Content uploaded by Don Munro
Author content
All content in this area was uploaded by Don Munro on May 01, 2020
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Values in action scale and the Big 5:
An empirical indication of structure
Craig Macdonald, Miles Bore
*
, Don Munro
School of Psychology, University of Newcastle, Callaghan, NSW 2308, Australia
Available online 4 December 2007
Abstract
Within this study we used self-report measures completed by 123 undergraduate students from an Australian university
to investigate the validity of Peterson and Seligman’s [Peterson, C., & Seligman, M. E. P (2004). Character strengths and
virtues. New York: Oxford.] classification system of 24 character strengths and six virtues. We also looked at how the 24
character strengths relate to the Five Factor Model of personality and to a measure of social desirability. Using a second
order factor analysis of the 24 character strengths, we found that these 24 character strengths did not produce a factor
structure consistent with the six higher order virtues as proposed by Peterson and Seligman [Peterson, C., & Seligman,
M. E. P (2004). Character strengths and virtues. New York: Oxford.]. Instead, the 24 character strengths were well repre-
sented by both a one and four factor solution. Patterns of significant relationships between each of the 24 character
strengths, the one and four factor solutions and the Five Factor Model of personality were found. The results have impli-
cations for [Peterson, C., & Seligman, M. E. P (2004). Character strengths and virtues. New York: Oxford.] classification.
Crown copyright !2007 Published by Elsevier Inc. All rights reserved.
Keywords: Values; VIA scale; Big 5; Five Factor Model
1. Introduction
The field of positive psychology has the goal of helping people achieve an above normal or optimal level
of functioning, leading to a happier existence (Gable & Haidt, 2005; Wallis, 2005). Wallis (2005) suggests
that this is because much of psychological practice and theory has focused on helping people to recover
from a diminished level of functioning, and has largely neglected helping people achieve a higher level
of functioning.
Two of the main proponents of positive psychology are Seligman and Csikszentmihalyi (2000), who see
positive psychology encompassing subjective experience, individual traits, and societal interactions. With
regard to the area of individual differences, Peterson and Seligman (2004) have developed a hierarchy of posi-
tive psychological character strengths. The hierarchy consists of 24 specific character strengths that are seen as
0092-6566/$ - see front matter Crown copyright !2007 Published by Elsevier Inc. All rights reserved.
doi:10.1016/j.jrp.2007.10.003
*
Corresponding author. Fax: +61 2 49216906.
E-mail address: Miles.Bore@newcastle.edu.au (M. Bore).
Available online at www.sciencedirect.com
Journal of Research in Personality 42 (2008) 787–799
www.elsevier.com/locate/jrp
Author's personal copy
the psychological ingredients that make up six ‘‘virtues”. These virtues are situated at a higher level of abstrac-
tion than character strengths, and are likened to constructs proposed by philosophers and religious figures
over many centuries. These six virtues and their associated character strengths are displayed in Table 1.
Peterson and Seligman (2004) analysed different religious, cultural and legal texts from around the world in
an attempt to achieve a universal classification for character strengths, and only included character strengths
and virtues that were found to be ubiquitous (Dahlsgaard, Peterson, & Seligman, 2005).
To measure and assess the 24 character strengths, Peterson and Seligman (2004) also developed the Virtues
In Action Scale (VIA). The VIA is a self-assessment measure of character strength requiring respondents to
rate how likely they are to participate in certain behaviours that are representative of the different character
strengths. It is important to note that the scale does not directly measure the six virtues they describe; these are
only linked conceptually to the character strengths by Peterson and Seligman (2004).
In addition to developing their classification system, Peterson and Seligman (2004) have also suggested how
their classification of character strengths and virtues is related to, but distinct from, already established the-
ories of values. For example, Peterson and Seligman (2004) see their classification of character strengths and
virtues as being related to Maslow’s (1973) idea of self-actualised individuals, the Five Factor Model (FFM) of
personality (McCrae & John, 1992; Costa & McCrae, 1994), Cawley’s virtue factors (Cawley, Martin, & John-
son, 2000), Buss’ evolutionary ideas about what is attractive in a mate [i.e. what character traits are essential
for survival and propagation, (Botwin, Buss, & Shackelford, 1997; Shackelford, Schmitt, & Buss, 2005)], and
Schwartz’s (1992) Universal Values.
Some research into establishing the validity of these claims has begun. Haslam, Bain, and Neal (2004)
found that both Schwartz’s (1992) Universal Values and the Five Factor Model (FFM) of personality were
conceptually linked to the 24 character strengths. However, as these constructs were defined and subsequently
measured by only one or two terms that were ranked and grouped together by participants on the basis of
conceptual likeness, more thorough research is needed before we can draw any firm conclusions.
Peterson and Seligman (2004) acknowledge that there are some clear correspondences between their
classification and the FFM. For example, Neuroticism could be seen as the conceptual opposite of Hope,
Table 1
Peterson and Seligman’s (2004) classification of character strengths and virtues
Virtue Character strengths
Wisdom and Knowledge Creativity
Curiosity
Open-mindedness
Love of Learning
Perspective
Courage Bravery
Persistence
Integrity
Vitality
Humanity Love
Kindness
Social Intelligence
Justice Citizenship
Fairness
Leadership
Temperance Forgiveness and Mercy
Humility/Modesty
Prudence
Self-control
Transcendence Appreciation of Beauty
Gratitude
Hope
Humour
Spirituality
Adapted from Table 1.1 in Peterson and Seligman (2004), pp 29–30.
