- Email: [email protected]
States reflecting the Big Five dimensions
Personality and Individual Differences 34 (2003) 591–603 www.elsevier.com/locate/paid
States reflecting the Big Five dimensions Nicola S. Schutte*,a, John M. Malouffa, Elida Segrerab, Amanda Wolfb, Larissa Rodgersb a
University of New England, School of Psychology, Armidale, NSW 2351 Australia Nova Southeastern University, Behavioral Sciences, Ft. Lauderdale, FL 33314, USA
b
Received 2 July 2001; received in revised form 15 December 2001; accepted 21 January 2002
Abstract Two studies explored the possibility that the Big Five dimensions, which extensive research has shown underlie most human traits, also provide a structure for transitory states. A confirmatory factor analysis showed an acceptable fit between responses on measures of transitory states and the Big Five dimensions. Further, the state measures of the Big Five dimensions had good internal consistency. As one would expect, each Big Five state was more related to the corresponding Big Five trait than to other Big Five traits. As expected on the basis of previous research, higher levels of state surgency were associated with higher levels of state positive mood, and higher levels of state emotional stability were associated with lower levels of state negative mood. Unexpectedly, state conscientiousness was also highly associated with state positive mood. Because one would expect states to be changeable, the second study used an experimental manipulation to attempt to change levels of the Big Five States. All states changed in the expected direction; however, only the changes in state surgency, agreeableness, and openness were statistically significant. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Big Five; States
1. States reflecting the Big Five dimensions A number of theories of human behavior (Allport, 1961; Buss & Craik, 1983; Costa & McCrae, 1992; Goldberg, 1993; Wakefield, 1989) emphasize the hierarchical organization of human functioning. According to these theories general and enduring characteristics are at the top of the hierarchy and more specific or passing characteristics are at the bottom of the hierarchy. These
* Corresponding author. E-mail address: [email protected] (N. S. Schutte). 0191-8869/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0191-8869(02)00031-4
592
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
characteristics at the bottom of the hierarchy are generally viewed as being less descriptive of an individual and as being prompted by certain situations or cognitive processes. A related conceptualization of human characteristics is their classification into traits or states (e.g. Spielberger & Sydeman, 1994; Watson, Clark, & Tellegen, 1988). In a hierarchical model, traits are conceptualized as higher-level and enduring characteristics, while states are lower level and less enduring characteristics. Trait anxiety, for instance, disposes individuals to feel chronic anxiety, while state anxiety is a situationally linked experience of anxiety that passes when the situation is no longer present (Spielberger & Sydeman 1994). A similar differentiation can be made between state and trait anger (Spielberger & Sydeman, 1994). Watson et al. (1988) studied characteristic (general) positive and negative affect and state (momentary) positive and negative mood. As one would expect, ratings of characteristic positive and negative affect were relatively unchanged over the course of an eight-week period, while ratings of momentary positive and negative mood showed some consistency, but more variability than characteristic affect. Similarly, Usala and Hertzog (1991) found that state anxiety was less stable over time than trait anxiety. Thus, one might conclude that these mood-related states are the result of characteristic affect interacting with situational influences. Characteristics such as states, which may be placed at the bottom of a hierarchical conceptualization of functioning, may be related to important outcomes. For example, state anxiety has been found to be associated with performance (Catanzano, 1996; Menzel & Carrell, 1994) and cognitive dissonance (Menasco & Hawkins, 1978). State mood that has been manipulated on the positive mood dimension has been found to be associated with helping behavior (Isen and Levine, 1972; Isen & Simmonds, 1978), memory (Teasdale & Barnard, 1993), perception (Forgas & Bower, 1987), and judgement (Forgas, 1995). State anxiety, state anger, and state positive and negative mood have been investigated. However, relatively little research has focused on other types of states. Five dimensions, surgency, agreeableness, conscientiousness, emotional stability, and openness, seem to underlie many characteristic traits (Goldberg, 1993; John & Srivastava, 1999; McCrae & Costa, 1999). These Big Five dimensions have been identified in many factor analytic studies (Costa & McCrae, 1992, 1996; Digman, 1990; John, 1990; John & Srivastava, 1999; McCrae & Costa, 1999), including cross-cultural studies (McCrae & Costa, 1997a; Saucier & Goldberg, 1996). Typically these studies have found that if individuals rate themselves or others on a wide variety of trait descriptors, five factors emerge. Different researchers have applied somewhat different terms to these factors, e.g. using the term ‘‘extraversion’’ instead of ‘‘surgency,’’ low ‘‘neuroticism’’ instead of ‘‘emotional stability,’’ and ‘‘intellect’’ or ‘‘imagination’’ instead of ‘‘openness’’ (Goldberg, 1992; Saucier & Goldberg, 1996). The consistent finding that many personality traits group into five dimensions might be in part due to biological predispositions to organize responses into these dimensions. Research showing moderate heritability of the Big Five traits and facets of the Big Five traits (e.g. Jang, McCrae, Angleitner, Riemann & Livesley, 1998; Loehlin, McCrae, Costa & John, 1998) supports the possibility of such a biological predisposition. In a review of research on the Big Five dimensions, John and Srivastava (1999) pointed out that as well as being important in understanding the organization of human functioning, the Big Five dimensions are related to important life outcomes. For example, research with adolescents has found that low agreeableness and low conscientiousness predict juvenile delinquency; neuroticism
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
593
and low conscientiousness predict internalizing psychopathologies; and conscientiousness and openness are associated with good school performance. Research with adults has shown that conscientiousness predicts good work performance and good health, low agreeableness and high neuroticism are associated with poor health, high agreeableness is associated with helping others, high extraversion predicts leadership, high neuroticism is associated with depression, and high openness is associated with creativity. In a factor analytic study, Borkenau and Ostendorf (1998) explored the relationship between daily states and the Big Five dimensions. In this study participants used adjective checklists to record their states for each day during 90 consecutive days. Two important findings were that there was substantial variability in the day-to-day responding of most participants and that the overall pattern of responding across the 90 days grouped into the Big Five dimensions. Similarly, using an experience sampling methodology, Fleeson (2001) found evidence for within-person variability, which indicated that situation-specific behavior may be grouped into the Big Five dimensions, as well as stable average tendencies for individuals, supporting the existence of stable Big Five traits. Nemanick and Munz (1997) examined the relationship between two Big Five traits and characteristics at different levels in a hierarchical conceptualization of personality. They conceptualized the Big-Five characteristic of extraversion as being at a high level of characteristic personality explanation, characteristic positive affectivity as being on an intermediate level, and affect on a given day as being on a relatively low level. Extraversion was associated more strongly with characteristic positive affectivity than with positive affect on a given day. Characteristic positive affectivity in turn was more strongly related to positive affect on a given day. A similar pattern was found for the relationship between the Big Five factor of neuroticism, conceptualized as being at a high level of explanation; characteristic negative affectivity, an intermediate characteristic; and negative affect on a given day. If personality characteristics group into five dimensions, and if these five dimensions are superordinate categories in a hierarchical organization of human functioning, then it is possible that the state level of functioning reflects these five dimensions. The research on daily states by Borkenau and Ostendorf (1998), density distributions of behaviors by Fleeson (2001) and the relationship between trait extraversion and neuroticism and daily mood states by Nemanick and Munz (1997) supports this view. The purpose of the current research was to investigate further whether there might be present states that correspond to the Big Five dimensions, these states’ relationship to the Big Five traits and to mood, and whether these states are indeed changeable as one would expect of states. A related purpose was to develop and validate present state measures of each of the Big Five characteristics. Study 1: Development of State Measures of the Big Five, Factor Structure and Internal Consistency of the States, Relationship Between the Big Five States and Traits, and Relationship Between the Big Five States and Positive and Negative State Mood. 1.1. Overview Saucier (1994) developed a short form of the Big-Five trait measure created by Goldberg (1992). We altered this short form to make it into present a state measure of the Big Five dimensions. This measure focused on immediate states. If the Big Five dimensions apply to states as well as to traits, one would expect the factor structure of immediate states to be similar to the
594
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
