J Happiness Stud

DOI 10.1007/s10902-014-9583-7

RESEARCH PAPER

Predicting Psychological and Subjective Well-Being

from Personality: Incremental Prediction from 30 Facets
Over the Big 5

Jeromy Anglim • Sharon Grant

Ó Springer Science+Business Media Dordrecht 2014

Abstract This study investigated the relationship between the Big 5, measured at factor
and facet levels, and dimensions of both psychological and subjective well-being. Three
hundred and thirty-seven participants completed the 30 Facet International Personality
Item Pool Scale, Satisfaction with Life Scale, Positive and Negative Affectivity Schedule,
and Ryff’s Scales of Psychological Well-Being. Cross-correlation decomposition presented
a parsimonious picture of how well-being is related to personality factors. Incremental
facet prediction was examined using double-adjusted r2 confidence intervals and semi-
partial correlations. Incremental prediction by facets over factors ranged from almost
nothing to a third more variance explained, suggesting a more modest incremental pre-
diction than presented in the literature previously. Examination of semi-partial correlations
controlling for factors revealed a small number of important facet-well-being correlations.
All data and R analysis scripts are made available in an online repository.

Keywords Subjective well-being  Psychological well-being  Personality  Big 5 

Personality facets

1 Introduction

Understanding the relationship between personality and well-being is of fundamental

importance for both theoretical and applied reasons. For one, the relationship of personality
to well-being may shed light on the temporal stability of well-being. It can also be helpful
to understand the role of personality when designing interventions targeted at increasing

J. Anglim (&)
School of Psychology, Deakin University, 221 Burwood Highway, Burwood, VIC 3125, Australia
e-mail: jeromy.anglim@deakin.edu.au

S. Grant
Faculty of Health, Arts and Design, Swinburne University of Technology, Melbourne, VIC, Australia

123
 J. Anglim, S. Grant

well-being. However, at present, it is unclear whether facet-level or factor-level personality

analysis is superior for understanding well-being. Researchers need an unbiased assess-
ment of this issue, given the reduction in parsimony that results when moving away from
broad models of personality such as the Five Factor Model. In addition, little is currently
known about the relationship between psychological well-being and personality facets. In
what follows, we describe subjective well-being (SWB), psychological well-being (PWB)
and the Five Factor Model (FFM), and review previous work on personality and well-being
before outlining the objectives for the current study, which illustrates a new approach for
assessing the incremental variance of personality facets over factors.

1.1 Subjective and Psychological Well-Being

The literature on SWB is expansive (Diener 2009; Diener et al. 1999, 2003; Lucas and
Diener 2008). SWB has been defined and measured in a variety of ways and can include
happiness and quality of life measures, but a common approach is to operationalize SWB
as a composite of satisfaction with life, high positive affect, and low negative affect (Deci
and Ryan 2008; Diener 1984; Lucas et al. 1996). This operationalization is adopted in the
current study.
The construct of PWB was developed in response to a perceived failure of SWB to
capture various humanistic concepts of well-being related to identity, meaning, and
relatedness (McGregor and Little 1998; Ryan and Deci 2001; Ryff and Keyes 1995). Ryff
(1989) proposed a six dimensional model of PWB composed of autonomy, environmental
mastery, personal growth, positive relations, purpose in life, and self-acceptance. The same
author also developed a measure of these six dimensions that has subsequently been used
in several studies (for a review see, Keyes et al. 2002). Studies have shown that envi-
ronmental mastery and self-acceptance overlap substantially with SWB (Compton 1998;
Keyes et al. 2002; McGregor and Little 1998; Ryff and Keyes 1995) but that the other
dimensions are more distinct, correlating only moderately with SWB measures (Compton
1998; Keyes et al. 2002; McGregor and Little 1998; Ryff and Keyes 1995).
While situational factors lead to short-term fluctuation and in some cases long-term
change in well-being, substantial research has supported a dispositional perspective of
well-being. Building on ideas such as the ‘‘hedonic treadmill’’ (Brickman and Campbell
1971). Headey and Wearing (1992) proposed that while life events can temporarily alter
well-being, well-being has a set point which varies between individuals. Genetic and twin
studies have established a hereditary basis for the stable component of well-being (Bou-
chard and Loehlin 2001; Lykken and Tellegen 1996; Weiss et al. 2008). Furthermore, a
17 year longitudinal study (Fujita and Diener 2005) found that satisfaction with life
showed substantial stability over time, albeit at about half the level of personality traits.
Thus, personality traits provide an important means of understanding the stability in well-
being.