788 C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799
Author's personal copy
and Extroversion could be a key to Leadership (Peterson & Seligman, 2004). They also show how their clas-
sification system as a whole corresponds to the FFM, by conceptually equating a factor analysis of their 24
character strengths to the five factors, although the FFM does not account for all of their classification
(see Table 2).
It is important to note that Peterson and Seligman (2004) did not empirically correlate their value strength
factors with the five factors of the FFM but only make these links conceptually. They also acknowledge that
the five factors found in their factor analysis of the VIA do not exactly reflect their hypothesized hierarchical
classification of the six virtues. Moreover, it is also interesting to note that only 19 of the 24 character strengths
are reported, raising the question of where the other five would load. This ambiguity brings into doubt both
the hierarchical link between the 24 character strengths and six virtues and the conceptual links between the
FFM and the character strengths. The proposed relationships are further brought into doubt when one
reviews other research into character strengths and the FFM. This research suggests that some of the character
strengths are related to combinations of FFM traits and not individual traits. For instance, creative people
have been shown to be high in Openness (O) and low in Agreeableness [A (King, Walker, & Broyles,
1996)]; honest and humble people have been found to be high in Agreeableness (Ashton & Lee, 2005) and also
high in Conscientiousness [C (Paunonen, 2003)]; Brose, Rye, Lutz-Zois, and Ross (2005) found forgiveness to
be both negatively correlated with Neuroticism (N) and positively correlated with Agreeableness and some-
times Extroversion (E). Also, a meta-analysis of a number of different studies by Suroglou (2000) found that
religiosity was related to high A, C and (to some extent) E.
Although most of the theoretical correlates predicted by Peterson and Seligman (2004) are reflected in this
research, there are often multiple predictors present as shown in the studies above. Our research pursues this
idea by investigating which combinations of FFM traits, rather than a single trait, are best related to each of
the 24 character strengths found within Peterson and Seligman’s (2004) classification.
In consideration of the factor analysis carried out by Peterson and Seligman (2004) it is important to reit-
erate that their results did not support their theory of particular combinations of character strengths as being
represented as the higher order virtues. Consequently, this issue is also examined, with the expectation that a
factor analysis of the 24 character strengths will not produce the six virtues proposed by Peterson and Selig-
man (2004).
A further area of interest we considered was whether the VIA is effected by social desirability. Peterson and
Seligman (2004) state that the 24 character strengths are socially desirable constructs themselves and as a
result the VIA should not be affected by social desirability. We take this to mean that the VIA will not be
affected by individual differences in socially desirable responding. Although this may be true, the opposing
argument could also be made: as the 24 character strengths are socially desirable constructs, the VIA will
be highly affected by social desirability.
To summarise, the aim of this study is to further the understanding of Peterson and Seligman’s (2004) clas-
sification of 24 character strengths by examining the relationships between the character strengths themselves
and their relationship to the Five Factor Model of personality. To achieve this, the 24 character strengths are
Table 2
Factors found in the virtues in action scale and their correlates
Character strengths contained within the factor Name given to
factor
Reflected virtue Theoretical FFM correlate
Fairness, Humility, Mercy, Prudence Strengths of
restraint
Temperance Conscientiousness
Creativity, Curiosity, Love of Learning,
Appreciation of Beauty
Intellectual
strengths
Wisdom and
Knowledge
Openness
Kindness, Love, Leadership, Teamwork,
Playfulness
Interpersonal
strengths
Humanity and
Justice
Agreeableness
Bravery, Hope, Self-control, Zest Emotional
strengths
Courage Opposite of Neuroticism (Emotional
Stability)
Gratitude, Spirituality Theological
strengths
Transcendence No FFM correlate
Constructed from information found in Peterson and Seligman (2004), pp. 632–633.
C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799 789
Author's personal copy
first factor analysed with the expectation that the extracted factors will not neatly represent the six virtues pro-
posed by Peterson and Seligman (2004). The 24 character strengths will then be compared to the FFM in order
to investigate the relationships between them. It is hypothesized that there will not be a one-to-one relation-
ship between for the majority of the character strengths and the FFM personality traits. Rather it is expected
that most of the character strengths will show relationships to more than one of the FFM constructs. Follow-
ing Peterson and Seligman’s (2004) view, we expect that the VIA will not be influenced by socially desirable
responding.
2. Method
2.1. Participants
The participants for this study were 123 first year psychology students enrolled in an Australian university.
There were 28 males and 86 females, with 9 participants not indicating their gender. Ages ranged from 18 to 57
years, with a mean of 21.51 years and a standard deviation of 6.57 years. All participants volunteered to take
part in the study and were given course credit for doing so.