factor structure of traits. If there were a hierarchical relationship between traits and states, then one would expect each of the Big Five states to relate more closely to the corresponding Big Five trait than to any other Big Five trait. Further, because Nemanick and Munz (1997) found surgency (extraversion) to be associated with positive affectivity, which in turn was associated with daily positive mood, and neuroticism to be associated with negative affectivity, which in turn was associated with daily negative mood, one would expect a moderate positive relationship between present-state surgency and state positive mood and a moderate negative relationship between present-state emotional stability and state negative mood. 1.2. Method 1.2.1. Development of the present state measures of the big five characteristics Goldberg (1992) created a set of 100 adjectives for use in measuring the big-five dimensions: surgency, agreeableness, conscientiousness, emotional stability, and intellect. Applying factor analysis techniques to hundreds of English trait terms Goldberg (1982, 1990) obtained the usual Big Five factor pattern. After exploring the usefulness of bipolar items, Goldberg (1992) developed a 100-item set of unipolar adjectives (e.g. fearful, innovative) to measure the Big Five dimensions. He selected 20 items for each of the five dimensions on the basis of factor analysis results on hundreds of adjectives in prior research (Goldberg, 1982). Goldberg (1992) found that the Big Five dimensions on the Unipolar Marker scale correlated in expected ways with scores on the NEO Personality Inventory, which is a commonly used measure of the Big Five: Surgency correlated most highly with NEO extroversion, Agreeableness most highly with NEO agreeableness, Conscientiousness most highly with NEO conscientiousness, Emotional Stability most highly with low NEO neuroticism, and Intellect most highly with NEO openness. Saucier (1994) shortened the 100-item Unipolar Marker measure to 40 items, with eight items representing each dimension. He used the following criteria for item selection: (a) high factor loadings on the relevant dimension, (b) that an item be easily understood, and (c) high correlation with the full set of items for the dimension. We adapted the 40-item Unipolar Marker measure to a state measure of the Big-Five dimensions. On the original and brief Unipolar Marker measures of the Big-Five traits, respondents rate themselves as they are generally or characteristically. We modified these instructions to ask respondents to rate themselves at the present moment. The instructions were: ‘‘Describe yourself as you feel right now, that is, at the present moment. Describe yourself as you see yourself at the present time, not as you wish to be in the future or as you were in the past.’’ As was done on the original version of the measure, respondents then rated themselves on 40 adjectives using a 9-point scale on which ‘‘1’’ means extremely inaccurate and ‘‘9’’ means extremely accurate. The adjectives comprise five subscales, each of which represents one of the BigFive characteristics. Examples of adjectives from each of the subscales are as follows, surgency: ‘‘bold;’’ agreeableness: ‘‘kind;’’ conscientiousness: ‘‘efficient;’’ emotional stability: ‘‘relaxed;’’ and openness: ‘‘creative.’’ 1.2.2. Participants A sample of 189 participants was recruited from a university in the southeastern United States and from community settings, such as sports centers and workplaces, in the southeastern United
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
595
States. Participants’ average age was 23.26, SD=7.00; 132 were women and 57 were men. All participated on a voluntary basis and none were paid for their participation. 1.2.3. Procedure All participants completed the present-state measure of the Big Five dimensions. Sixty-eight of the participants also completed the Big-Five Trait Inventory (John & Srivastava, 1999), which has evidence of good reliability and validity. Items are in phrase form and individuals rate themselves on how well they fit the description given in each item; a typical example of an item is ‘‘helpful and unselfish with others’’ for the agreeableness scale. The other 121 participants completed the state version of the Positive and Negative Affect Scales (PANAS; Watson & Clark, 1997; Watson et al., 1988), which assesses positive and negative mood states as independent dimensions and which has evidence of good reliability and validity. Items on these scales are in adjective form and typical items are ‘‘enthusiastic’’ for positive affect and ‘‘distressed’’ for negative affect. 1.3. Results 1.3.1. Descriptive statistics The average scores and standard deviations on the Big Five state measures were as follows: surgency, M=47.84, SD=11.22; agreeableness, M=54.41, SD=10.02; conscientiousness, M=51.29, SD=10.75; openness, M=50.73, SD=10.51; emotional stability, M=44.33, SD=11.38. Average scores on the Big-Five Trait Inventory were as follows: extraversion, M=27.46, SD=6.38; agreeableness, M=35.87, SD=5.16; conscientiousness, M=35.12, SD=5.79; openness, M=37.06; SD=6.80, neuroticism, M=26.71, SD=6.82. The average scores and standard deviations for mood were M=31.80, SD=10.38 for positive mood and M=16.29, SD=7.15 for negative mood. 1.3.2. Confirmatory factor analysis The Big Five present state measures are theoretically separate factors, although they may intercorrelate to some degree as do Big Five trait measures. In order to test whether the items on the Big Five state measures actually load on the Big Five dimensions, we performed a maximum likelihood confirmatory factor analysis. Following the guidelines set out by Floyd and Widaman (1995), we created two parcels of four items for each hypothesized state factor. This reduces the instability associated with the use of individual items (see Floyd & Widaman, 1995). Following the suggestions of Kishton and Widaman (1994), we randomly assigned four scale items to each parcel. Fig. 1 shows the model. To avoid nonidentifiability of the model, we followed the suggestions of Arbuckle and Wothke (1995) and constrained to unity the path from every latent error variable to its measured variable (parcel) and the path from each latent construct (e.g. surgency) to its first measured variable (parcel). In the initial testing of the model, the solution indicated that agreeableness parcel 2 had negative variance. This common problem in structural equation modeling can be eliminated by setting the variance for that variable to 0 (Loehlin, 1998). Using Amos software, we followed the suggestions of Arbuckle (personal communication), who created the software, and set the variance for that parcel to 0.000001 to be as close to 0 as the software allows. Setting that parameter had no appreciable effect on the fit indicators. The model was evaluated with the chi-square statistic, which is a test of absolute fit that is difficult to pass in analyses based on large samples or in models with many observed variables (Floyd & Widaman,
596
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
Fig. 1. Confirmatory factor analysis model of the Big Five states. Large ellipses represent latent factors, boxes represent measured variables comprised of item parcels, and small ellipses represent error of the measured variables.
1995); three relative fit statistics, the Bollen incremental fit index (IFI), the Tucker Lewis index (TLI), and the Bentler comparative fit index (CFI); and one measure based on population discrepancies, the root mean square error of approximation (RMSEA). There are no universally accepted standards for determining the appropriateness of a model, so we used the commonly followed rules (see Lucas & Fujita, 2000) that a model fits well if the chisquare is nonsignificant, the three fit indices are greater than 0.90, and the RMSEA does not exceed 0.05 The results were that the chi-square (df=25) equaled 62.7, P<0.01; the three fit indices were all at 0.99, and the RMSEA was <0.09. Hence, the model (a) failed to meet the standard of the absolute fit statistic, chi-square, (b) met the standard of the IFI, (c) met the standard of the TLI, (d) met the standard of the CFI, and (e) almost met the RMSEA population discrepancy standard, falling in an area, <0.09, that is usually called marginal (e.g. Loehlin, 1998). See Fig. 1 for the parameter estimates. 1.3.3. Internal consistency and relationship of the states with each other Table 1 shows the internal consistency of each of the Big Five state measures and the relationships of the states with each other. The Cronbach alpha coefficients for each of the state measures
597
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603 Table 1 Alpha coefficients of the Big Five state measures and correlations between the Big Five states (N=189)
Surgency Agreeableness Conscientiousness Openness Emotional Stability
Surgency
Agreeableness
Conscientiousness
Openness
Emotional stability
(0.78) 0.20a 0.22a 0.30a 0.15
(0.73) 0.37a 0.33a 0.29a
(0.79) 0.32a 0.30a
(0.75) 0.05
(0.75)
Note. Alpha coefficients are presented on the diagonal; correlations are presented below the diagonal. a p < 0.01.
showed reasonable internal consistency for the eight items comprising each of the scales. The states showed low to moderate correlations with each other. The median correlation was 0.295, with a range of 0.05 to 0.37. 1.3.4. Relationship between Big-Five states and Big-Five traits Each of the Big-Five states was most highly associated with the corresponding Big-Five trait (see Table 2). While all of the correlations between states and their corresponding traits were statistically significant, none, with possibly the exception of surgency, had so much common variance that they seemed to be measuring the same construct. The median within-dimension (e.g. state agreeableness and trait agreeableness) correlation was 0.58, with a range of 0.39 to 0.77. The median inter-dimension correlation was 0.09, with a range of 0.00 to 0.28. 1.3.5. Relationship of state surgency and state emotional stability with mood As predicted, higher levels of state surgency were significantly associated with higher levels of state positive mood. Surprisingly, higher levels of state conscientiousness were also strongly associated with more state positive mood. As predicted higher levels of state emotional stability were significantly associated with lower levels of state negative mood. See Table 3 for the correlations.
