«Running Head: TWO-FACTOR MODEL OF TEMPERAMENT A Two-Factor Model of Temperament David E. Evans H. Lee Moffitt Cancer Center & Research Institute, ...»
Two-Factor Model 1
Running Head: TWO-FACTOR MODEL OF TEMPERAMENT
A Two-Factor Model of Temperament
David E. Evans
H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL
Mary K. Rothbart
University of Oregon
Key words: temperament, personality, Big Five
Two-Factor Model 2
The higher order structure of temperament using the Adult Temperament Questionnaire was
examined in three studies. Two undergraduate samples included 468 participants, and a more
diverse community sample included 700 participants. Factor analysis of theoretically generated scales and subscales consistently supported a higher order two-factor model of temperament.
Scales associated with the broad factors of extraversion/positive emotionality and orienting sensitivity loaded positively on one factor; scales associated with negative affect loaded positively and effortful control negatively on the second factor. These findings are discussed in relation to experimental research on attention and emotion. We also suggest that the higher order two-factor model of temperament might shed light on psychobiological substrates of Digman’s higher order two-factor model of personality derived from Big Five measures.
Two-Factor Model 3 A Two-Factor Model of Temperament Evidence from a number of research areas suggests the existence of at least two high level affective-motivational temperament systems: childhood temperament (Putnam, Ellis, & Rothbart, 2001); adult temperament (Strelau & Zawadzki, 1997); personality (Eysenck & Eysenck, 1985; Tellegen, 1985); neuroscience (Carver & White, 1994; Depue & Collins, 1999;
Derryberry & Tucker, 1992; Gray, 1990; Panksepp, 1998); and individual differences in emotionality (Watson & Clark, 1992; Watson & Walker, 1996; Watson, Wiese, Vaidya, & Tellegen, 1999). The first system includes dimensions of negative emotionality-neuroticism (e.g., negative emotionality [Harkness, Tellegen, & Waller, 1995; Tellegen, 1985]; negative affectivity [Ahadi, Rothbart, & Ye, 1993; Rothbart, Ahadi, Hershey, & Fisher, 2001; Watson et al., 1999]; neuroticism [Eysenck & Eysenck, 1985; McCrae & John, 1992]; and emotional stability [Goldberg, 1993]). The second system is an extraversion/positive emotionality dimension (e.g., positive emotionality [Harkness et al., 1995; Tellegen, 1985]; positive affect [Watson et al., 1999]; extraversion [Eysenck & Eysenck, 1985; McCrae & John, 1992]; and extraversion/surgency [Goldberg, 1993]).
Since variants of these two constructs are common across a number of models, researchers have discussed possible mechanisms for extraversion/positive emotionality and negative emotionality. However, the attentional constructs of effortful control (Rothbart & Rueda, 2005) and orienting sensitivity (Evans & Rothbart, 2007) have not been included in most theories of temperament, and have thereby received little attention. The orienting sensitivity construct includes awareness and thinking peripheral to current tasks, whereas effortful control includes the capacity to control attention under conditions of conflict, to plan and to detect
Evans and Rothbart (2007; Rothbart, Ahadi, & Evans, 2000) have developed a theorydriven model of temperament, extracting broad factors of orienting sensitivity, effortful control, affiliativeness, extraversion/positive emotionality, and negative affect. Data supported both fiveand six-factor models, with the six-factor model dividing negative affect into separate aggressive and nonaggressive negative affect factors. Two of the factors were related to attention (orienting sensitivity and effortful control), and the other three involve affective and motivational processes (affiliativeness, extraversion/positive emotionality, and negative affect). In these factor analyses, because we were primarily interested in a more fine-grained discrimination of temperament constructs, we explored four-, five-, and six-factor solutions. However, we also reported (Evans & Rothbart, 2007) correlations among factors, finding that negative affect and effortful control were consistently negatively associated while being relatively independent of the other factors.
These other factors were also somewhat consistently positively intercorrelated, suggesting to us a broader two-factor structure of temperament, which might be indicative of differential associations between emotional and attentional constructs at this general level. In the current research, we explored further the possibility of a higher order two-factor model, investigating relations between attentional and affective-motivational scales.
Our primary goal in these studies was to explore relations among the broad temperament constructs at a higher level. This higher order factor analytic approach to personality was followed by Digman (1997), who discovered the Big Five model of personality traits may be further reduced to two higher order factors. He labeled the factors alpha and beta. As evidenced by the Greek factor labels Digman (1997) chose for the higher factors, their psychological meaning was not selfevident. However, Digman offered speculations about their meaning. Whereas the first factor, containing emotional stability (or neuroticism in reverse), conscientiousness, and agreeableness,
second factor, combining intellect/openness and extraversion, seemed to be more related to dynamic, creative, and expressive qualities.
The higher order factor alpha (neuroticism in reverse, conscientiousness, and agreeableness,) was therefore interpreted as reflecting effects of socialization. The idea was that an early social environment encouraging stability, secure attachments, and responsibility, would result in higher levels of emotional stability (neuroticism in reverse), conscientious, and agreeable behavior. Digman suggested that the second higher order factor (beta) might have to do with personal growth and self-fulfillment related to self-actualization. Indeed, being extraverted, outgoing, and open to experience is congruent with humanistic ideals emphasizing individual self-expression. Paulhus and John (1998), however, suggested that the higher-order Digman model reflected moralistic (i.e., Digman’s alpha) and egoistic (i.e., Digman’s beta) biases, with the latter resulting from unconscious defense mechanisms, not temperament. To avoid labeling confusion, it is important to note that Paulhus and John’s research interest cites an earlier literature that used different Greek labels. Digman’s alpha was gamma, Digman’s beta was alpha according to this labeling.