2.2. Materials
The materials used were question booklets containing a battery of ten psychometric tests and answer book-
lets. Four versions of the question booklets were used with the questionnaires presented in different orders in
each version to account for any possible order effects. The psychometric tests not relevant to the current
research will be reported elsewhere.
To measure the 24 character strengths, Peterson and Seligman’s (2004) Virtues In Action Scale (VIA) was
used. This test uses a five-point Likert scale (A = very much like me, B = like me, C = neutral, D = unlike me,
E = very much unlike me) to measure how frequently one perceives oneself as exhibiting certain behaviours.
These behaviours are representative of the 24 individual character strengths. The actual VIA scale used was a
213-item edition obtained from the IPIP web site (International Personality Item Pool, 2001). Items were com-
piled in the test booklet pseudo-randomly to ensure that items representing individual character strengths and
positively and negatively worded items were distributed evenly throughout the questionnaire. Using data
obtained from over 150,000 adult respondents completing the VIA over the Internet, the measure was found
to have acceptable internal reliability (all alphas >.7) and temporal reliability (test/retest >.7) (Peterson &
Seligman, 2004).
The measure of the Five Factor Model of Personality was Goldberg’s (1999) Big Five scale obtained from
the IPIP web site (International Personality Item Pool, 2001). This measure has 20 self-report items per scale
and participants respond using a four-point Likert scale (F = definitely false, f = false on the whole, t = true
on the whole, T = definitely true). This instrument has been found to be both a reliable and valid measure of
the FFM (see Goldberg, 1999; Goldberg et al., 2006).
The 20 item Marlowe-Crowne Social Desirability scale (Strahan & Gerbasi, 1972) was also included in the
battery in order to examine and potentially account for social desirability effects. Strahan and Gerbasi
reported internal reliabilities ranging from.73 to.83 for the 20 item short-form of the scale. The scale uses a
true/false response option for each question. The scale is balance with half the items being reverse scored.
2.3. Procedure
Testing was done in multiple testing sessions. Upon entering the testing room the participants were ran-
domly given one of the four differently ordered questionnaire booklets and its accompanying answer booklet.
Participants sat where they chose and were instructed to complete all the tests presented the question booklet
in the order they appeared. They were also told that all information recorded was confidential. All tests were
answered in a separate answer booklet using either a pen or pencil. Both booklets were collected and partic-
ipants were allowed to leave when they had finished. The time taken to complete the entire battery was
between 90 and 120 min.
790 C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799
Author's personal copy
3. Results
Each questionnaire was first assessed for missing data; 18 item non-responses from the VIA scale and 17
from the FFM scale were replaced with the middle response of their corresponding Likert scale. Two VIA
respondents and one FFM scale respondent were removed for not responding correctly. Each measure was
scored and descriptive statistics and histograms were generated. These were then inspected for normality
and were all found to be within acceptable limits. T-tests were calculated between mean scores for each
FFM scale and norms derived by the second author (MB) from data collected for teaching purposes from
undergraduate psychology students (n= 565). No significant differences were found (p> .05). VIA norms
have not been published therefore no comparison was possible. In addition, t-tests were calculated to assess
differences in age and gender; and all tests conducted were non-significant. Order effects possibly resulting
from the different presentation order of the tests were also found to be non-significant.
Internal reliabilities for all the measures were then assessed via the generation of Cronbach Alpha coeffi-
cients. These were compared to alpha coefficients reported in the literature for each of the individual measures
(see Table 3).
The alpha coefficient for all 213 VIA scale items taken together (reverse scored where appropriate) was .96.
This showed that the behaviours related to the character strengths were responded to in a highly consistent
Table 3
Alpha coefficients for the VIA scale, the IPIP FFM scale, and the social desirability scale compared to the expected alpha coefficients
obtained from published findings
Character strength Our sample Previously reported No. items
Appreciation of Beauty .70 .77 8
Capacity for Love .68 .70 9
Citizenship/Teamwork .62 .78 9
Curiosity .83 .78 10
Equity/Fairness .80 .70 9
Forgiveness/Mercy .85 .76 9
Gratitude .76 .79 8
Hope/Optimism .58 .73 8
Humor/Playfulness .86 .84 9
Industry/Perseverance/Persistence .83 .81 8
Integrity/Honesty/Authenticity .56 .72 9
Judgment/Open-mindedness .80 .80 9
Kindness/Generosity .75 .72 10
Leadership .43 .77 7
Love of Learning .74 .77 10
Modesty/Humility .68 .70 9
Originality/Creativity .83 .85 8
Perspective/Wisdom .76 .75 9
Prudence .57 .73 9
Self-regulation/Self-control .68 .75 11
Social/Personal/Emotional Intel .75 .76 7
Spirituality/Religiousness .89 .91 7
Valor/Bravery/Courage .61 .75 10
Zest/Enthusiasm/Vitality .80 .78 9
Total character strength .96 n.a 213
IPIP Big 5
Agreeableness .88 .91 20
Conscientiousness .90 .88 20
Extroversion .92 .88 20
Neuroticism .91 .91 20
Openness .87 .90 20
Marlowe-Crowne Social Desirability .74 .73 to .83 20
Notes: For published alpha reliabilities see: VIA, Peterson and Seligman (2004); IPIP Big 5, Goldberg et al. (2006); Marlowe-Crowne
Social Desirability, Strahan and Gerbasi (1972).