Table 2 Correlations between Big Five states and Big Five traits (n=68) States Traits
Surgency
Surgency Agreeableness Conscientiousness Openness Emotional stability
0.77a 0.02 0.17 0.28 0.14
Agreeableness 0.00 0.42a 0.08 0.13 0.24
Conscientiousness 0.15 0.06 0.65a 0.05 0.23
Openness 0.10 0.09 0.02 0.58a 0.01
Emotional stability 0.16 0.23 0.09 0.09 0.39a
Note. Hypothesized correlations are underlined. Traits were measured with the Big Five trait inventory. Responses on trait emotional stability were scored so that high scores indicated greater emotional stability. a p < 0.01.
598
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
Table 3 Correlations between the Big Five states and positive and negative mood (N=121) States
Positive mood
Surgency Agreeableness Conscientiousness Openness Emotional stability
0.45a 0.33a 0.56a 0.36a 0.31a
Negative mood 0.18 0.23 0.29a 0.01 0.46a
Note. Hypothesized correlations are underlined. a p< 0.01.
Regression analyses investigated the relative value of the five states in predicting positive and negative mood. With an R squared of 0.32, F (5,115)=54.75, P<0.001, state conscientiousness best predicted positive mood. State surgency significantly added to this prediction with an R squared increase of 0.10, F (5,115)=19.89, P<0.001. With an R squared 0.21, F (5,115)=29.79, P<0.001, state emotional stability was the best predictor of negative mood. None of the other states added significantly to these predictions.
2. Study 2: Changeability of the Big Five states 2.1. Overview One would expect that states change as conditions change. To assess whether the Big Five states are indeed changeable, participants completed the Big Five state measures, then went through an induction designed to increase one of the states, and then again completed the scale assessing that state. 2.2. Methods 2.2.1. Participants A group of 143 participants was recruited from a university in the Southeastern United States and two work settings, a human relations organization and staff of a primary school, in the Northeastern United States. The average age of the participants was 33.92, SD=15.30; 104 were women and 39 were men. All participated on a volunteer basis and none were paid for their participation. 2.2.2. Procedure All participants completed the Big Five state measure described in Study 1 and then went through a state induction. Using the approach developed by Velten (1968), each of the inductions was designed to increase one of the states. Participants read 12 statements and were asked to spend 10 seconds to concentrate on and imagine what was described in each statement. To ensure that the domain encompassed by each of the Big Five characteristics was reflected by the statements, subfactors or facets which have been found for the Big Five dimensions (Costa & McCrae,
599
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
1992; McCrae & Costa, 1997b) provided the conceptual basis for the development of the statements for each set of Big Five characteristic induction items. For ethical reasons, the statements were designed to change the target state in the direction that most individuals would find desirable, e.g. towards greater agreeableness or more emotional stability. Following are examples of items for each of the inductions. Examples of surgency-state induction items are ‘‘Picture yourself as being the center of attention at a party’’ and ‘‘Picture yourself doing something adventurous.’’ Examples of agreeableness-state induction items are ‘‘Imagine praising another person’’ and ‘‘Focus on feeling that most of the people you deal with are honest and trustworthy.’’ Examples of conscientiousness-state induction items are ‘‘Picture yourself being an efficient and effective person at work’’ and ‘‘Picture yourself finishing a project that you have started.’’ Examples of emotional-stability-state induction items are ‘‘Focus on feeling that everything in your life is going right’’ and ‘‘Focus on thinking things through completely before making a decision.’’ Examples of openness-state induction items are ‘‘Imagine yourself completely absorbed listening to music’’ and ‘‘Picture yourself having intellectual conversations with others.’’ After reading the state-induction items, participants completed the state measure that matched their state-induction condition. For instance, participants who went through the agreeableness state induction completed the agreeableness state scale. 2.3. Result The internal reliability of each of the scales was again calculated. The Cronbach’s alphas were as follows: for the surgency items, 0.81; for the agreeableness items, 0.77; for the conscientiousness items, 0.78; for the openness items, 0.81, and for the emotional stability items, 0.83. Table 4 shows the mean scores from before to after the induction for each of the state measures. Each of the inductions changed the state scores in the expected direction; however the changes in conscientiousness and emotional stability did not reach significance. There was a significant increase in surgency, agreeableness, and openness from pre- to post-induction.
3. Discussion Two studies provided evidence that the Big Five dimensions may extend to states as well as traits. Study 1 examined the fit between present state items and the Big Five dimensions. The Table 4 State changes from pre to post-induction Pre-induction
Surgency Agreeableness Conscientiousness Emotional stability Openness
Post-induction
DF
M
SD
M
SD
t
p
21 22 27 20 25
47.71 58.64 52.70 49.00 52.08
11.45 8.02 10.14 13.26 13.88
53.86 62.09 54.26 52.60 54.00
10.33 7.84 11.13 12.69 13.60
3.53 2.86 1.39 1.47 2.37
0.002 0.009 ns ns 0.026
600
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
three relative fit indexes suggested a good fit between responses and the Big Five dimensions; the RMSEA, the population discrepancy statistic, suggested a marginal fit; and the chi square suggested a poor fit, as it tends to do with samples as large as this one (N=189; see Floyd & Widaman, 1995). Overall, one could reasonably describe the fit as acceptable. The inter-correlation of state factors ranged from 0.07 to 0.48, with a median of 0.28 and the intercorrelation of the state measures ranged from 0.05 to 0.37, with a median of 0.295. John and Srivastava (1999) found a similar pattern in a confirmatory factor analysis of trait measures of the Big Five with a median factor inter-correlation of 0.22 and a range of ‘‘below 0.20’’ (page 119) to 0.37. As one would expect, each Big Five state was more related to the corresponding Big Five trait than to other Big Five traits. Further, as expected on the basis of previous research, higher levels of state surgency were associated with higher levels of state positive mood, and higher levels of state emotional stability were associated with lower levels of state negative mood. Because one would expect states to be changeable, Study 2 used an experimental manipulation to attempt to change levels of the Big Five States. All states changed in the expected direction; however, only the changes in state surgency, agreeableness, and openness were statistically significant. The present studies suggest the usefulness of a hierarchical conceptualization of functioning in which the Big Five dimensions provide an organizational template. The findings of the current studies extend the previous findings of Borkenau and Ostendorf (1998), Fleeson (2001) and Nemanick and Munz (1997) regarding states and traits. Now evidence exists supporting the value of a Big Five conceptualization of functioning from present states, to daily average states, to enduring traits. Both biological and learning processes might explain the relationship between Big Five traits, towards the top of the hierarchy, and Big Five present states, towards the bottom of the hierarchy. First, some research (Jang et al., 1998; Loehlin et al., 1998) suggests that the Big Five dimensions have a biological basis. If this is the case then humans may have a predisposition to respond and organize behavior along the Big Five dimensions at different levels of functioning. In those who have a stronger predisposition towards a certain Big Five characteristic, for example towards greater conscientiousness, this predisposition might be manifested both at the trait and more weakly at the state level. Second, the concept of reciprocal determinism from social learning theory (Bandura, 1986) may provide a vehicle for explaining the link between Big Five characteristics at a higher and lower level in a hierarchical model. Situations and models may prompt states at the lower levels of the hierarchy. If situational exposure is long-term, the states at the lower levels of the hierarchy may help build a permanent predisposition, or higher-level trait. Conversely, the presence of a strong higher-level trait may make it more likely that cognitions and emotions arise that prime the state related to that trait, or that the trait sensitizes the individual to aspects of the environment that relate to that trait. The present studies also provide preliminary reliability and validity evidence for a measure of Big Five states derived from the Unipolar Markers (Goldberg, 1992; Saucier, 1994). The internal consistencies found for the five state scales ranged from 0.73 to 0.83 in two studies. Validity evidence was found in the convergent and divergent correlation patterns with trait Big Five measures. Also, the surgency and emotional stability state scales correlated as predicted with positive mood and negative mood respectively. Finally, three of the state scales, surgency, agreeableness, and openness, showed significant sensitivity to change targeted at these states. The other scales showed only non-significant change.
