustration
.00
.00
.00
.44
Note. Abbreviations: EA = Effortful Attention, RA = Cognitive Sensitivity, E = Extraversion,
NA = Negative Affect. Loadings associated with hypothesized model in bold. Modified model
in parenthesis.
Temperament and the Big Five 41
Table 5
Correlations of Temperament Factor Scores with Big Five Scales (Study One)
Big Five Scales
Temperament
C
I/O
E
N
A
Effortful Attention
.43
.21
.06
-.34
-.05
Cognitive sensitivity
.17
.55
.19
.16
.21
Extraversion
.10
.41
.59
-.09
.27
Negative Affect
-.15
-.12
-.20
.53
.04
Effortful Attention
.41
.24
.08
-.33
-.04
Cognitive sensitivity
.15
.57
.16
.18
.15
Extraversion
.12
.31
.64
-.10
.28
Negative Affect
-.05
-.03
-.08
.55
.00
Factor Scores
Scale Scores
Note. Abbreviations: C = Conscientiousness, I/O = Intellect/Openness,
E = Extraversion, N = Neuroticism, A = Agreeableness.
Correlations greater than .30 are printed in bold.
Temperament and the Big Five 42
Table 6
Study Two: General Constructs and Associated Scales
General Constructs
Associated Scales
Affiliativeness
Emotional Empathy, Empathic Guilt, Social Closeness
Aggressive Negative Affect
Aggression Control, Frustration, Social Anger
Effortful Control
Activation Control, Effortful Attention, Inhibitory
Control
Extraversion/Surgency
High Intensity Pleasure, Positive Affect, Sociability
Non-Aggressive Negative Affect
Discomfort, Fear, Sadness
Cognitive Sensitivity
Affective Perceptual Sensitivity, Associative Sensitivity,
General Perceptual Sensitivity
Temperament and the Big Five 43
Table 7
Means, Standard Deviations, and Alpha Reliability Coefficients of Adult Temperament
Questionnaire (Version 2) and Big Five Scales for Study Two
Scales
Mean
SD
Alpha
Frustration
4.23
.82
.80
Social Anger
3.83
.88
.81
Aggression Control
4.92
.91
.84
Sadness
4.39
.86
.80
Discomfort
3.94
.76
.72
Fear
4.27
.87
.76
Effortful Attention
3.70
.95
.88
Inhibitory Control
3.95
.74
.66
Activation Control
4.17
.97
.84
General Perceptual Sensitivity
4.96
.70
.81
Affective Perceptual Sensitivity
5.02
.88
.90
Associative Sensitivity
5.16
.79
.85
Sociability
5.41
.89
.89
High Pleasure
5.10
.73
.77
Positive Affect
5.12
.88
.84
Emotional Empathy
5.08
.71
.75
Empathetic Guilt
5.61
.86
.84
Social Closeness
5.80
.60
.81
ATQ Scales
Temperament and the Big Five 44
Table 7 - continued
Big Five Scales
Neuroticism
4.94
1.21
.82
Intellect/Openness
6.70
1.36
.85
Extraversion
5.68
1.53
.88
Agreeableness
7.23
1.06
.85
Conscientiousness
6.08
1.40
.85
Temperament and the Big Five 45
Table 8
Pattern Matrix for EFA of Adult Temperament Scales for Study Two
Factor Loadings
Adult Temperament Scales
NA
OR
E
Aff
EC
Frustration
.79
-.22
.11
-.01
.11
Aggression Control
-.73
-.04
-.17
.47
-.17
Social Anger
.68
.22
-.13
-.32
.01
Fear
.64
-.04
-.20
.31
-.14
Discomfort
.58
-.03
-.39
.15
.16
Sadness
.53
.28
-.22
.19
-.10
Affective Perceptual Sensitivity
-.06
.87
-.03
.09
-.04
General Perceptual Sensitivity
-.05
.74
.06
.05
.11
Associative Sensitivity
.09
.72
.13
-.01
-.04
Sociability
-.05
-.06
.76
.17
-.05
High Pleasure
-.10
.26
.60
-.07
-.19
Positive Affect
-.20
-.04
.57
.29
.16
Emotional Empathy
.15
.04
.33
.65
.11
Empathetic Guilt
-.01
.06
.03
.65
-.05
Social Closeness
.07
.21
.48
.47
.08
Activation Control
.17
-.05
.03
.12
.87
Effortful Attention
-.39
.15
-.04
-.21
.52
Inhibitory Control
-.34
.11
-.41
.10
.40
Note: Abbreviations: NA = Negative Affect, OR = Cognitive Sensitivity, E = Extraversion, Aff =
Affiliation, EC = Effortful Control. Loadings less than .25 not reported.