C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799 791
Author's personal copy
fashion regardless of the separate scales with which they are theoretically associated. This overall measure was
included in subsequent analyses as a variable called Total Character Strength.
A second order factor analysis of the 24 character strengths derived from the VIA scale was then carried out
using a Principle Components analysis with a Varimax rotation. An initial extraction revealed five compo-
nents with an eigenvalue >1, however, the scree plot (shown in Fig. 1) indicated that three components might
also yield a simple solution.
The loadings of the character strengths in the five factor solution were not easily interpretable and many
high cross-loadings between the factors were evident. This was also observed when using an oblique (Direct
Oblimin) rotation. A four factor solution was then generated (see Table 4) and found to be more interpretable.
The four factors were tentatively labeled Positivity, Intellect, Conscientiousness and Niceness. The only nota-
ble reduction in communality between the five and four factor solutions was with the variable of Spirituality/
Religiosity. This four factor structure also had the clearest pattern of correlations with the FFM constructs
(see Table 4).
Social Desirability scores were correlated with each of the VIA character strength scores and the five factor
scores (also shown in Table 4). The correlations indicated that the factors were differentially related to Social
Desirability. Factors 2 and 3 (Intellect and Conscientiousness) were not significantly related to Social Desir-
ability scores, however, Factor 1 (Positivity) was weakly and positively related to Social Desirability and Fac-
tor 3 (Niceness) moderately positively related. The correlations between Social Desirability and each of the 24
character strengths reflect this pattern, with the six character strengths that loaded on Factor 4 (Niceness) all
correlating significantly with Social Desirability.
The two and three factor solutions were also run but were found to produce pronounced high cross-load-
ings and disjointed relationships to the FFM scales. Therefore these solutions were not considered further. A
single component solution was produced on which all the character strengths except Modesty and Prudence
loaded strongly (Table 5). However, eight variables showed very low communalities (<.3) and are therefore
not well represented by this solution.
Given the significant correlations between Social Desirability and several of the VIA character strength
scores, the data was reanalysed after ‘centering’ the 24 character strength scores for each participant. This
was done by dividing each character strength score by the number of items that made up each score, calcu-
lating the mean score for each participant across all 24 character strengths and then subtracting this mean
from each participant’s score for each character strength. The effect of centering is to remove the influence
of mean response set differences between participants (but not individual variability); in this case, individual
differences in the tendency to agree more strongly with socially desirable items.
Fig. 1. Scree plot for the 24 character strength raw scores.
792 C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799
Author's personal copy
A second order Principle Components factor analysis of the centered scores of the 24 character strengths
derived from the VIA scale was conducted. The initial extraction revealed nine factors with an eigenvalue >1
(see Fig. 2). Several solutions were examined, but none were clearly interpretable. Given the elbow in the Scree
Plot at factor 6 a five factor solution is presented here as an example (see Table 6). We could arrive at no
meaningful interpretation of this factor structure, particularly with regard to some of the negative loadings
observed. The correlations observed between social desirability and each of the centred character strength
scores appeared to be systematically reduced compared to the same correlations with character strength
raw scores reported above. Thus, the technique of centering reduced but did not eliminate the influence of
social desirability.
Stepwise regression analyses were conducted with the FFM traits and social desirability as predictor vari-
ables and the raw and the centralised character strength scores as criterion variables. After adjustment to
account for family-wise error a working significance level of p6.01 was used. Only values for the constructs
that had a pvalue 601 are presented in the results of this analysis (see Table 7).
Table 4
Varimax rotated four factor solution of 24 character strength scores and correlations with social desirability and Big 5 scores
Variable Factor 1 Factor 2 Factor 3 Factor 4 Communality SocDes correlation
Positivity
Citizenship/Teamwork .78 .64 .13
Capacity for Love .74 .58 .19
*
Hope/Optimism .73 .68 .26
**
Humor/Playfulness .63 !.43 .60 .18
Zest/Enthusiasm/Vitality .62 !.40 !.48 .77 .28
**
Leadership .50 !.44 .38 .58 .15
Intellect
Originality/Creativity !.78 .69 .02
Appreciation of Beauty !.72 .55 .15
Curiosity .49 !.61 .69 .32
**
Love of Learning !.60 !.43 .61 .15
Social/Personal/Emotional Intel .55 !.58 .67 .23
*
Perspective/Wisdom .35 !.53 !.47 .65 .12
Valor/Bravery/Courage .42 !.44 .46 .22
*
Conscientiousness
Self-regulation/Self-control !.75 .62 .12
Industry/Perseverance/Persist !.73 .66 .17
Judgment/Open-mindedness !.59 .34 .58 !.05
Integrity/Honesty/Authenticity !.58 .39 .57 .33
***
Prudence !.53 .50 .60 .16
Niceness
Modesty/Humility .75 .72 .28
**
Equity/Fairness .36 !.31 .64 .68 .53
***
Kindness/Generosity .35 .64 .65 .45
***
Forgiveness/Mercy .35 .62 .53 .55
***
Spirituality/Religiousness .50 .34 .25
**
Gratitude .42 !.41 .45 .59 .29
**
Proportion of Variance .19 .16 .13 .13 .61
Factor correlations
Social Desirability .24
**
!.04 !.06 .46
***
Agreeableness .20
*
!.08 .14 .57
***
.60
***
Conscientiousness .12 .14 .71
***
.15 .18
*
Extraversion .71
***
.26
**
!.06 !.10 .20
*
Neuroticism !.50
***
!.17 !.12 !.30 !.27
**
Openness .03 .68
***
.11 .04 .16
Notes: Loadings <.3 not shown. SocDes = Social Desirability.