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
601
One un-hypothesized finding was that higher state conscientiousness was correlated significantly with higher positive mood state. Why this correlation occurred is unclear, but the association may be worthy of further research. It could be that a positive mood state makes a person more conscientious. This interpretation is consistent with prior findings that clinically depressed individuals tend to have lower trait conscientiousness (Anderson & McLean, 1997) and that mood enhancing events lead to better task performance (Nantais & Schellenberg, 1999), perhaps through increased conscientiousness. Alternatively, the correlation found in the present study could be due to a third variable influencing both state conscientiousness and state positive affect. It is also possible that state conscientiousness, by prompting more careful mood regulation, causes state positive mood. Some cautions are in order regarding the findings of the second study. First, it could be that the significant change in three of the five states apparently produced by the state imagery manipulation actually occurred as an effect of repeated assessment or as a results of the research participants trying to give the researchers the effect the manipulation seemed intended to produce. Second, because the effects of each manipulation were measured with only the scale assessing the targeted state, it is not possible to determine if all the states might have increased with an induction aimed at one. Future research could help clarify to what degree an extension of the Big Five to present states is theoretically or practically useful. Because Big Five states may be relatively easily manipulated compared to Big Five traits, the effect of the Big Five dimensions on various important realms of functioning might be experimentally examined. For example, experimental research might explore the effect of the Big Five states on behavior and cognitive processes. Experimental research might also explore what phenomena affect the Big Five states. For instance, how might schools, employers, and public health officials prompt the conscientious state likely to help students, employees, and other individuals conscientiously study, work, and engage in health-protective behaviors? Perhaps higher positive mood also prompts state surgency and higher negative mood prompts lower state emotional stability. Future research might also further define the levels of functioning in a Big Five hierarchical organization of functioning. For example, the relationships between trait facets (McCrae & Costa, 1997b) and Big Five states, and whether there are facets of states (breaking Big Five states into even smaller components), remain to be investigated. References Allport, G. (1961). Holt. New York: Reinhart and Winston. Anderson, K. W., & McLean, P. D. (1997). Conscientiousness in depression: Tendencies, predictive utility, and longitudinal stability. Cognitive Therapy and Research, 21, 223–238. Arbuckle, J. L., & Wothke, W. (1995). Amos 4.0: User’s Guide. Chicago: Small Waters. Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall. Borkenau, P., & Ostendorf, F. (1998). The Big Five as states: How useful is the Five-Factor Model to describe intraindividual variations over time. Journal of Research in Personality, 32, 202–221. Buss, D. M., & Craik, K. H. (1983). The act frequency approach to personality. Psychological Review, 90, 105–126. Catanzano, S. J. (1996). Negative mood regulation expectancies, emotional distress, and examination performance. Personality and Social Psychology Bulletin, 22, 1023–1059.
602
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
Costa, P. T., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO Personality Inventory. Psychological Assessment, 4, 5–13 20–22. Costa, P. T., & McCrae, R. R. (1996). Mood and personality in adulthood. In C. Magai, & S. H. McFadden (Eds.), Handbook of emotion, adult development, and aging (pp. 369–383). San Diego: Academic Press. Digman, J. M. (1990). Personality structure: emergence of the five-factor model. Annual Review of Psychology, 41, 417– 440. Fleeson, W. (2001). Toward a structure- and process-integrated view of personality: Traits as density distributions of states. Journal of Personality and Social Psychology, 80, 1011–1027. Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7, 286–299. Forgas, J. P. (1995). Mood and judgement: The Affect Infusion Model (AIM). Psychological Bulletin, 117, 39–66. Forgas, J. P., & Bower, G. H. (1987). Mood effects on person perception judgements. Journal of Personality and Social Psychology, 53, 53–60. Goldberg (1982). From Ace to Zombie: Some explorations in the language of personality. In: C. D. Spielberger, & J. N. Butcher (Eds.), Advances in personality assessment (Vol. 1, pp. 203–234). Erlbaum, Hillsdale, NJ. Goldberg, L. R. (1990). An alternative ‘‘description of personality’’: The Big-Five factor structure. Journal of Personality and Social Psychology, 59, 1216–1229. 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 personality traits. American Psychologist, 48, 26–34. Isen, A. M., & Levine, P. F. (1972). The effect of feeling good on helping: Cookies and kindness. Journal of Personality and Social Psychology, 21, 384–388. Isen, A. M., & Simmonds, S. F. (1978). The effect of feeling good on a helping task that is incompatible with good mood. Social Psychology, 41, 345–349. Jang, K. L., McCrae, R. R., Angleitner, A., Riemann, R., & Livesley, W. J. (1998). Heritability of facet-level traits in a cross-cultural twin sample: Support for a hierarchical model of personality. Journal of Personality and Social Psychology, 74, 1556–1565. John, O. P. (1990). factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In L. A. Pervin (Ed.), Handbook of personality: theory and research factor taxonomy: dimensions of personality in the natural language and in questionnaires (pp. 66–100). New York: Guilford. John, O. J., & 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 (pp. 102–138). New York: Guilford. Kishton, J., & Widaman, K. F. (1994). Unidimensional versus domain representative parceling of questionnaire items: An empirical example. Educational and Psychological Measurement, 54, 757–765. Loehlin, J. C. (1998). Latent variable models: An introduction to factor, path, and structural models (3d ed.). Mahwah, NJ: Erlbaum. Loehlin, J. C., McCrae, R. R., Costa, P. T., & John, O. P. (1998). Heritabilities of common and measure-specific components of the Big Five personality factors. Journal of Research in Personality, 32, 431–453. Lucas, R. E., & Fujita, F. (2000). Factors influencing the relation between extraversion and pleasant affect. Journal of Personality and Social Psychology, 79, 1039–1056. McCrae, R. R., & Costa, P. T. (1997a). Personality trait structure as a human universal. American Psychologist, 52, 509–516. McCrae, R. R., & Costa, P. T. (1997b). Stability and change in personality assessment: The Revised NEO Personality Inventory in the Year 2000. Journal of Personality Assessment, 68, 86–94. McCrae, R. R., & Costa, P. T. (1999). A five-factor theory of personality. In L. A. Pervin, & O. P. John (Eds.), Handbook of personality: theory and research (pp. 139–153). New York: Guilford. Menasco, M. B., & Hawkins, D. I. (1978). A field test of the relationship between cognitive dissonance and state anxiety. Journal of Marketing Research, 15, 650–655. Menzel, K. E., & Carrell, C. J. (1994). The relationship between preparation and performance in public speaking. Communication Education, 43, 17–27.
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
603
Nantais, K. M., & Schellenberg, E. G. (1999). The Mozart effect: An artifact of preference. Psychological Science, 10, 370–373. Nemanick, R. C., & Munz, D. C. (1997). Extraversion and neuroticism, trait mood and state affect. Journal of Social Behavior and Personality, 12, 1079–1092. Saucier, G. (1994). Mini-Markers: A brief version of Goldberg’s Unipolar Big-Five Markers. Journal of Personality Assessment, 63, 506–516. Saucier, G., & Goldberg, L. R. (1996). The language of personality: Lexical perspectives on the five-factor model. In J. S. Wiggins (Ed.), The five-factor model of personality: Theoretical perspectives (pp. 21–50). New York: Guilford Press. Saucier, G., Hampson, S. E., & Goldberg, L. R. (2000). Cross-language studies of lexical personality factors. In: S. E. Hampson (Ed.), Advances in personality psychology, Vol. 1 (pp. 1–36). Psychology Press, Philadelphia, PA. Spielberger, D. D., & Sydeman, S. J. (1994). State-Trait Anxiety Inventory and State-trait Anger Expression Inventory. In E. M. Maruish (Ed.), The use of psychological testing for treatment planning and outcome assessment (pp. 292–321). Hillsdale, NJ: Lawrence Erlbaum Associates. Teasdale, J. D., & Barnard, P. J. (1993). Affect, cognition, and change. Hove: Lawrence Erlbaum. Usala, P. D., & Hertzog, C. (1991). Evidence of differential stability of state and trait anxiety in adults. Journal of Personality and Social Psychology, 60, 471–479. Velten, E. (1968). A laboratory task for induction of mood states. Behavior Research and Therapy, 6, 475–482. Wakefield, J. C. (1989). Levels of explanation in personality theory. In D. Buss, & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 333–346). New York: Springer. Watson, D., & Clark, L. A. (1997). Measure and mismeasurement of mood: Recurrent and emergent issues. Journal of Personality Assessment, 68, 267–269. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070.