Temperament and the Big Five 46
Table 9
CFA of Adult Temperament Scales with Affiliativeness Scales for Study Two
Factor Loadings
Adult Temperament Scales
NA
OR
E
Aff
EC
Social Anger
.78
.00
.00
.00
.00
Aggression Control
-.69
.00
.00
.00
.00
Frustration
.69
.00
.00
.00
.00
Fear
.62
.00
.00
.00
.00
Sadness
.55
.00
.00
.00
.00
Discomfort
.49
.00
.00
.00
.00
Affective Perceptual Sensitivity
.00
.88
.00
.00
.00
General Perceptual Sensitivity
.00
.82
.00
.00
.00
Associative Sensitivity
.00
.72
.00
.00
.00
Sociability
.00
.00
.76
.00
.00
Positive Affect
.00
.00
.70
.00
.00
High Pleasure
.00
.00
.58
.00
.00
Social Closeness
.00
.00
.00
.88
.00
Emotional Empathy
.00
.00
.00
.73
.00
Empathetic Guilt
.00
.00
.00
.53
.00
Inhibitory Control
.00
.00
.00
.00
.78
Effortful Attention
.00
.00
.00
.00
.76
Activation Control
.00
.00
.00
.00
.58
Note: Abbreviations: NA = Negative Affect, OR = Cognitive Sensitivity, E = Extraversion,
Aff = Affiliation, EC = Effortful Control. Loadings less than .25 not reported.
Temperament and the Big Five 47
Table 10
Correlations of Temperament Factor Scores and General Construct
Scores with Big Five Scales (Study Two)
Big Five Scales
Temperament
N
I/O
E
A
C
Negative Affect
.74
-.04
-.14
-.30
-.24
Cognitive Sensitivity
.22
.65
.13
.22
.02
Extraversion
.04
.21
.67
.31
-.06
Affiliativeness
-.09
.18
.26
.69
.25
Effortful Control
-.41
.13
.14
.12
.64
Negative Affect
.70
-.05
-.22
-.33
-.18
Cognitive Sensitivity
.18
.64
.16
.22
.02
Extraversion
-.09
.20
.68
.38
.01
Affiliativeness
-.03
.27
.36
.67
.21
Effortful Control
-.43
.15
.05
.11
.60
Factor Scores
Scale Scores
Note: Abbreviations: I/O = Intellect/Openness, C = Conscientiousness, E = Extraversion,
A = Agreeableness, N = Neuroticism. Correlations greater than .35 are printed in bold.
Temperament and the Big Five 48
Table 11
Study Three: General Scales and Associated Scales from the
IPIP Version of the 100-item ATQ
General Constructs
Affiliativeness
Associated Scales
Emotional Empathy and Empathic Guilt
Affective Perceptual Sensitivity, Associative
Cognitive Sensitivity
Sensitivity, and General Perceptual Sensitivity
Effortful Control
Activation Control and Effortful Attention
High Intensity Pleasure, Positive Affect, and
Extraversion/Surgency
Sociability
Negative Affect
Fear, Sadness, and Frustration
Temperament and the Big Five 49
Table 12
Study Three: Means, Standard Deviations, and Alpha Reliability Coefficients
for General Scales and Scales from Shortened Version of ATQ
Mean
SD
Alpha
# of Items
Effortful Control
3.77
.54
.60
2
Activation Control
3.83
.66
.76
7
Effortful Attention
3.70
.60
.74
8
Cognitive Sensitivity
3.58
.58
.69
3
Affective Perceptual Sensitivity
3.40
.84
.79
6
General Perceptual Sensitivity
3.98
.62
.68
6
Associative Sensitivity
3.36
.73
.66
6
Extraversion/Surgency
3.48
.55
.59
3
High Intensity Pleasure
3.10
.76
.62
6
Positive Affect
3.75
.66
.71
6
Sociability
3.60
.79
.79
6
Negative Affect
2.69
.54
.75
4
*Non-aggressive Negative Affect
2.76
.67
.74
2
Fear
2.69
.77
.77
8
Sadness
2.82
.73
.72
7
*Aggressive Negative Affect
2.63
.59
.73
2
Frustration
2.74
.65
.72
8
Social Anger
2.52
.69
.78
8
Temperament and the Big Five 50
Table 12 continued
Affiliativeness w/o Soc. Close.