1.2 Personality

Historically, trait research began with a proliferation of traits which was later followed by
various attempts at data reduction and eventually a movement to the Big 5 (Costa and
McCrae 1992; Goldberg 1993; McCrae and John 1992), typically labeled neuroticism,
extraversion, openness, conscientiousness, and agreeableness. More recently, researchers
responding to the success of the Big 5 have called for even higher level factor models
(Digman 1997; Musek 2007) and more detailed lower level models (Paunonen and Ashton

123
Incremental Prediction from 30 Facets Over the Big 5

2001; Paunonen and Jackson 2000). Several test publishers have developed facet-level
models of the Big 5, which aim to capture both the Big 5 and their constituent lower-level
facets (Costa and McCrae 1992; Goldberg 1992; John and Srivastava 1999). Despite some
discontent over the dominance of the Big 5, the taxonomy provides an organizing
framework for understanding different traits.

1.3 Personality Factors and Well-Being

The relationship between personality and SWB has received substantial research attention,
with neuroticism and extraversion emerging as important correlates. DeNeve and Cooper
(1998) conducted a large meta-analysis of correlations between SWB and personality
traits. As most of the included studies were conducted prior to the emergence of the Big 5
as a unifying framework, they categorized the studies according to the Big 5. More
recently, Steel et al. (2008) conducted an updated meta-analysis presenting separate results
using common measures of the Big 5. Both meta-analyses found that neuroticism had the
highest correlation with life satisfaction and negative affectivity while extraversion had the
highest correlation with positive affect.
Only a few studies have examined correlations between personality and PWB (Bardi
and Ryff 2007; Butkovic et al. 2012; Grant et al. 2009; Keyes et al. 2002; Schmutte and
Ryff 1997). Such studies suggest that neuroticism, extraversion and conscientiousness are
the major correlates for most PWB dimensions. More generally, dimensions of PWB tend
to be better predicted by personality than are SWB dimensions, and PWB tends to correlate
with more of the Big 5 dimensions.
In order to compare Big 5 correlates of SWB versus PWB, Grant et al. (2009) used a
model constraints approach. After reversing neuroticism and negative affect, they found
support for a model where personality traits varied in their average correlation with well-
being, and well-being dimensions varied in their average correlation with personality.
Average well-being correlations were largest for neuroticism (r = -.44), followed by
extraversion (r = .31), conscientiousness (r = .29), openness (r = .12), and agreeableness
(r = .11). PWB variables tended to correlate more with personality than did SWB vari-
ables. There were also several unique combinations of personality factors and well-being
variables that had larger correlations than would be expected from their component
averages. These were neuroticism with negative affect, extraversion with positive affect
and positive relations, conscientiousness with personal growth and purpose in life,
agreeableness with positive relations, and openness with personal growth. In the opposite
direction, autonomy correlated less with agreeableness than would be expected by com-
ponent averages.

1.4 Personality Facets and Well-Being

While the Big 5 provides a useful organizing framework for personality research, several
researchers have raised concerns that a more detailed model of personality is required to
adequately predict criteria of interest such as well-being (Paunonen and Ashton 2001;
Paunonen and Jackson 2000). To provide a richer model of personality, many personality
tests include both high-level factors, such as the Big 5, and nested lower-level facets. For
example, the NEO-PI-R (Costa and McCrae 1992) includes six facets for each factor of the
Big 5. So for example, the neuroticism factor is composed of the facets of anxiety,
hostility, depression, self-consciousness, impulsiveness, and vulnerability to stress. As a
result, many studies have examined facet-level correlations with a range of outcome