For correlations: "p< .05; ""p< .01; """p< .001.
C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799 793
Author's personal copy
Table 5
One factor solution of 24 character strength scores and correlations with social desirability and Big 5 scores
Character strength Factor 1 Communality
Zest/Enthusiasm/Vitality .81 .65
Curiosity .79 .63
Perspective/Wisdom .76 .57
Social/Personal/Emotional Intelligence .75 .56
Gratitude .74 .55
Hope/Optimism .74 .54
Equity/Fairness .72 .52
Kindness/Generosity .70 .49
Humor/Playfulness .68 .46
Leadership .68 .46
Industry/Perseverance/Persistence .63 .39
Integrity/Honesty/Authenticity .63 .39
Valor/Bravery/Courage .63 .39
Capacity for Love .57 .32
Citizenship/Teamwork .57 .33
Originality/Creativity .56 .31
Love of Learning .54 .29
Forgiveness/Mercy .53 .28
Appreciation of Beauty .52 .27
Self-regulation/Self-control .49 .24
Judgment/Open-mindedness .45 .20
Spirituality/Religiousness .39 .16
Prudence .24 .06
Modesty/Humility .19 .04
Proportion of Variance .38 .38
Factor correlations
Social Desirability .37
***
Agreeableness .37
***
Conscientiousness .52
***
Extroversion .51
***
Neuroticism !.45
***
Openness .45
***
Notes: For correlations "p< .05; ""p< .01; """p< .001.
Fig. 2. Scree plot for the 24 character strengths centred scores.
794 C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799
Author's personal copy
The results of this analysis found that no consistent pattern of predictors was evident. For example, in the
raw score analysis, Neuroticism was only predictive of Hope/Optimism and Valour/Bravery/Courage while
Spirituality/Religiosity was only weakly predicted by Agreeableness. Social Desirability was found to be a sig-
nificant predictor of just 3 of the 24 character traits. Conducting the stepwise regression analysis on the char-
acter strength centred scores did produce somewhat different predictor patterns but no consistency in the
pattern of predictors was apparent.
4. Discussion
As hypothesised, the results of a second order Principal Components factor analysis of the 24 character
strengths were not consistent with Peterson and Seligman’s (2004) theory of how they relate to the six virtues.
For their theory to have been supported the 24 character strengths needed to have produced a clean six com-
ponent solution, whereas only five components with eigenvalues greater than one were found. At first glance
the five component solution appears to be similar to the five factor solution obtained by Peterson and Selig-
man (2004), but it is apparent on closer inspection that they are made up of different clusters of character
Table 6
Varimax rotated four component solution of centred scores and correlations with social desirability and Big 5 scores
Variable Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Communality SocDes correlation
Curiosity .75 .65 .12
Prudence !.70 !.36 .64 .09
Zest/Enthusiasm/Vitality .67 .31 !.32 .67 .08
Modesty/Humility !.59 !.47 !.30 .66 .03
Judgment/Open-mindedness !.59 .32 .54 !.32
***
Hope/Optimism .55 !.41 .50 .01
Forgiveness/Mercy !.72 .54 .44
***
Equity/Fairness !.72 .54 .43
***
Kindness/Generosity !.55 .38 .27
**
Perspective/Wisdom .47 .30 !.20
*
Social/Personal/Emotional Intel .62 .47 !.04
Leadership .58 !.31 .53 !.21
*
Self-regulation/Self-control !.57 .46 !.13
Industry/Perseverance/Persist .43 !.56 .51 !.04
Humor/Playfulness .51 !.30 .48 !.04
Originality/Creativity .36 .44 .35 !.40 .67 !.21
*
Citizenship/Teamwork !.75 .59 !.17
Love of Learning .66 .47 !.10
Capacity for Love !.63 .43 !.06
Appreciation of Beauty .30 .39 .34 !.06
Gratitude .70 .55 .09
Spirituality/Religiousness .69 .50 .12
Integrity/Honesty/Authenticity !.34 !.42 .42 .05
Valor/Bravery/Courage !.38 .21 !.07
Proportion Of Variance .12 .11 .10 .10 .07 .50
Factor Correlations
Social Desirability .25
**
!.49
***
!.09 .06 .14
Agreeableness .16 !.57
***
!.20
*
.04 .11
Conscientiousness .10 .18
*
!.49
***
.17 .06
Extroversion .49
***
.14 .28
**
!.33
***
.07
Neuroticism !.49
***
.04 .04 .13 .07
Openness .31
***
.09 .24
**
.48
***
!.10
Notes: Loadings <.3 not shown. SocDes = Social Desirability.