4.18
.50
.65
2
Emotional Empathy
4.09
.58
.68
6
Empathic Guilt
4.27
.58
.64
6
**Social Closeness
4.24
.51
.61
6
Note: The number of items for general scales refers to the number of scales
(i.e., subscales within the general scales. General Scales printed in italics.
Temperament and the Big Five 51
Table 13
Pattern Matrix for EFA of Adult Temperament Scales for Study Three
Factor Loadings
Adult Temperament Scales
CS
NE
Aff
EC
PA
Affective Perceptual Sensitivity
.76
-.07
.08
-.09
-.09
General Perceptual Sensitivity
.67
-.07
.02
.10
.04
Associative Sensitivity
.55
.10
-.14
.03
.11
Fear
-.07
.82
.11
.05
.06
Sadness
.22
.69
.14
-.07
-.11
Frustration
-.14
.49
-.19
-.03
.11
Emotional Empathy
.08
.06
.73
.09
.06
Empathetic Guilt
-.10
.05
.68
.04
-.02
Effortful Attention
.18
-.10
-.08
.69
-.08
Activation Control
-.15
.06
.13
.66
.08
High Intensity Pleasure
.26
.15
-.15
.01
.59
Sociability
-.06
.03
.06
.02
.57
Positive Affect
.05
-.35
.21
-.06
.43
Note: Abbreviations: NA = Negative Affect, OR = Cognitive Sensitivity, E = Extraversion,
Aff = Affiliation, EC = Effortful Control.
Temperament and the Big Five 52
Table 14
Confirmatory Factor Analysis Loadings from Community Adult Temperament
Scales for Study Three
Factor Loadings
Adult Temperament Scales
CS
NE
Aff
EC
PA
Affective Perceptual Sensitivity
.76
.00
.00
.00
.00
General Perceptual Sensitivity
.67
.00
.00
.00
.00
Associative Sensitivity
.55
.00
.00
.00
.00
Fear
.00
.82
.00
.00
.00
Sadness
.00
.69
.00
.00
.00
Frustration
.00
.49
.00
.00
.00
Emotional Empathy
.00
.00
.73
.00
.00
Empathetic Guilt
.00
.00
.68
.00
.00
Effortful Attention
.00
.00
.00
.69
.00
Activation Control
.00
.00
.00
.66
.00
High Intensity Pleasure
.00
.00
.00
.00
.59
Sociability
.00
.00
.00
.00
.57
Positive Affect
.00
.00
.00
.00
.43
Note: Abbreviations: NA = Negative Affect, OR = Cognitive Sensitivity, E = Extraversion,
Aff = Affiliativeness, EC = Effortful Control.
Temperament and the Big Five 53
Table 15
Correlations of Temperament Factor Scores and General Construct Scores with
NEO-PI Five Factor Model Scales (Study Three)
Big Five Scales
Temperament
N
I/O
C
E
A
Negative Affect
.70
-.08
-.25
-.22
-.13
Cognitive Sensitivity
.06
.61
-.05
.28
.08
Effortful Control
-.46
-.02
.59
.19
.09
Extraversion/Surgency
-.20
.40
.06
.65
.17
Affilliativeness
.05
.20
.09
.13
.49
Negative Affect
.67
-.06
-.19
-.15
-.16
Cognitive Sensitivity
.05
.61
-.04
.23
.01
Effortful Control
-.38
-.02
.60
.17
.06
Extraversion/Surgency
-.07
.33
-.01
.67
.09
Affiliativeness
.03
.15
.10
.08
.50
Factor Scores
Construct Scores
Note: Abbreviations: I/O = Intellect/Openness, C = Conscientiousness, E = Extraversion,
A = Agreeableness, N = Neuroticism. Correlations greater than .35 are printed in bold.
Temperament and the Big Five 54
Authors’ Note
This research was supported by the Emotion Research Training grant 5 T32 MH18934
funded by NIMH and awarded to David Evans as a trainee, and by NIMH grants MH43361 and
MH40662 awarded to Mary Rothbart.
We thank Doug Derryberry, Carmen Gonzalez, Catharina Hartman, Bertram Malle,
Tommie Mobbs, Michael Posner, Myron Rothbart, Gerard Saucier, and anonymous reviewers
for their helpful comments on previous drafts of this paper.