123
 J. Anglim, S. Grant

variables (e.g., Paunonen and Ashton 2001), and a few have examined facet-level corre-
lations with well-being (e.g., Quevedo and Abella 2011; Schimmack et al. 2004).
Initial research on facet-level prediction of SWB suggested that facets enable a dramatic
increase in prediction of SWB (Quevedo and Abella 2011; Schimmack et al. 2004; Steel
et al. 2008), yet critical analysis suggests that incremental prediction may be more modest.
For example, Quevedo and Abella (2011) examined prediction of SWB by NEO Big 5 and
30 facets as well as additional scales of optimism, self-esteem, and perceived social
support. Using stepwise regression predicting life satisfaction they found that adjusted r2
was .16 with the NEO Big 5 as predictors, .22 with NEO 30 facets as predictors, and .29
with NEO 30 facets and additional non-NEO scales including self-esteem and perceived
social support as predictors. They interpreted their results as indicating that facets explain
double the variance of factors. However, an alternative interpretation is that self-esteem
and perceived social support are not personality traits in the traditional sense, and thus the
adjusted r2 comparison of .16 for factors versus .22 (i.e., 37.5 % increase) for facets is a
more reasonable estimate. A second example is provided in the meta-analysis of Steel et al.
(2008), based on pooled correlations, which suggested very large incremental prediction
for extraversion and neuroticism facets over corresponding factors. However, the authors
acknowledged that the obtained estimates of incremental prediction were unreasonably
large, suggesting that the process of pooling correlations across studies may have led to
unreliable estimates.
In relation to PWB, to our knowledge, Siegler and Brummett (2000) provide the only
facet-level analysis to date. They used data from a pre-existing study, and although this
included items for the 30 NEO facets, dimensions of PWB were approximated based on
available items rather than established PWB scales. The study reported facet-level corre-
lations with the constructed indices of PWB. No estimate of incremental prediction of
facets over factors was provided.

1.5 The Current Study

There are several problems with existing approaches to performing facet-level analysis.
First, much of the broader facet-analysis literature has relied on small samples with fewer
than 200 respondents, which has produced uncertain estimates of incremental prediction
and increased the biasing effect that can result from having many more facets than factors
as predictors. Second, methods for assessing the incremental prediction of facets have been
employed without explicit articulation of the population parameter being estimated. Thus,
it has been difficult to evaluate the potential bias and uncertainty in parameter estimates
due to stepwise regression with different p-entry rules, adjusted or non-adjusted r2, and use
of only some or all factors. Third, existing research has involved different types of factor-
facet comparisons. Specifically, studies vary in their use of facet-level test inventories, the
number of Big 5 factors included, and their inclusion of variables that are arguably not
personality traits. Finally, many studies have only reported zero-order correlations between
facets and criteria, instead of controlling for variance explained by factors. This leads to a
dramatic loss of parsimony without evidence of whether a facet-level analysis is superior.
In summary, there is a need to provide a realistic picture of the value of a facet-level
analysis for understanding the relationship between personality and SWB and PWB. Some
existing studies suggest that facets may explain double the variance of factors, yet the
combination of methods used and minimal research suggest that this may be an overes-
timate, at least when limited to facets within a Big 5 inventory.