For correlations: "p< .05; ""p< .01; """p< .001.
C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799 795
Author's personal copy
strengths (compare Table 2 with Table 4). Furthermore, none of the clusters of character strengths contained
within the components were completely representative of the clusters needed to make up any of the six virtues,
thus providing additional evidence against Peterson and Seligman’s (2004) theoretical relationships.
Many of the VIA character strengths loaded strongly on multiple components making interpretability ques-
tionable. Similar problems of cross-loading may have been the reason that Peterson and Seligman (2004) orig-
inally left out five of the character strengths when reporting the results of their factor analysis. Even so, they
must have found cross-loadings different to those contained in our component solution, because many of the
character strengths remaining in Peterson and Seligman’s (2004) solution showed cross-loadings in ours.
In an attempt to find a solution that was representative of independent higher order constructs, two, three
and four component solutions of the 24 character strengths were produced. Each of these solutions also had
multiple character strengths cross-loading on different factors. The two and three factor solutions were unin-
terpretable, whereas the four-factor solution made conceptual sense and is discussed below. An oblique rota-
tion was also carried out to allow the factors to correlate more strongly with each other, but the results did not
dramatically enhance the interpretability of the solution.
Lastly a single component solution was generated. Only two character strengths did not load strongly: Pru-
dence and Modesty. Several others were also poorly represented, as indicated by the squared loadings. How-
Table 7
Stepwise regression analysis with FFM scales and social desirability as predictor variables of raw and centred character strength scores
Character strength Proportion of raw score accounted for Proportion of centred score accounted for
A C E N O Soc
Des
A C E N O Soc
Des
Appreciation of Beauty — — — — 27.6
**
— 4.41
*
— — — 10.1
**
—
Capacity for Love — — 24.97
**
— — — — 8.01
*
6.49
**
— 4.79
*
—
Citizenship/Teamwork — — 34.7
**
— — — — 7.53
**
17.49
**
— 13.59
**
—
Curiosity — 5.06
*
19.06
**
— 11.28
**
— — — — — 5.62
*
—
Equity/Fairness 30.46
**
— — — — 5.52
*
20.02
**
— — 5.17
*
— 3.79
*
Forgiveness/Mercy 44.89
**
— — — — 2.66
*
33.43
**
3.4
*
————
Gratitude — 11.24
**
6.55
*
—————————
Hope/Optimism — — 3.21
*
35.51
**
— — — — — 18.18
**
——
Humour/Playfulness — — 34.73
**
— 4.22
**
— — 8.22
**
20.43
**
———
Industry/Perseverance/
Persistence
— 58.67
**
— — — — — 37.66
**
— 3.7
*
——
Integrity/Honesty/
Authenticity
— 17.28
**
— — — 6.44
*
— — 5.28
*
———
Judgment/Open-
mindedness
— 22.83
**
2.8
*
— 4.58
*
— — 6.82
**
15.26
**
— — 7.11
*
Kindness/Generosity 20.67
**
— 3.79
*
— — — 6.98
**
— — 7.34
*
——
Leadership — — 19.75
**
— — — — 6.95
*
————
Love of Learning — 9.22
**
— — 25.67
**
— — — 12.04
**
— 8.01
**
—
Modesty/Humility 22.85
**
— 17.79
**
— — — 11.39
**
— 32.12
**
— 6.35
**
—
Originality/Creativity 34.2
**
— — — — — 10.97
**
— — — 17.8
**
—
Perspective/Wisdom — 21.82
**
— — 13.5
**
— 7.13
*
— — — 5.07
*
—
Prudence — 18.8
**
11.21
**
— — — — 4.72
*
29.41
**
2.9
*
6.66
**
—
Self-control — 29.02
**
— — — — — 5.82
*
————
Social/Personal/
Emotional Intel
— — 32.9
**
— 6.46
**
— — — 15.46
**
8.42
**
——
Spirituality/
Religiousness
5.72
*
———————————
Valour/Bravery/
Courage
— — — 16.59
**
5.84
*
— 6.96
**
— — 6.77
*
——
Zest/Enthusiasm/
Vitality
— 25.99
**
17.76
**
— — — — 5.37
**
10.78
**
———
Total Character
Strength
— 29.56
**
7.13
**
— 12.08
**
4.61
*
Note: ""p< .001, "p6.01.
796 C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799
Author's personal copy
ever, the idea that the character strengths are best represented by one overarching factor was further sup-
ported by the very high Cronbach Alpha coefficient of .96 that was obtained for the Total Character Strength
score. Although this variable is not the one suggested for the VIA scale by Peterson and Seligman (2004), it
may provide a useful addition to a theory of character.
In order to further investigate the VIA structure we centered the data and again conducted factor analyses.