Correspondence concerning this article should be addressed to the first author:
David E. Evans, H. Lee Moffitt Cancer Center & Research Institute: Tobacco Research and
Intervention Program, 4115 E. Fowler Ave., Tampa, FL 33617. Email: evansde@moffitt.usf.edu
Temperament and the Big Five 55
References
Ahadi, S. A,. & Rothbart, M .K. (1994). Temperament, development, and the Big Five. In C. F.
Halverson, Jr., G. A. Kohnstamm & R. P. Martin (Eds.), The developing structure of
temperament and personality from infancy to adulthood (pp. 189-207). Hillsdale, NJ:
Erlbaum.
Ahadi, S. A., Rothbart, M. K., & Ye, R. (1993). Children’s temperament in the U.S. and China:
Similarities and differences. European Journal of Personality, 7, 359-377.
Allport, G. W., & Odbert, H. S. (1936). Trait names: A psycho-lexical study. Psychological
Monographs, 47, 171.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders
(4th). Washington, DC: Author.
Angleitner, A., & Ostendorf, F. (1994). Temperament and the Big Five factors of personality.
In C. F. Halverson, Jr., G. A. Kohnstamm & R. P. Martin (Eds.), The developing
structure of temperament and personality from infancy to adulthood (pp. 69-90).
Hillsdale, NJ: Erlbaum.
Ashby, F. G., Isen, A. M., & Turken, A. U. (1999). A neuropsychological theory of positive
affect and its influence on cognition. Journal of Personality and Social Psychology, 106,
529-550.
Block, J. (1995). A contrarian view of the five-factor approach to personality description.
Psychological Bulletin, 117, 187-215.
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen
& J. S. Long (Eds.), Testing structural equation models (pp. 136-162). Newbury Park,
CA: Sage.
Temperament and the Big Five 56
Buss, A. H. (1991). The EAS theory of temperament. In J. Strelau & A. Angleitner (Eds.),
Explorations in temperament, (pp. 45-60). New York: Plenum Press.
Buss, A. H., & Plomin, R. (1975). A temperament theory of personality development. New
York: Wiley-Interscience.
Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale,
NJ: Erlbaum.
Capaldi, D., & Rothbart, M. K. (1992). Development and validation of an early adolescent
temperament measure. Journal of Early Adolescence, 12, 153-173.
Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective
responses to impending reward and punishment: The BIS/BAS scales. Journal of
Personality and Social Psychology, 67, 319-333.
Clark, L. A., & Watson, D. (1999). Temperament: A new paradigm for trait psychology. In L.
A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd
Edition; pp. 399-423). New York: Guilford.
Cloninger, R. C. (1998). The genetics and psychobiology of the seven-factor model of
personality. In K. R. Silk (Ed.), Biology of personality disorders (pp. 63-92).
Washington, DC: American Psychiatric Press.
Cloninger, C. R., Przybeck, T. R., Svrakic, D. M., & Wetzel, R. D. (1994). The Temperament
and Character Inventory (TCI): A guide to its development and use. St. Louis, MO:
Center for Psychobiology of Personality, Washington University.
Costa, P. T., Jr., & McCrae R. R. (1994). Stability and change in personality from adolescence
through adulthood. In C. F. Halverson, Jr., G. A. Kohnstamm, & R. P. Martin (Eds.), The
developing structure of temperament and personality from infancy to adulthood (pp. 139150). Hillsdale, NJ: Erlbaum.
Temperament and the Big Five 57
Davidson, R. J. (1993). The neuropsychology of emotion and affective style. In M. Lewis & J.
M. Haviland (Eds.), Handbook of emotions (pp. 143-154). New York: Guilford.
Dawkins, K., & Furnham, A. (1989). The colour naming of emotional words. British Journal of
Psychology, 80, 383-389.
Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine,
facilitation of incentive motivation and extraversion. Behavioral and Brain Sciences, 22,
491-569.
Derryberry, D., & Rothbart, M. K. (1988). Arousal, affect, and attention as components of
temperament. Journal of Personality and Social Psychology, 55, 958-966.
Derryberry, D., & Rothbart, M. K. (1997). Reactive and effortful processes in the organization
of temperament. Development and Psychopathology, 9, 633-652.
Derryberry, D., & Tucker, D. M. (1992). Neural mechanisms of emotion. Journal of Consulting
and Clinical Psychology, 60, 329-338.