States reflecting the Big Five dimensions Nicola S. Schutte*,a, John M. Malouffa, Elida Segrerab, Amanda Wolfb, Larissa Rodgersb a
University of New England, School of Psychology, Armidale, NSW 2351 Australia Nova Southeastern University, Behavioral Sciences, Ft. Lauderdale, FL 33314, USA
b
Received 2 July 2001; received in revised form 15 December 2001; accepted 21 January 2002
Abstract Two studies explored the possibility that the Big Five dimensions, which extensive research has shown underlie most human traits, also provide a structure for transitory states. A confirmatory factor analysis showed an acceptable fit between responses on measures of transitory states and the Big Five dimensions. Further, the state measures of the Big Five dimensions had good internal consistency. As one would expect, each Big Five state was more related to the corresponding Big Five trait than to other Big Five traits. As expected on the basis of previous research, higher levels of state surgency were associated with higher levels of state positive mood, and higher levels of state emotional stability were associated with lower levels of state negative mood. Unexpectedly, state conscientiousness was also highly associated with state positive mood. Because one would expect states to be changeable, the second study used an experimental manipulation to attempt to change levels of the Big Five States. All states changed in the expected direction; however, only the changes in state surgency, agreeableness, and openness were statistically significant. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Big Five; States
1. States reflecting the Big Five dimensions A number of theories of human behavior (Allport, 1961; Buss & Craik, 1983; Costa & McCrae, 1992; Goldberg, 1993; Wakefield, 1989) emphasize the hierarchical organization of human functioning. According to these theories general and enduring characteristics are at the top of the hierarchy and more specific or passing characteristics are at the bottom of the hierarchy. These
* Corresponding author. E-mail address: [email protected] (N. S. Schutte). 0191-8869/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0191-8869(02)00031-4
592
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
characteristics at the bottom of the hierarchy are generally viewed as being less descriptive of an individual and as being prompted by certain situations or cognitive processes. A related conceptualization of human characteristics is their classification into traits or states (e.g. Spielberger & Sydeman, 1994; Watson, Clark, & Tellegen, 1988). In a hierarchical model, traits are conceptualized as higher-level and enduring characteristics, while states are lower level and less enduring characteristics. Trait anxiety, for instance, disposes individuals to feel chronic anxiety, while state anxiety is a situationally linked experience of anxiety that passes when the situation is no longer present (Spielberger & Sydeman 1994). A similar differentiation can be made between state and trait anger (Spielberger & Sydeman, 1994). Watson et al. (1988) studied characteristic (general) positive and negative affect and state (momentary) positive and negative mood. As one would expect, ratings of characteristic positive and negative affect were relatively unchanged over the course of an eight-week period, while ratings of momentary positive and negative mood showed some consistency, but more variability than characteristic affect. Similarly, Usala and Hertzog (1991) found that state anxiety was less stable over time than trait anxiety. Thus, one might conclude that these mood-related states are the result of characteristic affect interacting with situational influences. Characteristics such as states, which may be placed at the bottom of a hierarchical conceptualization of functioning, may be related to important outcomes. For example, state anxiety has been found to be associated with performance (Catanzano, 1996; Menzel & Carrell, 1994) and cognitive dissonance (Menasco & Hawkins, 1978). State mood that has been manipulated on the positive mood dimension has been found to be associated with helping behavior (Isen and Levine, 1972; Isen & Simmonds, 1978), memory (Teasdale & Barnard, 1993), perception (Forgas & Bower, 1987), and judgement (Forgas, 1995). State anxiety, state anger, and state positive and negative mood have been investigated. However, relatively little research has focused on other types of states. Five dimensions, surgency, agreeableness, conscientiousness, emotional stability, and openness, seem to underlie many characteristic traits (Goldberg, 1993; John & Srivastava, 1999; McCrae & Costa, 1999). These Big Five dimensions have been identified in many factor analytic studies (Costa & McCrae, 1992, 1996; Digman, 1990; John, 1990; John & Srivastava, 1999; McCrae & Costa, 1999), including cross-cultural studies (McCrae & Costa, 1997a; Saucier & Goldberg, 1996). Typically these studies have found that if individuals rate themselves or others on a wide variety of trait descriptors, five factors emerge. Different researchers have applied somewhat different terms to these factors, e.g. using the term ‘‘extraversion’’ instead of ‘‘surgency,’’ low ‘‘neuroticism’’ instead of ‘‘emotional stability,’’ and ‘‘intellect’’ or ‘‘imagination’’ instead of ‘‘openness’’ (Goldberg, 1992; Saucier & Goldberg, 1996). The consistent finding that many personality traits group into five dimensions might be in part due to biological predispositions to organize responses into these dimensions. Research showing moderate heritability of the Big Five traits and facets of the Big Five traits (e.g. Jang, McCrae, Angleitner, Riemann & Livesley, 1998; Loehlin, McCrae, Costa & John, 1998) supports the possibility of such a biological predisposition. In a review of research on the Big Five dimensions, John and Srivastava (1999) pointed out that as well as being important in understanding the organization of human functioning, the Big Five dimensions are related to important life outcomes. For example, research with adolescents has found that low agreeableness and low conscientiousness predict juvenile delinquency; neuroticism
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
593
and low conscientiousness predict internalizing psychopathologies; and conscientiousness and openness are associated with good school performance. Research with adults has shown that conscientiousness predicts good work performance and good health, low agreeableness and high neuroticism are associated with poor health, high agreeableness is associated with helping others, high extraversion predicts leadership, high neuroticism is associated with depression, and high openness is associated with creativity. In a factor analytic study, Borkenau and Ostendorf (1998) explored the relationship between daily states and the Big Five dimensions. In this study participants used adjective checklists to record their states for each day during 90 consecutive days. Two important findings were that there was substantial variability in the day-to-day responding of most participants and that the overall pattern of responding across the 90 days grouped into the Big Five dimensions. Similarly, using an experience sampling methodology, Fleeson (2001) found evidence for within-person variability, which indicated that situation-specific behavior may be grouped into the Big Five dimensions, as well as stable average tendencies for individuals, supporting the existence of stable Big Five traits. Nemanick and Munz (1997) examined the relationship between two Big Five traits and characteristics at different levels in a hierarchical conceptualization of personality. They conceptualized the Big-Five characteristic of extraversion as being at a high level of characteristic personality explanation, characteristic positive affectivity as being on an intermediate level, and affect on a given day as being on a relatively low level. Extraversion was associated more strongly with characteristic positive affectivity than with positive affect on a given day. Characteristic positive affectivity in turn was more strongly related to positive affect on a given day. A similar pattern was found for the relationship between the Big Five factor of neuroticism, conceptualized as being at a high level of explanation; characteristic negative affectivity, an intermediate characteristic; and negative affect on a given day. If personality characteristics group into five dimensions, and if these five dimensions are superordinate categories in a hierarchical organization of human functioning, then it is possible that the state level of functioning reflects these five dimensions. The research on daily states by Borkenau and Ostendorf (1998), density distributions of behaviors by Fleeson (2001) and the relationship between trait extraversion and neuroticism and daily mood states by Nemanick and Munz (1997) supports this view. The purpose of the current research was to investigate further whether there might be present states that correspond to the Big Five dimensions, these states’ relationship to the Big Five traits and to mood, and whether these states are indeed changeable as one would expect of states. A related purpose was to develop and validate present state measures of each of the Big Five characteristics. Study 1: Development of State Measures of the Big Five, Factor Structure and Internal Consistency of the States, Relationship Between the Big Five States and Traits, and Relationship Between the Big Five States and Positive and Negative State Mood. 1.1. Overview Saucier (1994) developed a short form of the Big-Five trait measure created by Goldberg (1992). We altered this short form to make it into present a state measure of the Big Five dimensions. This measure focused on immediate states. If the Big Five dimensions apply to states as well as to traits, one would expect the factor structure of immediate states to be similar to the
594
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