123
Incremental Prediction from 30 Facets Over the Big 5

The aim of the current study was to examine the relationship between personality and
well-being, focusing particularly on the degree to which 30 personality facets provide
incremental prediction of well-being over and above Big 5 personality factors. To provide
a more comprehensive perspective, both SWB and PWB were examined. To provide more
accurate estimates, we applied new methods for obtaining unbiased estimates and confi-
dence intervals of incremental facet prediction.
Using a moderately large sample, we measured the Big 5 factors and 30 facets of
personality, SWB, and Ryff’s (1989) six dimensions of PWB. To overcome issues with
previous studies, we applied methods to get both point estimates and confidence intervals
for incremental prediction of facets over factors for SWB and PWB. We also assessed
incremental facet prediction using semi-partial correlations controlling for the Big 5. In
general, we predicted a more modest prediction of facets over factors in the range of almost
none to a third more variance explained. We expected facet-level semi-partial correlations
to highlight a small number of meaningful incremental facets, with a factor-level expla-
nation capturing most of the main story.
Specifically, to overcome previous limitations, we define the parameter of interest as the
population incremental variance explained, Dq2 , by facets, q2ðfacetsÞ , over factors, q2ðfactorsÞ .
Thus, Dq2 ¼ q2ðfacetsÞ  q2ðfacetsÞ . Since adjusted r2 is designed to provide an unbiased esti-
mate of q2 , we recommend using R2adjðfacetsÞ  R2adjðfacetsÞ as the estimator for Dq2 . We use a
double-adjusted-r2 bootstrap procedure for providing confidence intervals on the incre-
mental population prediction of facets. Finally, we examine semi-partial correlations
between facets and criteria, where facets are adjusted for factors, in order to assess
incremental contribution of facets in a more parsimonious way than only reporting zero-
order correlations.
In addition to the facet-level analysis, we also examined the relationship between
personality factors and well-being. We decomposed cross-correlations between personality
factors and well-being in order to identify the unique profile of personality correlations for
each type of well-being, thereby replicating and extending previous work by Grant et al.
(2009), who included only four dimensions of PWB. We examined cross-correlations using
all six PWB dimensions. In particular, we were interested in whether the unique combi-
nations of correlations between personality factors and well-being would replicate.

2 Method

2.1 Participants and Procedure

The sampling method was based on convenience sampling and, as such, participants were
mostly undergraduate psychology students drawn from two Australian universities. The
final cleaned sample for this study included 337 participants (24 % male, 76 % female).
Ages ranged from 16 to 55 (M = 21, SD = 8.8). The study was completed online using
Inquisit 3.0 (2012). After reading a plain language statement and providing informed
consent for participation, participants completed demographics, the IPIP-NEO, SWLS,
PANAS, and PWB. The final sample was drawn from an initial sample of 420 participants.
Participants were excluded if any of the following criteria were met: (1) they took less than
500 ms to respond to 10 or more items out of the 409 personality and well-being items
(n = 72), or (2) they failed to answer one or more personality or well-being items
(n = 14).

123
 J. Anglim, S. Grant

2.2 Instruments

2.2.1 International Personality Item Pool (IPIP) Scales Measuring Constructs Similar
to 30 NEO-PI-R Facet Scales

This inventory provides a measure of both Big 5 personality factors (neuroticism, extra-
version, conscientiousness, agreeableness, openness) as well as 30 facets representing six
facets per factor (Goldberg 1999; Goldberg et al. 2006). The 30 facets are closely aligned
with those of the 30 item NEO-PI-R (Costa and McCrae 2008). The IPIP measure has the
advantage of being in the public domain permitting full disclosure of item content and
sharing of raw data. The test is composed of 300 items, 10 items per facet, and 60 items per
factor. Each item is rated on a 5-point scale measuring the degree to which it accurately
describes the participant (1 = very inaccurate, 2 = moderately inaccurate, 3 = neither
inaccurate nor accurate, 4 = moderately accurate, 5 = very accurate). Scales were com-
puted as the mean of items after any required item-reversal. Initial evidence regarding the
reliability and predictive validity of the IPIP scales is favorable (Goldberg 1999). The
scales have an average coefficient alpha of .80 and an average correlation with corre-
sponding NEO-PI-R scales of .73, or .94 when corrected for attenuation due to the unre-
liability of the scales in each pair (Goldberg 1999). The IPIP scales show good predictive
utility for health-related criterion variables. Johnson’s (2000) factor analysis (principal
components) of the IPIP facet-level scales showed that a five-factor solution accounted for
64.9 % of the variance. Facets generally loaded as expected, and the five factors were
clearly defined by the five sets of six facet scales, with the facet scales within a given
domain showing primary loadings on the domain factor in 27 out of 30 cases.