Diener, E. (2000). Introduction to the special section on personality development. Journal of
Personality and Social Psychology, 78, 120-121.
Digman, J. M. (1997). Higher-order factors of the Big Five. Journal of Personality and Social
Psychology, 73, 1246-1256.
Digman, J. M., & Shmelyov, A. G. (1996). The structure of temperament and personality in
Russian children. Journal of Personality and Social Psychology, 71, 341-351.
Ellis, L. K., & Rothbart, M. K. (2001). Revision of the Early Adolescent Temperament
Questionnaire. Manuscript in preparation.
Eysenck, H. J. (1991). Four ways five-factors are not basic. Journal of Personality and
Individual Differences, 13, 667-673.
Temperament and the Big Five 58
Eysenck, H. J., & Eysenck, M. W. (1985). Personality and individual differences: A natural
science approach. New York: Plenum Press.
Gartstein, M. A., & Rothbart, M. K. (2003). Studying infant temperament via the Revised Infant
Behavior Questionnaire. Infant Behavior and Development, 26, 64-86.
Goldberg, L. R. (1992). The development of markers for the Big Five-factor structure.
Psychological Assessment, 4, 26-42.
Goldberg, L. R. (1993). The structure of phenotypic traits. American Psychologist, 48, 26-34.
Goldsmith, H. H., Buss, A., Plomin, R., Rothbart, M. K., Thomas, A., Chess, S., Hinde, R. A., &
McCall, R. B. (1987). What is temperament? Four approaches. Child Development, 58,
505-529.
Gray, J. A. (1990). Brain systems that mediate both emotion and cognition. Cognition and
Emotion, 4, 269-288.
Harkness, A. R., Tellegen, A., & Waller, N. (1995). Differential convergence of self-report and
informant data for multidimensional personality questionnaire: Implications for the
construct of negative emotionality. Journal of Personality Assessment, 64, 185-204.
Hogan, R., & Horan, J. (1995). Hogan Personality Inventory manual: Second edition. Tulsa,
OK: Hogan Assessment Systems.
Isen, A. M., & Daubman, K. A. (1984). The influence of affect on categorization. Journal of
Personality and Social Psychology, 47, 1206-1217.
Isen, A. M., Johnson, M. S., Mertz, E., & Robinson, G. F. (1985). The influence of positive
affect on the unusualness of word associations. Journal of Personality and Social
Psychology, 48, 1413-1426.
Temperament and the Big Five 59
John, O. P. (1990). The “Big Five” factor taxonomy: Dimensions of personality in the natural
language and in questionnaires. Handbook of Personality Theory and Research (pp. 66100). New York: Guilford Press.
John, P. O., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and
theoretical perspectives. In L. A. Pervin, & O. P. John (Eds.), Handbook of personality:
Theory and research (2nd edition., pp. 102-138). NY: Guilford.
Lucas, R. E., Diener, E., Grob, A., Suh, M. E., & Shao, L. (2000). Cross-cultural evidence for
the fundamental features of extraversion. Journal of Personality and Social Psychology,
79, 452-468.
MacLeod, C., & Hagan, R. (1992). Individual differences in the selective processing of
threatening information, and emotional responses to a stressful life event. Behavior
Research and Therapy, 30, 151-161.
McCrae, R. R. (1993-1994). Openness to experience as a basic dimension of personality.
Imagination, Cognition and Personality, 13, 39-55.
McCrae, R. R., & Costa, P. T., Jr. (1996). Toward a new generation of personality theories:
Theoretical contexts for the Five Factor Model. In J. S. Wiggins (Ed), The Five Factor
Model of personality: Theoretical perspectives (pp. 51-87). New York: Guilford Press.
McCrae, R. R., Costa, P. T., Jr., Ostendorf, F., Angleitner, A., Hrebickova, M., Avia, M. D.,
Sanz, J., Sanchez-Bernardos, M. L., Kusdil, M. E., Woodfield, R., Saunders, P. R., &
Smith, P. B. (2000). Nature over nurture: Temperament, personality, and life span
development. Journal of Personality and Social Psychology, 78, 173-186.
McCrae, R. R., & John, O. P. (1992). An introduction to the Five Factor Model and its
applications. Journal of Personality, 60, 175-215.
Temperament and the Big Five 60
McCrae, R. R., Zonderman, A. B., Costa, P. T., Jr., Bond, M. H., & Paunonen, S. V. (1996).