factor structure of traits. If there were a hierarchical relationship between traits and states, then one would expect each of the Big Five states to relate more closely to the corresponding Big Five trait than to any other Big Five trait. Further, because Nemanick and Munz (1997) found surgency (extraversion) to be associated with positive affectivity, which in turn was associated with daily positive mood, and neuroticism to be associated with negative affectivity, which in turn was associated with daily negative mood, one would expect a moderate positive relationship between present-state surgency and state positive mood and a moderate negative relationship between present-state emotional stability and state negative mood. 1.2. Method 1.2.1. Development of the present state measures of the big five characteristics Goldberg (1992) created a set of 100 adjectives for use in measuring the big-five dimensions: surgency, agreeableness, conscientiousness, emotional stability, and intellect. Applying factor analysis techniques to hundreds of English trait terms Goldberg (1982, 1990) obtained the usual Big Five factor pattern. After exploring the usefulness of bipolar items, Goldberg (1992) developed a 100-item set of unipolar adjectives (e.g. fearful, innovative) to measure the Big Five dimensions. He selected 20 items for each of the five dimensions on the basis of factor analysis results on hundreds of adjectives in prior research (Goldberg, 1982). Goldberg (1992) found that the Big Five dimensions on the Unipolar Marker scale correlated in expected ways with scores on the NEO Personality Inventory, which is a commonly used measure of the Big Five: Surgency correlated most highly with NEO extroversion, Agreeableness most highly with NEO agreeableness, Conscientiousness most highly with NEO conscientiousness, Emotional Stability most highly with low NEO neuroticism, and Intellect most highly with NEO openness. Saucier (1994) shortened the 100-item Unipolar Marker measure to 40 items, with eight items representing each dimension. He used the following criteria for item selection: (a) high factor loadings on the relevant dimension, (b) that an item be easily understood, and (c) high correlation with the full set of items for the dimension. We adapted the 40-item Unipolar Marker measure to a state measure of the Big-Five dimensions. On the original and brief Unipolar Marker measures of the Big-Five traits, respondents rate themselves as they are generally or characteristically. We modified these instructions to ask respondents to rate themselves at the present moment. The instructions were: ‘‘Describe yourself as you feel right now, that is, at the present moment. Describe yourself as you see yourself at the present time, not as you wish to be in the future or as you were in the past.’’ As was done on the original version of the measure, respondents then rated themselves on 40 adjectives using a 9-point scale on which ‘‘1’’ means extremely inaccurate and ‘‘9’’ means extremely accurate. The adjectives comprise five subscales, each of which represents one of the BigFive characteristics. Examples of adjectives from each of the subscales are as follows, surgency: ‘‘bold;’’ agreeableness: ‘‘kind;’’ conscientiousness: ‘‘efficient;’’ emotional stability: ‘‘relaxed;’’ and openness: ‘‘creative.’’ 1.2.2. Participants A sample of 189 participants was recruited from a university in the southeastern United States and from community settings, such as sports centers and workplaces, in the southeastern United
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
595
States. Participants’ average age was 23.26, SD=7.00; 132 were women and 57 were men. All participated on a voluntary basis and none were paid for their participation. 1.2.3. Procedure All participants completed the present-state measure of the Big Five dimensions. Sixty-eight of the participants also completed the Big-Five Trait Inventory (John & Srivastava, 1999), which has evidence of good reliability and validity. Items are in phrase form and individuals rate themselves on how well they fit the description given in each item; a typical example of an item is ‘‘helpful and unselfish with others’’ for the agreeableness scale. The other 121 participants completed the state version of the Positive and Negative Affect Scales (PANAS; Watson & Clark, 1997; Watson et al., 1988), which assesses positive and negative mood states as independent dimensions and which has evidence of good reliability and validity. Items on these scales are in adjective form and typical items are ‘‘enthusiastic’’ for positive affect and ‘‘distressed’’ for negative affect. 1.3. Results 1.3.1. Descriptive statistics The average scores and standard deviations on the Big Five state measures were as follows: surgency, M=47.84, SD=11.22; agreeableness, M=54.41, SD=10.02; conscientiousness, M=51.29, SD=10.75; openness, M=50.73, SD=10.51; emotional stability, M=44.33, SD=11.38. Average scores on the Big-Five Trait Inventory were as follows: extraversion, M=27.46, SD=6.38; agreeableness, M=35.87, SD=5.16; conscientiousness, M=35.12, SD=5.79; openness, M=37.06; SD=6.80, neuroticism, M=26.71, SD=6.82. The average scores and standard deviations for mood were M=31.80, SD=10.38 for positive mood and M=16.29, SD=7.15 for negative mood. 1.3.2. Confirmatory factor analysis The Big Five present state measures are theoretically separate factors, although they may intercorrelate to some degree as do Big Five trait measures. In order to test whether the items on the Big Five state measures actually load on the Big Five dimensions, we performed a maximum likelihood confirmatory factor analysis. Following the guidelines set out by Floyd and Widaman (1995), we created two parcels of four items for each hypothesized state factor. This reduces the instability associated with the use of individual items (see Floyd & Widaman, 1995). Following the suggestions of Kishton and Widaman (1994), we randomly assigned four scale items to each parcel. Fig. 1 shows the model. To avoid nonidentifiability of the model, we followed the suggestions of Arbuckle and Wothke (1995) and constrained to unity the path from every latent error variable to its measured variable (parcel) and the path from each latent construct (e.g. surgency) to its first measured variable (parcel). In the initial testing of the model, the solution indicated that agreeableness parcel 2 had negative variance. This common problem in structural equation modeling can be eliminated by setting the variance for that variable to 0 (Loehlin, 1998). Using Amos software, we followed the suggestions of Arbuckle (personal communication), who created the software, and set the variance for that parcel to 0.000001 to be as close to 0 as the software allows. Setting that parameter had no appreciable effect on the fit indicators. The model was evaluated with the chi-square statistic, which is a test of absolute fit that is difficult to pass in analyses based on large samples or in models with many observed variables (Floyd & Widaman,
596
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
Fig. 1. Confirmatory factor analysis model of the Big Five states. Large ellipses represent latent factors, boxes represent measured variables comprised of item parcels, and small ellipses represent error of the measured variables.
1995); three relative fit statistics, the Bollen incremental fit index (IFI), the Tucker Lewis index (TLI), and the Bentler comparative fit index (CFI); and one measure based on population discrepancies, the root mean square error of approximation (RMSEA). There are no universally accepted standards for determining the appropriateness of a model, so we used the commonly followed rules (see Lucas & Fujita, 2000) that a model fits well if the chisquare is nonsignificant, the three fit indices are greater than 0.90, and the RMSEA does not exceed 0.05 The results were that the chi-square (df=25) equaled 62.7, P<0.01; the three fit indices were all at 0.99, and the RMSEA was <0.09. Hence, the model (a) failed to meet the standard of the absolute fit statistic, chi-square, (b) met the standard of the IFI, (c) met the standard of the TLI, (d) met the standard of the CFI, and (e) almost met the RMSEA population discrepancy standard, falling in an area, <0.09, that is usually called marginal (e.g. Loehlin, 1998). See Fig. 1 for the parameter estimates. 1.3.3. Internal consistency and relationship of the states with each other Table 1 shows the internal consistency of each of the Big Five state measures and the relationships of the states with each other. The Cronbach alpha coefficients for each of the state measures
597
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603 Table 1 Alpha coefficients of the Big Five state measures and correlations between the Big Five states (N=189)
Surgency Agreeableness Conscientiousness Openness Emotional Stability
Surgency
Agreeableness
Conscientiousness
Openness
Emotional stability
(0.78) 0.20a 0.22a 0.30a 0.15
(0.73) 0.37a 0.33a 0.29a
(0.79) 0.32a 0.30a
(0.75) 0.05
(0.75)
Note. Alpha coefficients are presented on the diagonal; correlations are presented below the diagonal. a p < 0.01.
showed reasonable internal consistency for the eight items comprising each of the scales. The states showed low to moderate correlations with each other. The median correlation was 0.295, with a range of 0.05 to 0.37. 1.3.4. Relationship between Big-Five states and Big-Five traits Each of the Big-Five states was most highly associated with the corresponding Big-Five trait (see Table 2). While all of the correlations between states and their corresponding traits were statistically significant, none, with possibly the exception of surgency, had so much common variance that they seemed to be measuring the same construct. The median within-dimension (e.g. state agreeableness and trait agreeableness) correlation was 0.58, with a range of 0.39 to 0.77. The median inter-dimension correlation was 0.09, with a range of 0.00 to 0.28. 1.3.5. Relationship of state surgency and state emotional stability with mood As predicted, higher levels of state surgency were significantly associated with higher levels of state positive mood. Surprisingly, higher levels of state conscientiousness were also strongly associated with more state positive mood. As predicted higher levels of state emotional stability were significantly associated with lower levels of state negative mood. See Table 3 for the correlations.