2.2.2 Scales of Psychological Well-Being (Ryff 1989)

This inventory measures six dimensions of PWB: positive relations, autonomy, environ-
mental mastery, personal growth, purpose in life, and self-acceptance. Each item was rated
on a 6-point Likert-style response scale (1 = strongly disagree, 2 = disagree somewhat,
3 = disagree slightly, 4 = agree slightly, 5 = agree somewhat, 6 = strongly agree).
Responses were scored as the mean after any required item-reversal. The 14-item per scale
version was used to ensure reliability for high quality measurement. Specifically, for the
14-item version, Ryff and Essex (1992) report internal consistency alpha coefficients
ranging from .86 to .93. Factor analytic evidence suggests that (a) self-acceptance and
environmental mastery are closely related to traditional SWB measures, (b) personal
growth, positive relations with others and purpose in life share a higher order factor, and
(c) autonomy is more distinct, being more related to variables concerned with power and
control (Ryff 1989).

2.2.3 Satisfaction with Life Scale (Diener et al. 1985)

This well-established 5-item scale was used to measure global life satisfaction. Each item
was rated on a 7-point Likert-style response scale (1 = strongly disagree, 2 = disagree,
3 = slightly disagree, 4 = neither agree nor disagree, 5 = slightly agree, 6 = agree,
7 = strongly agree). The scale scores were computed as the mean of the items. Diener
et al. (1985) reported high internal consistency and high temporal reliability for the scale.
The two-month test–retest reliability in their study was .82 with a Cronbach’s alpha of .87.
Item loadings ranged from .61 to .84, with a single factor accounting for 66 % of the

123
Incremental Prediction from 30 Facets Over the Big 5

variance. In addition, the scale correlated significantly with related measures (e.g., per-
sonality, self-esteem, symptom checklist) and was uncontaminated by social desirability.

2.2.4 Positive and Negative Affectivity Schedule (PANAS; Watson et al. 1988)

The PANAS consists of two subscales that measure positive and negative affect. In the
current study, the instrument was administered using ‘‘past few weeks’’ time instructions.
Participants rated the extent to which they had experienced each of 20 emotions over the
past few weeks on a 5-point scale (1 = very slightly or not at all, 2 = a little,
3 = moderately, 4 = quite a bit, 5 = extremely). Scales were scored as the mean of the
items. Watson et al. (1988) reported that reliabilities (Cronbach’s alpha) were within an
acceptable range for both positive and negative affect (.86 to .90) and were unaffected by
the time instructions used. Both subscales demonstrated satisfactory test–retest reliability
over a 2-month period. The same authors reported a low (negative) correlation between
positive and negative affect, with adjectives loading on the appropriate factor. The sub-
scales showed good external validity, correlating significantly with measures of anxiety,
depression, and distress.

2.3 Data Analysis

Data was analyzed using R 3.0.1 (R Core Team 2013). In the interests of reproducible
research, all code used to perform the analysis and all data and metadata is available from
figshare.com (Analysis for ‘‘Predicting Psychological and Subjective Well-Being’’; doi:10.
6084/m9.figshare.972885).

3 Results

3.1 Descriptive Statistics, Reliabilities, and Factor Analysis

Tables 1 and 2 report descriptive statistics and reliability for all scales used in the study.
Reliability was generally very good with mean Cronbach’s alpha of .81 for personality
factors, .80 for personality facets, and .88 for well-being scales.
Exploratory factor analysis was performed on facet scale scores to examine whether the
30 facets loaded on the proposed five factors. Five factors were extracted using maximum
likelihood estimation with Promax rotation. Overall, the factor solution showed good
correspondence to the theorized structure. There was a clear drop in the scree plot after five
factors and the parallel analysis also suggested five factors. Five factors explained 58.9 %
of the variance. Of the 30 facets, 28 facets loaded above .35, and 25 loaded maximally on
their theorized factor. Out of the 120 cross-loadings of facets on non-theorized factors,
only 12 (10 %) loaded above .35. Prominent cross-loadings included self-consciousness
(-.51), trust (.56), and altruism (.50) on extraversion; activity level (.59) on conscien-
tiousness; dutifulness (.51) on agreeableness; and emotionality (.51) on neuroticism.