Evaluating replicability of factors in the Revised NEO Personality Inventory:
Confirmatory factor analysis versus procrustes rotation. Journal of Personality and
Social Psychology, 70, 552-566.
Mervielde, I. (1998). Validity of results obtained by analyzing free personality descriptions. In
G. A. Kohnstamm, C. F. Halverson, Jr., I. Mervielde, & V. L. Havill (Eds.), Parental
descriptions of child personality: Developmental antecedents of the Big Five (pp. 189203). Mahwah, NJ: Erlbaum.
Mogg, K., Bradley, B. P., & Williams, R. (1995). Attentional bias in anxiety and depression:
The role of awareness. British Journal of Clinical Psychology, 34, 17-36.
Myers, L. B., & McKenna, F. P. (1996). The colour naming of socially threatening words.
Personality and Individual Differences, 20, 801-803.
Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions.
New York: Oxford University Press.
Posner, M. I., & Dehaene, S. (1994). Attentional Networks. Trends in Neurosciences, 17, 7579.
Posner, M.I., & Fan, J. (in press). Attention as an organ system. In J. Pomerantz (Ed.),
Neurobiology of perception and communication: From synapse to society. The IVth De
Lange Conference. Cambridge UK: Cambridge University Press.
Posner, M. I., & Raichle, M. E. (1994). Images of the mind. New York: Scientific American
Library.
Pratto, F., & John, O. P. (1991). Automatic vigilance: The attention grabbing power of negative
social information. Journal of Personality and Social Psychology, 61, 380-391.
Temperament and the Big Five 61
Putnam, S. P., Ellis, L. K., & Rothbart, M. K. (2001). The structure of temperament from
infancy through adolescence. In A. Eliasz & A. Angleitner (Eds.), Advances in research
on temperament (pp. 165-182). Germany: Pabst Scientific.
Robins, R. W., John, O. P., & Caspi, A. (1994). Major dimensions of personality in early
adolescence: The Big Five and beyond. In C. F. Halverson, Jr., G. A. Kohnstamm, & R.
P. Martin (Eds.), The developing structure of temperament and personality from infancy
to adulthood (pp. 267-291). Hillsdale, NJ: Erlbaum.
Rothbart, M. K. (1994). Broad dimensions of temperament and personality. In P. Ekman, & R.
J. Davidson (Eds.), The nature of emotion (pp. 337-341). New York: Oxford University
Press.
Rothbart, M. K., Ahadi, S. A., & Evans D. E. (2000). Temperament and personality: Origins and
outcomes. Journal of Personality and Social Psychology, 78, 122-135.
Rothbart, M. K., Ahadi, S. A., Hershey, K. L., Fisher, P. (2001). Investigations of temperament
at three to seven years: The Children’s Behavior Questionnaire. Child Development, 72,
1394-1408.
Rothbart, M. K., & Bates, J. (1998). Temperament. In W. Damon (Series Ed.) & N. Eisenberg
(Vol. Ed.), Handbook of child psychology: Social, emotional and personality
development, 5th edition (pp. 105-176). New York: Wiley.
Rothbart, M. K., & Derryberry, D. (1981). Development of individual differences in
temperament. In M. E. Lamb, & A. L. Brown (Eds.), Advances in developmental
psychology (Vol. 1; pp. 37-86). Hillsdale, NJ: Erlbaum.
Temperament and the Big Five 62
Rothbart, M. K., & Derryberry, D. (2002). Temperament in children. In C. von Hofsten & L.
Bäckman (Eds.), Psychology at the turn of the millennium. Vol. 2: Social,
developmental, and clinical perspectives, (pp. 17-35). East Sussex, UK: Psychology
Press.
Rothbart, M. K., Derryberry, D., & Posner, M. (1994). A psychobiological approach to the
development of temperament. In J. Bates & T. D. Wachs (Eds.), Temperament:
Individual differences at the interface of biology and behavior (83-116). Washington,
DC: American Psychological Association.
Saucier, G. (1994). Mini-markers: A brief version of Goldberg's unipolar Big-Five markers.
Journal of Personality Assessment, 63, 506-516.
Saucier, G. (2001, August). Going beyond the Big Five. Paper presented at the annual meeting
of the American Psychological Association, San Francisco, CA.