None of the resulting factor solutions were supportive of the proposed relationship between the 24 character
strengths and the six virtues. Therefore, regardless of which factor solution might be considered most appro-
priate from these results, none of them supports at the next level of abstraction the six virtues as proposed by
Peterson and Seligman (2004). But that does not mean that these six virtues are meaningless. Like the 24 char-
acter strengths, they were originally found ubiquitously in cross-cultural and cross-generational literature. The
interpretation of the virtues as higher order representations of a large number of character strengths was
purely theoretical. The results of the factor analyses within this study, and even the results of Peterson and
Seligman’s (2004) own factor analyses, reveals that their theory does not hold up under empirical investiga-
tion. The question now open for consideration is what to do with the six virtues Peterson and Seligman
(2004) found to be ubiquitous in various forms of literature. Potentially the six virtues could be better dealt
with as a stand-alone classification with a separate instrument needing to be developed to measure them, or
simply treated as additional character strengths themselves that could be added to Peterson and Seligman’s
classification of character strengths.
Another question raised by the factor analyses is how to interpret the overarching factor that was found in
the raw data factor structure but not found (or minimized) in the centered scores factor structure. Centering
the data appeared to removed the influence of the large first factor. This then raises the question of what this
factor was representing. Centering is generally thought to remove the influence of socially desirable respond-
ing. However, one suggestion we make here is that social desirability in this instance is itself a character
strength (rather than a response set) with anti-social tendencies being its antithesis. As such, it would seem
inappropriate to remove a critical aspect of the construct the VIA sets out to measure.
The most simple interpretation might be that this factor is representing some sort of ‘‘goodness”as a
whole, thus raising the question of whether a person possessing large amounts of one character strength is
also likely to possess large amounts of multiple character strengths. Another interpretation may be that
because these character strengths are measured via self-assessment they might be subject to self-enhancing
halo effects. Therefore, self-perception of one’s overall character might be what is being measured by this
factor.
The correlations between the Marlow-Crowne Social Desirability scores and the character strengths and
factor scores partly support these suggestions. However, Social Desirability was found to be significantly
related to some character strengths but not all character strengths. This clearly does not support Peterson
and Seligman’s (2004) assertion that the VIA is free of social desirability effects because all the items of the
VIA are socially desirable. Rather the differential relationships observed suggest that some character strengths
are more socially desirable than others. Perhaps not surprisingly all the character traits that loaded on the fac-
tor of Niceness were all significantly related to social desirability scores.
A further complication is that centering the character strength scores reduced but did not remove the influ-
ence of social desirability. This finding, together with the differential relationships between social desirability
scores and the character strengths and factor scores, and the presence of a possible single overriding factor,
strongly suggest that research into the predictive validity of the VIA is required. Findings from such research
might be able to indicate which scores are predictive of positive behaviour: raw character strength scores, cen-
tered scores or scores with social desirability partialed out.
Within considering various factor analyses of the VIA it is important not to lose sight of the individual
character strengths. Given that the majority of the 24 constructs had alpha coefficients over .6 (as shown in
Table 3) and that they are all highly recognized constructs within society, one should not ignore them as indi-
vidual constructs. The only character strength with very low internal consistency was Leadership. The reason
for this may have been the young age of the majority of participants and their lack of opportunities to exhibit
leadership. Prudence, Integrity and Hope also had reliabilities of less than .60. One may argue, for example,
that Prudence is not a concept that is familiar to the current generation of younger people. This leads to the
suggestion that the cultural context of character needs to be considered.
C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799 797
Author's personal copy
As hypothesised, some of the predictors of the character strengths consisted of more than one FFM con-
struct, which brings into doubt Peterson and Seligman’s (2004) idea that each character strength should be
represented by one FFM construct. Furthermore, some FFM scales were only weakly related to the character
strengths. In particular, Neuroticism was only predictive (negatively) of Hope/Optimism and Valour/Bravery/
Courage. A similar trend was seen with the regressions conducted on the centralized data. Within this, some
differences in predictors were seen between the two analyses although the stronger predictors remained con-
stant between the two data sets. Moreover, changes between the data sets did not show any consistent pattern,
with some FFM traits increasing in predictiveness when moving from the original to the centralized data sets,
whereas others decreased. A better understanding of what latent construct is being removed when centering
the data is needed before this can be discussed further. What is evident from both of these data sets is that
as hypothesized, for a majority of the character strengths there was more than one FFM predictor.
None of the factor analyses were entirely ‘clean’ in terms of cross-loadings. The cross-loadings suggest that
such character strengths might be positioned in factor space between two orthogonal factors or between two
(or more) of the FFM traits as suggested in the AB5C personality model (Hofstee, de Raad, & Goldberg,
1992). Other strengths, such as spirituality, might be related to a sixth ‘big trait’ (e.g., see Piedmont,
1999).
1
Another way of relating the FFM to character strength is to look at how the FFM relates to the fac-
tors obtained in each solution attempted. When this was done with the four factor solution of the character
strength scores (not centered) one could tentatively interpret the factors as representing higher traits such as
Positivity, Intellect, Conscientiousness and Niceness. The first component was positively correlated with
Extroversion and negatively with Neuroticism, with the character strengths reflecting the idea of being positive
(or engaged) and active in life (and the workplace). The second component reflects character strengths mainly
related to the activity of the mind (except Valour/Bravery/Courage) and was positively correlated with Open-
ness. The third component was highly positively correlated with Conscientiousness and reflected self-regula-
tion of one’s attitudes and actions. The fourth component (Niceness) was highly positively correlated with
Agreeableness and the individual character strengths are all related to interaction between individuals.