3.2 Correlations Between Big 5 Personality and Well-Being

The full correlation matrix between factors, facets, and well-being measures is available
from the online repository mentioned in the Method. Table 3 shows the correlations

123
 J. Anglim, S. Grant

Table 1 Descriptive statistics

Variable a M SD
and Cronbach’s alpha reliability
for personality factors and well-
being scales Neuroticism .85 2.91 0.59
Extraversion .84 3.37 0.54
Openness .72 3.64 0.41
Agreeableness .79 3.64 0.45
Conscientiousness .84 3.49 0.51
Satisfaction with life .88 4.47 1.44
Positive affect .89 3.47 0.79
Negative affect .87 2.21 0.80
Positive relations .89 4.42 0.90
Autonomy .85 4.08 0.83
Environmental mastery .88 4.07 0.87
Personal growth .86 4.85 0.69
Purpose in life .89 4.36 0.91
Self-acceptance .94 4.06 1.07

between personality factors and well-being scales. To better understand the cross-corre-
lations between personality factors and well-being, a decomposition was performed. First,
neuroticism and negative affect were reversed, so that all variables were positively framed.
Second, cross-correlations were obtained between personality factors and well-being
variables denoted by rij where i ¼ 1; . . .; I, and j ¼ 1; . . .; J indexing the I ¼ 5 personality
factors and J ¼ 9 well-being variables respectively. Then, the overall average cross-cor-
relation was obtained as
1 XX
r:: ¼ rij ;
IJ i j

as well as the average deviation for cross-correlations for each well-being variable
1X
r:j ¼ rij
I i

and personality factor

1X
ri: ¼ rij :
J j

Thus, observed correlations can be decomposed into the overall average correlation,
average deviation of the personality correlations, average deviation of the well-being
correlations, and a residual.
rij ¼ r:: þ ð
r:j  r:: Þ þ ð
ri:  r:: Þ þ uij :

So for example, the expected correlation between openness and personal growth was the
grand mean (.39) plus the deviation from the grand mean of the average openness cor-
relation (-.16) plus the deviation from the grand mean of the average personal growth
correlation (.03) which equals .26, but the obtained correlation was .55; the residual was

123
Incremental Prediction from 30 Facets Over the Big 5

Table 2 Descriptive statistics

Variable a M SD
and Cronbach’s alpha reliability
for personality facets
N1: Anxiety .83 3.10 0.73
N2: Anger .88 2.85 0.80
N3: Depression .90 2.61 0.88
N4: Self-consciousness .83 2.94 0.77
N5: Immoderation .77 3.24 0.70
N6: Vulnerability .85 2.73 0.75
E1: Friendliness .88 3.60 0.78
E2: Gregariousness .87 3.32 0.84
E3: Assertiveness .84 3.24 0.74
E4: Activity level .72 3.01 0.57
E5: Excitement seeking .82 3.30 0.72
E6: Cheerfulness .82 3.75 0.67
O1: Imagination .82 3.73 0.71
O2: Artistic interests .75 3.96 0.62
O3: Emotionality .77 3.76 0.62
O4: Adventurousness .76 3.47 0.59
O5: Intellect .79 3.70 0.64
O6: Liberalism .65 3.21 0.60
A1: Trust .85 3.43 0.70
A2: Morality .77 3.85 0.62
A3: Altruism .79 4.03 0.57
A4: Cooperation .75 3.55 0.67
A5: Modesty .80 3.28 0.70
A6: Sympathy .73 3.68 0.59
C1: Self-efficacy .80 3.69 0.57
C2: Orderliness .84 3.33 0.78
C3: Dutifulness .69 3.94 0.50
C4: Achievement striving .84 3.67 0.70
C5: Self-discipline .89 3.01 0.83
C6: Cautiousness .80 3.26 0.67

therefore .55 - .26 = .29. Table 4 reports this analysis. Large positive residual correla-
tions indicate that the two variables correlate more with each other than would be expected
based on how much the variables correlate generally with other variables. Thus, such
correlations help to highlight the unique personality profile of each well-being variable.
The average correlation between personality and well-being was moderately large (.39).
On the well-being side, there was not a lot of variation in average correlations with
personality, although autonomy was lower than the others. The average absolute cross-
correlation was larger for PWB scales (.41) than for SWB scales (.36). In terms of per-
sonality, the common ordering emerged of neuroticism being most important by some
margin followed by extraversion and conscientiousness, and then with much weaker
average cross-correlations for openness and agreeableness. In addition to these general
patterns, there were several notable residual cross-correlations shown in Table 4. For