Saucier, G., & Goldberg, L. R. (2003). The structure of personality attributes. In M. Barrick &
A. M. Ryan (Eds.), Personality and Work: Reconsidering the role of personality in
organizations (pp. 1-29). San Francisco, CA: Jossey-Bass/Pfeiffer.
Strelau, J., Angleitner, A., Bantelmann, J., & Ruch, W. (1990). The Strelau Temperament
Inventory-Revised (STI-R). Theoretical considerations and scale development.
European Journal of Personality, 4, 209-235.
Strelau, J., & Zawadzki, B. (1993). The formal characteristics of Behavior-Temperament
Inventory (FCB-TI): Theoretical assumptions and scale construction. European Journal
of Personality, 7, 313-336.
Temperament and the Big Five 63
Strelau, J., & Zawadzki, B. (1997). Temperament and personality: Eysenck’s three superfactors
as related to temperamental dimensions. In H. Nyborg (Ed.), The scientific study of
human nature: Tribute to Hans J. Eysenck at eighty (pp. 68-91). Oxford, UK:
Pergamon/Elsevier.
Svrakic, N. M., Svrakic, D. M., & Cloninger, C. R. (1996). A general quantitative theory of
personality development: Fundamentals of a self-organizing psychobiological complex.
Development and Psychopathology, 8, 247-272.
Tellegen, A. (1985). Structure of mood and personality and their relevance to assessing anxiety,
with an emphasis on self-report. In A. J. Tuma & J. D. Maser (Eds.), Anxiety and the
anxiety disorders (pp. 681-706). Hillsdale, NJ: Erlbaum.
Tellegen, A. (1993). Folk concepts and psychological concepts of personality and personality
disorder. Psychological Inquiry, 4, 122-130.
Thomas, A., & Chess, S. (1977). Temperament and development. New York: Brunner/Mazel.
Victor, J. B. (1994). The Five Factor Model applied to individual differences in school behavior.
In C. F. Halverson, Jr., G. A. Kohnstamm, & R. P. Martin (Eds.), The developing
structure of temperament and personality from infancy to adulthood (pp. 169-187).
Mahwah, NJ: Erlbaum.
Watson, D., & Clark, L. A. (1992). Affects separable and inseparable: On the hierarchical
arrangements of the negative affects. Journal of Personality and Social Psychology, 62,
489-505.
Watson, D., & Walker, L. M. (1996). The long-term stability and predictive validity of trait
measures of affect. Journal of Personality and Social Psychology, 70, 567-577.
Temperament and the Big Five 64
Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems of
affect: Structural findings, evolutionary considerations, and psychobiological evidence.
Journal of Personality and Social Psychology, 76, 820-838.
Windle, M., & Lerner, R. M. (1986). Reassessing the dimensions of temperamental individuality
across the life span: The Revised Dimensions of Temperament Survey (DOTS-R).
Journal off Adolescent Research, 1, 213-229.
Zuckerman, M. (1992). What is a basic factor and which factors are basic? Turtles all the way
down. Journal of Personality and Individual Differences, 13, 675-681.
Zuckerman, M. (1997). The psychobiological basis of personality. In H. Nyborg (Ed), The
scientific study of human nature: Tribute to Hans J. Eysenck at eighty (pp. 3-16).
Oxford, UK: Pergamon/Elsevier.
Zuckerman, M., Eysenck, S. B., & Eysenck, H. J. (1978). Sensation seeking in England and
America: Cross-cultural, age, and sex comparisons. Journal of Clinical & Consulting
Psychology, 46, 139-148.
Zuckerman, M., Kuhlman, D. M., Joireman, J., Teta, P., & Kraft, M. (1993). A comparison of
three structural models of personality: The Big Three, the Big Five, and the Alternative
Five. Journal of Personality and Social Psychology, 65, 757-768.
Temperament and the Big Five 65
Appendix A
General Constructs and Definitions of Scales with Sample Items
AFFILIATIVENESS
Emotional Empathy: Affective response congruent with what others are perceived to feel. I am
rarely bothered by the apparent suffering of strangers (coded in reverse).
Empathic Guilt: Distress in response to negatively affecting other people. Whenever I believe
that I have hurt someone’s feelings, I feel guilty.
Social Closeness: Feelings of warmth, closeness, interest, and involvement with others. There
are some people that I feel very close to.
EFFORTFUL CONTROL (Effortful Attention in Study One)
Activation Control: Capacity to perform an action when there is a strong tendency to avoid it. I
hardly ever finish things on time (coded in reverse).