A limitation of our research was the small sample size. It would be interesting to see how the 213 individual
items of the VIA scale factored together as opposed to the 24 character strengths analysed here. To do this,
over 1000 participants would be needed to allow for at least five participants per item. From this analysis one
could further investigate the influence of the overarching factor. Furthermore it is important that relationships
between the VIA and other psychological constructs continue to be investigated in order to elaborate the con-
struct validity of the character strengths themselves. Peer ratings of character strength could also be used in
future research to provide some validation for the VIA. In our study, both a single factor and a four factor
solution of the non-centered VIA character strength scores were most interpretable perhaps suggesting that
the latter represents the influence of the Big 5 on character strengths whereas the former may represent some
other global influence.
References
Ashton, M. C., & Lee, K. (2005). Honesty–humility, the big five and the five-factor model. Journal of Personality, 73(5), 1321–1353.
Botwin, M. D., Buss, D. M., & Shackelford, T. K. (1997). Personality and mate preferences: Five factors in mate selection and marital
satisfaction. Journal of Personality, 65(1), 107–136.
Brose, L. A., Rye, M. S., Lutz-Zois, C., & Ross, S. R. (2005). Forgiveness and personality traits. Personality and Individual Differences, 39,
35–46.
Cawley, M. J., Martin, J. E., & Johnson, J. A. (2000). A virtues approach to personality. Personality and Individual Differences, 28(5),
997–1013.
Costa, P. T., Jr., & McCrae, R. R. (1994). Stability and change in personality from adolescence through adulthood. In C. F. Halverson,
Jr., R. P. Kohnstamm, & R. P. Martin (Eds.), The developing structure of temperament and personality from infancy to adulthood
(pp. 138–149). New Jersey: Lawrence Erlbaun Associates.
Dahlsgaard, K., Peterson, C., & Seligman, M. E. P. (2005). Shared virtue: The convergence of valued human strengths across culture and
history. Review of General Psychology, 9(3), 203–213.
Gable, S. L., & Haidt, J. (2005). What (and why) is positive psychology? General Review of Psychology, 9(2), 103–110.
1
We are indebted to one of the reviewers for these two suggestions.
798 C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799
Author's personal copy
Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the lower-level facets of several five-factor
models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.). Personality Psychology in Europe (vol. 7, pp. 7–28). Tilburg,
The Netherlands: Tilburg University Press.
Goldberg, L. R., Johnson, J. A., Eber, H. W., Hogan, R., Ashton, M. C., Cloninger, C. R., et al. (2006). The international personality item
pool and the future of public-domain personality measures. Journal of Research in Personality, 40, 84–96.
Hofstee, W. K. B., de Raad, B., & Goldberg, L. R. (1992). Integration of the big five and circumplex approaches to trait structure. Journal
of Personality and Social Psychology, 63, 146–163.
Haslam, N., Bain, P., & Neal, D. (2004). The implicit structure of positive characteristics. Personality and Social Psychology Bulletin, 30(4),
529–541.
International Personality Item Pool (2001).A scientific collaboratory for the development of advanced measures of personality traits and
other individual differences. Available from <http://ipip.ori.org>, accessed:19/03/06.
King, L. A., Walker, L. M., & Broyles, S. J. (1996). Creativity and the five-factor model. Journal of Research in Personality, 30, 189–203.
Paunonen, S. V. (2003). Big five factors of personality and replicated predictions of behavior. Journal of Personality and Social Psychology,
84(2), 411–424.
Peterson, C., & Seligman, M. E. P. (2004). Character strengths and virtues. New York: Oxford.
Piedmont, R. L. (1999). Does spirituality represent the sixth factor of personality? Spiritual transcendence and the five-factor model.
Journal of Personality, 67, 985–1013.
Seligman, M. E. P., & Csikszentmihalyi, M. (2000). Positive psychology: An introduction. American Psychologist, 55, 5–14.
Shackelford, T. K., Schmitt, D. P., & Buss, D. M. (2005). Universal dimensions of human mate preferences. Personality and Individual
Differences, 39(2), 447–458.
Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries.
Advances in Experimental Social Psychology, 25, 1–65.
Strahan, R., & Gerbasi, K. (1972). Short, homogenous version of the Marlowe-Crowne social desirability scale. Journal of Clinical
Psychology, 28, 191–193.
Suroglou, V. (2000). Religion and the five factors of personality: A meta-analytic review. Personality and Individual Differences, 32, 15–25.
Maslow, A. H. (1973). Dominance, self-esteem, self-actualization: Germinal papers of A.H. Maslow. Monterey: Brooks/Cole.
McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60(2), 175–215.
Wallis, C. (2005). The new science of happiness. Reprinted from Time, January 17.
C. Macdonald et al. / Journal of Research in Personality 42 (2008) 787–799 799