Ashton, Lee, and Goldberg (2004) have also questioned the Digman two-factor model. A 30-degree rotation of Ashton et al.’s lexically derived two-factor model resulted in a factor very similar to Digman’s alpha. However, the second factor included extraversion-related content, and did not include the creative and imagination components of Digman’s beta. Ashton et al.’s model weights content loading on the earlier extracted Big Five factors (i.e., factors accounting for greater proportion of variance) more heavily than later extracted factors that account for less variance (e.g., intellect/openness). In contrast, Digman’s higher order factor structure was based
intellect/openness, as well as the other Big Five domains, was given equal weight in Digman’s model. We would argue that the factors should be equally weighted if the goal is to discover higher order relations among constructs.
Biesanz and West (2004) examined the Digman model at the scale level. Their findings showed support for the Digman model using self-report data at multiple time points. In contrast, Digman’s model was not supported using multiple informants. It was suggested that correlations among Big Five/FFM domains were method specific (i.e., self-report only), which in turn might reflect egoistic and moralistic biases, as opposed to reflecting real associations between the Big Five/FFM domains. However, DeYoung (2006) noted that the cross-informant correlations were very low (.18 to.43) in the Biesanz and West (2004) study, which constrained their correlations, as well as the reliability of correlations across observers. DeYoung (2006) used instead the Big Five Inventory, which has substantially greater cross-informant correlations. The resulting higher order factor analysis replicated the Digman alpha and beta two-factor structure.
Rothbart and Derryberry (1981) defined temperament as constitutionally based individual differences in reactivity and self-regulation. In addition to emphasizing emotional reactivity, these theorists viewed attentional processes as core self-regulatory components of temperament (Derryberry & Rothbart, 1988; 1997; Rothbart & Derryberry, 1981; 2002). This approach emphasizes identifying and characterizing basic temperamental processes while also exploring the interrelations among these processes (see Derryberry & Rothbart, 1997; Rothbart & Bates, 2006). Complex facets of human personality extending beyond temperament, such as attitudes, beliefs, goals, and values, develop out of evolutionary conserved temperament systems while also incorporating higher-level cognitive functioning that is relatively unique to human beings
people differ in social interaction, it is thus important to study core temperament processes.
The Adult Temperament Questionnaire (ATQ) is a measure that built on the earlier Physiological Reactivity Questionnaire (PRQ) developed by Derryberry and Rothbart (1988).
Scales previously developed as part of the PRQ were included in the first version of the ATQ, as well as additional temperament constructs. The PRQ emphasized definitional specificity of subscales loosely grouped under more general constructs (e.g., arousal, affect, and attention), and the ATQ included the same rigor in definitional specificity while investigating structure at a higher level. For example, the subscales attentional shifting, attentional focusing, attentional shifting from distress, and attentional shifting from reward included in Study 1 of this paper (also see Evans, 2004) all fit the higher-level definition of effortful attention. Appendix A contains operational definitions and sample items for all the subscales included in the three studies presented in this paper. The second version of the ATQ (Study 2) included additional conceptual breadth and empirical refinement. Finally, the third version of the ATQ (Study 3) is a short form (i.e., both fewer number of items, and fewer subscales) of the measure from Study 2. Each of the ATQ measures have reliable subscales, and the subscales show good convergent and divergent validity, as indicated by the factor analyses showing subscales loading on factors primarily indicative of their parent more general constructs (see Evans, 2004; Evans & Rothbart, 2007; Rothbart et al., 2000). Five- and six-factor models included factors labeled Orienting Sensitivity, Effortful Control, Extraversion/Positive Emotionality, Affiliativeness, and Negative Affect (negative affect was sub-factored into aggressive and nonaggressive negative affect in the six-factor model).
Relations of both the five- and six-factor temperament models converged substantially with the Big Five/FFM factors. Among a college sample, there was one-to-one convergence
negative affect factor scores were highly correlated with Big Five neuroticism (r =.74), orienting sensitivity with Big Five intellect/openness (r =.65), temperamental extraversion/surgency with Big Five extraversion (r =.67), and affiliativeness with Big Five agreeableness (r =.69). The effortful control factor score was highly correlated with Big Five conscientiousness (r =.64), while also having a substantial negative correlation (r = -.41) with Big Five neuroticism. Among a larger community sample, these findings were essentially replicated and extended. Five of the factor scores from the six temperament factors converged with the FFM scales, with correlations ranging from.52 to.69. Non-aggressive negative affect correlated highest with neuroticism (r =.69), while aggressive negative affect was correlated with both neuroticism (r =.57) and agreeableness (r = -.43).
The studies presented in this paper explore the organization of temperament at a higher level. In Study 1, two-factor solutions of the subscales were explored and the resulting twofactor structure is reported for each of three data sets. In Study 1, we also studied effects of removing subscales involving affect from the attentional constructs. In Studies 2 and 3, affiliative content was added to the analyses. Study 3 afforded the opportunity to explore relations between ATQ and Big Five/FFM two-factor models, using Goldberg’s (1992) 100
Participants A sample of 210 University of Oregon undergraduate psychology students filled out the first version of the ATQ (ATQ-1). Subjects received credit toward their psychology courses for participating. Participants completed the questionnaire in group settings. Three subjects who did not complete the questionnaire were excluded; 207 subjects’ responses were analyzed.
Gender information was not collected for the first study.