Attentional Focusing: Capacity to focus attention on desired channels, thereby resisting
unintentional shifting to irrelevant or distracting channels. My concentration is rarely
disrupted if there is music in the room around me.
Attentional Shifting: Shifting of attentional focus to desired channels, thereby avoiding
unintentional focusing on particular channels. I often lose my train of thought when I have to
keep track of several things at once (coded in reverse).
Attentional Shifting from Punishment: Control over focusing and shifting of attention amid
negative thoughts and negative emotions. When I need to focus my attention upon a task,
fearful thoughts rarely distract me.
Attentional Shifting from Reward: Control over focusing and shifting of attention in the face
of distracting positive emotions and desires. If I am anticipating a rewarding activity, it is
difficult for me to think of anything else (coded in reverse).
Effortful Attention: Capacity to focus attention as well as to shift attention when desired. This
scale includes the above four scales (attentional focusing, attentional shifting, attentional
shifting from punishment, and attentional shifting from reward) collapsed together.
Temperament and the Big Five 66
Inhibitory Control: Capacity to inhibit inappropriate behavior. It is easy for me to hold back
my laughter in a situation where it is not appropriate.
EXTRAVERSION/SURGENCY
Activity Level: The rate of onset, level of intensity, frequency, endurance, and enjoyment of
behavioral activity. Sometimes I feel as though I'm full of energy.
High Intensity Pleasure: Pleasure related to situations involving high stimulus intensity, rate,
complexity, novelty, and incongruity. I would not enjoy the sensation of listening to loud
music with a laser light show (coded in reverse).
Positive Affect: Latency, threshold, intensity, duration, and frequency of experiencing pleasure.
It doesn't take much to evoke a happy response in me.
Sociability: Enjoyment derived from social interaction and being in the presence of others. I
usually enjoy being with people.
NEGATIVE AFFECT
Aggression Control: Capacity to inhibit the behavioral expression of anger. I do not have a
problem in controlling hostile impulses.
Discomfort: Unpleasant affect resulting from the sensory qualities of stimulation. I find loud
noises to be very irritating.
Fear: Unpleasant affect related to anticipation of pain or distress. Loud noises sometimes scare
me.
Frustration: Unpleasant affect related to the interruption of tasks and behavior. I seldom
become irritated when someone is late (coded in reverse).
Sadness: Unpleasant affect and lowered mood and energy related to object or person loss,
disappointment, and exposure to suffering. I rarely feel sad after saying good-bye to friends
or relatives (coded in reverse).
Social Anger: Hostility felt toward other people. I rarely feel angry at people (coded in
reverse).
Temperament and the Big Five 67
COGNITIVE SENSITIVITY
Affective Perceptual Sensitivity: Spontaneous emotionally valenced explicit cognition
associated with low intensity stimuli. I am often consciously aware of how the weather
seems to affect my mood.
Associative Sensitivity: Spontaneous cognitive content that is not related to standard
associations with the environment. When I am resting with my eyes closed, I sometimes see
visual images.
External Perceptual Sensitivity: Awareness of slight, low intensity stimulation arising from the
environment. I often notice visual details in the environment.
Internal Perceptual Sensitivity: Awareness of slight, low intensity stimulation arising from
within the body. I usually fail to notice it when my muscles tense up just slightly (coded in
reverse).
General Perceptual Sensitivity: Internal and External Perceptual Sensitivity: Combined.
Temperament and the Big Five 68
Appendix B
Mini-Markers for the Big Five
Note: Big Five scales listed in bold; associated trait-adjectives listed under the name of each
scale. Items with “R” in parenthesis indicate conceptual reverse of the associated scale’s label.
Extraversion
Agreeableness
Conscientiousness
Bashful (R)
Cold (R)
Careless (R)
Bold
Cooperative
Disorganized (R)
Energetic
Harsh (R)
Efficient
Extraversion
Kind
Inefficient (R)
Quiet (R)
Rude (R)
Organized
Shy (R)
Sympathetic
Practical
Talkative
Unsympathetic (R)
Sloppy (R)
Withdrawn (R)
Warm
Systematic
Neuroticism
Intellect
Envious
Complex
Fretful
Creative
Jealous
Deep
Moody
Imaginative
Relaxed (R)
Intellectual
Temperamental
Philosophical
Touchy
Uncreative (R)
Unenvious (R)
Unintellectual (R)