Table 2 Correlations between Big Five states and Big Five traits (n=68) States Traits
Surgency
Surgency Agreeableness Conscientiousness Openness Emotional stability
0.77a 0.02 0.17 0.28 0.14
Agreeableness 0.00 0.42a 0.08 0.13 0.24
Conscientiousness 0.15 0.06 0.65a 0.05 0.23
Openness 0.10 0.09 0.02 0.58a 0.01
Emotional stability 0.16 0.23 0.09 0.09 0.39a
Note. Hypothesized correlations are underlined. Traits were measured with the Big Five trait inventory. Responses on trait emotional stability were scored so that high scores indicated greater emotional stability. a p < 0.01.
598
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
Table 3 Correlations between the Big Five states and positive and negative mood (N=121) States
Positive mood
Surgency Agreeableness Conscientiousness Openness Emotional stability
0.45a 0.33a 0.56a 0.36a 0.31a
Negative mood 0.18 0.23 0.29a 0.01 0.46a
Note. Hypothesized correlations are underlined. a p< 0.01.
Regression analyses investigated the relative value of the five states in predicting positive and negative mood. With an R squared of 0.32, F (5,115)=54.75, P<0.001, state conscientiousness best predicted positive mood. State surgency significantly added to this prediction with an R squared increase of 0.10, F (5,115)=19.89, P<0.001. With an R squared 0.21, F (5,115)=29.79, P<0.001, state emotional stability was the best predictor of negative mood. None of the other states added significantly to these predictions.
2. Study 2: Changeability of the Big Five states 2.1. Overview One would expect that states change as conditions change. To assess whether the Big Five states are indeed changeable, participants completed the Big Five state measures, then went through an induction designed to increase one of the states, and then again completed the scale assessing that state. 2.2. Methods 2.2.1. Participants A group of 143 participants was recruited from a university in the Southeastern United States and two work settings, a human relations organization and staff of a primary school, in the Northeastern United States. The average age of the participants was 33.92, SD=15.30; 104 were women and 39 were men. All participated on a volunteer basis and none were paid for their participation. 2.2.2. Procedure All participants completed the Big Five state measure described in Study 1 and then went through a state induction. Using the approach developed by Velten (1968), each of the inductions was designed to increase one of the states. Participants read 12 statements and were asked to spend 10 seconds to concentrate on and imagine what was described in each statement. To ensure that the domain encompassed by each of the Big Five characteristics was reflected by the statements, subfactors or facets which have been found for the Big Five dimensions (Costa & McCrae,
599
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
1992; McCrae & Costa, 1997b) provided the conceptual basis for the development of the statements for each set of Big Five characteristic induction items. For ethical reasons, the statements were designed to change the target state in the direction that most individuals would find desirable, e.g. towards greater agreeableness or more emotional stability. Following are examples of items for each of the inductions. Examples of surgency-state induction items are ‘‘Picture yourself as being the center of attention at a party’’ and ‘‘Picture yourself doing something adventurous.’’ Examples of agreeableness-state induction items are ‘‘Imagine praising another person’’ and ‘‘Focus on feeling that most of the people you deal with are honest and trustworthy.’’ Examples of conscientiousness-state induction items are ‘‘Picture yourself being an efficient and effective person at work’’ and ‘‘Picture yourself finishing a project that you have started.’’ Examples of emotional-stability-state induction items are ‘‘Focus on feeling that everything in your life is going right’’ and ‘‘Focus on thinking things through completely before making a decision.’’ Examples of openness-state induction items are ‘‘Imagine yourself completely absorbed listening to music’’ and ‘‘Picture yourself having intellectual conversations with others.’’ After reading the state-induction items, participants completed the state measure that matched their state-induction condition. For instance, participants who went through the agreeableness state induction completed the agreeableness state scale. 2.3. Result The internal reliability of each of the scales was again calculated. The Cronbach’s alphas were as follows: for the surgency items, 0.81; for the agreeableness items, 0.77; for the conscientiousness items, 0.78; for the openness items, 0.81, and for the emotional stability items, 0.83. Table 4 shows the mean scores from before to after the induction for each of the state measures. Each of the inductions changed the state scores in the expected direction; however the changes in conscientiousness and emotional stability did not reach significance. There was a significant increase in surgency, agreeableness, and openness from pre- to post-induction.
3. Discussion Two studies provided evidence that the Big Five dimensions may extend to states as well as traits. Study 1 examined the fit between present state items and the Big Five dimensions. The Table 4 State changes from pre to post-induction Pre-induction
Surgency Agreeableness Conscientiousness Emotional stability Openness
Post-induction
DF
M
SD
M
SD
t
p
21 22 27 20 25
47.71 58.64 52.70 49.00 52.08
11.45 8.02 10.14 13.26 13.88
53.86 62.09 54.26 52.60 54.00
10.33 7.84 11.13 12.69 13.60
3.53 2.86 1.39 1.47 2.37
0.002 0.009 ns ns 0.026
600
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
three relative fit indexes suggested a good fit between responses and the Big Five dimensions; the RMSEA, the population discrepancy statistic, suggested a marginal fit; and the chi square suggested a poor fit, as it tends to do with samples as large as this one (N=189; see Floyd & Widaman, 1995). Overall, one could reasonably describe the fit as acceptable. The inter-correlation of state factors ranged from 0.07 to 0.48, with a median of 0.28 and the intercorrelation of the state measures ranged from 0.05 to 0.37, with a median of 0.295. John and Srivastava (1999) found a similar pattern in a confirmatory factor analysis of trait measures of the Big Five with a median factor inter-correlation of 0.22 and a range of ‘‘below 0.20’’ (page 119) to 0.37. As one would expect, each Big Five state was more related to the corresponding Big Five trait than to other Big Five traits. Further, as expected on the basis of previous research, higher levels of state surgency were associated with higher levels of state positive mood, and higher levels of state emotional stability were associated with lower levels of state negative mood. Because one would expect states to be changeable, Study 2 used an experimental manipulation to attempt to change levels of the Big Five States. All states changed in the expected direction; however, only the changes in state surgency, agreeableness, and openness were statistically significant. The present studies suggest the usefulness of a hierarchical conceptualization of functioning in which the Big Five dimensions provide an organizational template. The findings of the current studies extend the previous findings of Borkenau and Ostendorf (1998), Fleeson (2001) and Nemanick and Munz (1997) regarding states and traits. Now evidence exists supporting the value of a Big Five conceptualization of functioning from present states, to daily average states, to enduring traits. Both biological and learning processes might explain the relationship between Big Five traits, towards the top of the hierarchy, and Big Five present states, towards the bottom of the hierarchy. First, some research (Jang et al., 1998; Loehlin et al., 1998) suggests that the Big Five dimensions have a biological basis. If this is the case then humans may have a predisposition to respond and organize behavior along the Big Five dimensions at different levels of functioning. In those who have a stronger predisposition towards a certain Big Five characteristic, for example towards greater conscientiousness, this predisposition might be manifested both at the trait and more weakly at the state level. Second, the concept of reciprocal determinism from social learning theory (Bandura, 1986) may provide a vehicle for explaining the link between Big Five characteristics at a higher and lower level in a hierarchical model. Situations and models may prompt states at the lower levels of the hierarchy. If situational exposure is long-term, the states at the lower levels of the hierarchy may help build a permanent predisposition, or higher-level trait. Conversely, the presence of a strong higher-level trait may make it more likely that cognitions and emotions arise that prime the state related to that trait, or that the trait sensitizes the individual to aspects of the environment that relate to that trait. The present studies also provide preliminary reliability and validity evidence for a measure of Big Five states derived from the Unipolar Markers (Goldberg, 1992; Saucier, 1994). The internal consistencies found for the five state scales ranged from 0.73 to 0.83 in two studies. Validity evidence was found in the convergent and divergent correlation patterns with trait Big Five measures. Also, the surgency and emotional stability state scales correlated as predicted with positive mood and negative mood respectively. Finally, three of the state scales, surgency, agreeableness, and openness, showed significant sensitivity to change targeted at these states. The other scales showed only non-significant change.