123
123
Table 3 Correlations between Big 5 personality and well-being scales
Variable 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Neuroticism –
2. Extraversion 2.47 –
3. Openness 2.07 .26 –
4. Agreeableness 2.21 .02 .23 –
5. Conscientiousness 2.54 .14 .04 .36 –
6. Satisfaction with life 2.57 .51 .13 .11 .35 –
7. Positive affect 2.52 .56 .22 .20 .45 .51 –
8. Negative affect .76 2.40 2.06 2.22 2.42 2.51 2.35 –
9. Positive relations 2.48 .58 .25 .38 .28 .55 .43 2.45 –
10. Autonomy 2.53 .26 .27 .04 .34 .33 .27 2.49 .32 –
11. Environmental mastery 2.75 .56 .13 .22 .60 .69 .58 2.68 .64 .48 –
12. Personal growth 2.42 .43 .55 .31 .40 .44 .49 2.39 .56 .51 .55 –
13. Purpose in life 2.58 .50 .25 .28 .65 .63 .62 2.49 .63 .43 .78 .69 –
14. Self-acceptance 2.74 .58 .20 .18 .46 .78 .56 2.64 .68 .54 .80 .60 .76

jrj  :11 indicates p \ .05; jrj  :15 indicates p \ .01 and are bolded
J. Anglim, S. Grant
Incremental Prediction from 30 Facets Over the Big 5

Table 4 Decomposition of cross-correlations between well-being and Big 5 personality

Personality Mean well-being
deviation
N- E O A C

Residual Correlations

Satisfaction with life .04 .08 -.04 -.04 -.03 -.06

Positive affect -.07 .08 -.01 -.01 .02 .00
Reversed negative affect .19 -.06 -.15 .02 .00 -.02
Positive relations -.12 .10 .02 .17 -.16 .00
Autonomy .04 -.12 .14 -.08 .01 -.10
Environmental mastery .09 .02 -.16 -.06 .10 .06
Personal growth -.21 -.09 .29 .07 -.07 .03
Purpose in life -.07 -.05 -.04 .01 .15 .06
Self-acceptance .10 .06 -.06 -.08 -.02 .04

Mean personality deviation .20 .09 -.16 -.18 .05 .39a

Mean deviations are the average cross-correlation for cross-correlations containing the focal variable minus
the overall mean of cross-correlations. Residual correlations equal the actual cross-correlation minus the
correlation predicted by summing the overall mean cross-correlations and the two mean deviations for the
constituent variables
Absolute residual cross-correlations greater than or equal to 0.15 are bolded
N- reversed neuroticism; E Extraversion, O Openness, A Agreeableness, C Conscientiousness
a
Overall mean cross-correlations between personality and well-being variables

example, the following correlated substantially more than was implied by average cross-
correlations for constituent variables: personal growth with openness, positive relations
with agreeableness, neuroticism with negative affect, and purpose in life with conscien-
tiousness. Interestingly, there were several negative residual cross-correlations: openness
with negative affect, personal growth with neuroticism, environmental mastery with
openness, and positive relations with conscientiousness. Thus, for example, the profile of
correlations for positive relations indicates stronger relations with agreeableness and
weaker relations with conscientiousness than is typically the case for well-being variables.

3.3 Facet-Level Correlations with Well-Being

Table 5 reports zero-order correlations between personality facets and well-being scales
and semi-partial correlations between personality facets and well-being scales (see
parentheses), where facets have been adjusted for their shared variance with personality
factors. The zero-order correlations present a complex pattern with many large correlations
often consistent with patterns at the factor-level. The semi-partial correlations focus purely
on the incremental prediction of facets over factors. Notable semi-partial correlations are
depression with satisfaction with life (r = -.28), positive relations (r = -.20), purpose in
life (r = -.24), and self-acceptance (r = -.37); self-consciousness with autonomy
(r = .20); assertiveness with autonomy (r = .18); excitement seeking with positive rela-
tions (r = -.19); cheerfulness with satisfaction with life (r = .26), cooperation with
autonomy (r = -.19); and achievement striving with purpose in life (r = .21). The table
also reports the proportion of variance in the facet that is not explained by factors. The
mean proportion of unique variance was 35.6 %.