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
601
One un-hypothesized finding was that higher state conscientiousness was correlated significantly with higher positive mood state. Why this correlation occurred is unclear, but the association may be worthy of further research. It could be that a positive mood state makes a person more conscientious. This interpretation is consistent with prior findings that clinically depressed individuals tend to have lower trait conscientiousness (Anderson & McLean, 1997) and that mood enhancing events lead to better task performance (Nantais & Schellenberg, 1999), perhaps through increased conscientiousness. Alternatively, the correlation found in the present study could be due to a third variable influencing both state conscientiousness and state positive affect. It is also possible that state conscientiousness, by prompting more careful mood regulation, causes state positive mood. Some cautions are in order regarding the findings of the second study. First, it could be that the significant change in three of the five states apparently produced by the state imagery manipulation actually occurred as an effect of repeated assessment or as a results of the research participants trying to give the researchers the effect the manipulation seemed intended to produce. Second, because the effects of each manipulation were measured with only the scale assessing the targeted state, it is not possible to determine if all the states might have increased with an induction aimed at one. Future research could help clarify to what degree an extension of the Big Five to present states is theoretically or practically useful. Because Big Five states may be relatively easily manipulated compared to Big Five traits, the effect of the Big Five dimensions on various important realms of functioning might be experimentally examined. For example, experimental research might explore the effect of the Big Five states on behavior and cognitive processes. Experimental research might also explore what phenomena affect the Big Five states. For instance, how might schools, employers, and public health officials prompt the conscientious state likely to help students, employees, and other individuals conscientiously study, work, and engage in health-protective behaviors? Perhaps higher positive mood also prompts state surgency and higher negative mood prompts lower state emotional stability. Future research might also further define the levels of functioning in a Big Five hierarchical organization of functioning. For example, the relationships between trait facets (McCrae & Costa, 1997b) and Big Five states, and whether there are facets of states (breaking Big Five states into even smaller components), remain to be investigated. References Allport, G. (1961). Holt. New York: Reinhart and Winston. Anderson, K. W., & McLean, P. D. (1997). Conscientiousness in depression: Tendencies, predictive utility, and longitudinal stability. Cognitive Therapy and Research, 21, 223–238. Arbuckle, J. L., & Wothke, W. (1995). Amos 4.0: User’s Guide. Chicago: Small Waters. Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall. Borkenau, P., & Ostendorf, F. (1998). The Big Five as states: How useful is the Five-Factor Model to describe intraindividual variations over time. Journal of Research in Personality, 32, 202–221. Buss, D. M., & Craik, K. H. (1983). The act frequency approach to personality. Psychological Review, 90, 105–126. Catanzano, S. J. (1996). Negative mood regulation expectancies, emotional distress, and examination performance. Personality and Social Psychology Bulletin, 22, 1023–1059.
602
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
Costa, P. T., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO Personality Inventory. Psychological Assessment, 4, 5–13 20–22. Costa, P. T., & McCrae, R. R. (1996). Mood and personality in adulthood. In C. Magai, & S. H. McFadden (Eds.), Handbook of emotion, adult development, and aging (pp. 369–383). San Diego: Academic Press. Digman, J. M. (1990). Personality structure: emergence of the five-factor model. Annual Review of Psychology, 41, 417– 440. Fleeson, W. (2001). Toward a structure- and process-integrated view of personality: Traits as density distributions of states. Journal of Personality and Social Psychology, 80, 1011–1027. Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7, 286–299. Forgas, J. P. (1995). Mood and judgement: The Affect Infusion Model (AIM). Psychological Bulletin, 117, 39–66. Forgas, J. P., & Bower, G. H. (1987). Mood effects on person perception judgements. Journal of Personality and Social Psychology, 53, 53–60. Goldberg (1982). From Ace to Zombie: Some explorations in the language of personality. In: C. D. Spielberger, & J. N. Butcher (Eds.), Advances in personality assessment (Vol. 1, pp. 203–234). Erlbaum, Hillsdale, NJ. Goldberg, L. R. (1990). An alternative ‘‘description of personality’’: The Big-Five factor structure. Journal of Personality and Social Psychology, 59, 1216–1229. 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 personality traits. American Psychologist, 48, 26–34. Isen, A. M., & Levine, P. F. (1972). The effect of feeling good on helping: Cookies and kindness. Journal of Personality and Social Psychology, 21, 384–388. Isen, A. M., & Simmonds, S. F. (1978). The effect of feeling good on a helping task that is incompatible with good mood. Social Psychology, 41, 345–349. Jang, K. L., McCrae, R. R., Angleitner, A., Riemann, R., & Livesley, W. J. (1998). Heritability of facet-level traits in a cross-cultural twin sample: Support for a hierarchical model of personality. Journal of Personality and Social Psychology, 74, 1556–1565. John, O. P. (1990). factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In L. A. Pervin (Ed.), Handbook of personality: theory and research factor taxonomy: dimensions of personality in the natural language and in questionnaires (pp. 66–100). New York: Guilford. John, O. J., & 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 (pp. 102–138). New York: Guilford. Kishton, J., & Widaman, K. F. (1994). Unidimensional versus domain representative parceling of questionnaire items: An empirical example. Educational and Psychological Measurement, 54, 757–765. Loehlin, J. C. (1998). Latent variable models: An introduction to factor, path, and structural models (3d ed.). Mahwah, NJ: Erlbaum. Loehlin, J. C., McCrae, R. R., Costa, P. T., & John, O. P. (1998). Heritabilities of common and measure-specific components of the Big Five personality factors. Journal of Research in Personality, 32, 431–453. Lucas, R. E., & Fujita, F. (2000). Factors influencing the relation between extraversion and pleasant affect. Journal of Personality and Social Psychology, 79, 1039–1056. McCrae, R. R., & Costa, P. T. (1997a). Personality trait structure as a human universal. American Psychologist, 52, 509–516. McCrae, R. R., & Costa, P. T. (1997b). Stability and change in personality assessment: The Revised NEO Personality Inventory in the Year 2000. Journal of Personality Assessment, 68, 86–94. McCrae, R. R., & Costa, P. T. (1999). A five-factor theory of personality. In L. A. Pervin, & O. P. John (Eds.), Handbook of personality: theory and research (pp. 139–153). New York: Guilford. Menasco, M. B., & Hawkins, D. I. (1978). A field test of the relationship between cognitive dissonance and state anxiety. Journal of Marketing Research, 15, 650–655. Menzel, K. E., & Carrell, C. J. (1994). The relationship between preparation and performance in public speaking. Communication Education, 43, 17–27.
N.S. Schutte et al. / Personality and Individual Differences 34 (2003) 591–603
603
Nantais, K. M., & Schellenberg, E. G. (1999). The Mozart effect: An artifact of preference. Psychological Science, 10, 370–373. Nemanick, R. C., & Munz, D. C. (1997). Extraversion and neuroticism, trait mood and state affect. Journal of Social Behavior and Personality, 12, 1079–1092. Saucier, G. (1994). Mini-Markers: A brief version of Goldberg’s Unipolar Big-Five Markers. Journal of Personality Assessment, 63, 506–516. Saucier, G., & Goldberg, L. R. (1996). The language of personality: Lexical perspectives on the five-factor model. In J. S. Wiggins (Ed.), The five-factor model of personality: Theoretical perspectives (pp. 21–50). New York: Guilford Press. Saucier, G., Hampson, S. E., & Goldberg, L. R. (2000). Cross-language studies of lexical personality factors. In: S. E. Hampson (Ed.), Advances in personality psychology, Vol. 1 (pp. 1–36). Psychology Press, Philadelphia, PA. Spielberger, D. D., & Sydeman, S. J. (1994). State-Trait Anxiety Inventory and State-trait Anger Expression Inventory. In E. M. Maruish (Ed.), The use of psychological testing for treatment planning and outcome assessment (pp. 292–321). Hillsdale, NJ: Lawrence Erlbaum Associates. Teasdale, J. D., & Barnard, P. J. (1993). Affect, cognition, and change. Hove: Lawrence Erlbaum. Usala, P. D., & Hertzog, C. (1991). Evidence of differential stability of state and trait anxiety in adults. Journal of Personality and Social Psychology, 60, 471–479. Velten, E. (1968). A laboratory task for induction of mood states. Behavior Research and Therapy, 6, 475–482. Wakefield, J. C. (1989). Levels of explanation in personality theory. In D. Buss, & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 333–346). New York: Springer. Watson, D., & Clark, L. A. (1997). Measure and mismeasurement of mood: Recurrent and emergent issues. Journal of Personality Assessment, 68, 267–269. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070.