123
 Table 5 Zero-order correlations between facets and well-being (semi-partial correlations with variance explained by factors removed from facets shown in parentheses)
SWL PA NA PR AU EM PG PL SA Uniquea Meanb

123
N1: Anxiety -.47 (.02) -.39 (.00) .64 (-.01) -.35 (.07) -.41 (.06) -.58 (.02) -.29 (-.01) -.38 (.05) -.57 (.09) .21 .03
N2: Anger -.31 (.05) -.29 (-.03) .54 (-.02) -.27 (.08) -.25 (.17) -.43 (.05) -.22 (.04) -.30 (.04) -.39 (.13) .33 .06
N3: Depression -.67 (-.28) -.56 (-.09) .70 (.12) -.59 (-.20) -.45 (-.03) -.76 (-.14) -.48 (-.13) -.70 (-.24) -.86 (-.37) .27 -.18
N4: Self-consciousness -.46 (.14) -.45 (.11) .56 (-.02) -.47 (-.02) -.53 (-.20) -.59 (.08) -.40 (-.01) -.48 (.05) -.63 (.04) .24 .02
N5: Immoderation -.16 (.09) -.16 (.05) .32 (-.09) -.07 (.07) -.29 (.01) -.32 (.07) -.10 (.12) -.24 (.12) -.24 (.10) .49 .08
N6: Vulnerability -.49 (.00) -.46 (-.03) .65 (.03) -.37 (.02) -.49 (-.04) -.68 (-.08) -.37 (-.02) -.50 (-.01) -.61 (.04) .27 -.02
E1: Friendliness .44 (-.03) .49 (-.03) -.41 (.00) .67 (.17) .22 (-.02) .55 (.04) .37 (-.07) .49 (.04) .56 (.03) .23 .01
E2: Gregariousness .34 (-.05) .34 (-.11) -.27 (-.01) .46 (-.01) .03 (-.15) .36 (-.04) .19 (-.12) .28 (-.06) .37 (-.07) .27 -.07
E3: Assertiveness .39 (-.08) .46 (-.03) -.31 (.03) .39 (.01) .39 (.18) .48 (-.03) .41 (.10) .45 (.00) .50 (.01) .29 .01
E4: Activity level .33 (-.02) .49 (.08) -.30 (.01) .35 (.01) .26 (.05) .51 (.04) .38 (.07) .53 (.06) .44 (.03) .55 .03
E5: Excitement seeking .21 (-.03) .23 (.01) -.08 (.05) .15 (-.19) .05 (-.07) .10 (-.12) .15 (-.03) .04 (-.12) .16 (-.13) .38 -.08
E6: Cheerfulness .57 (.26) .52 (.09) -.41 (-.08) .57 (.06) .25 (.02) .54 (.12) .46 (.07) .49 (.11) .59 (.16) .30 .11
O1: Imagination .06 (.07) .07 (.01) .06 (.00) .07 (-.04) .08 (-.10) -.04 (.00) .25 (-.09) .04 (-.02) .02 (-.04) .41 -.02
O2: Artistic interests .06 (-.01) .19 (.06) -.05 (-.02) .16 (-.04) .08 (-.10) .05 (-.06) .32 (-.08) .16 (-.06) .13 (.01) .54 -.03
O3: Emotionality .01 (.04) .13 (.02) .16 (.03) .25 (.14) .10 (.14) -.01 (.02) .41 (.12) .22 (.12) .08 (.13) .40 .08
O4: Adventurousness .27 (-.01) .29 (-.02) -.27 (.01) .30 (-.07) .27 (-.04) .31 (.00) .51 (.11) .32 (.02) .34 (-.04) .53 .00
O5: Intellect .18 (-.06) .25 (-.04) -.19 (.04) .21 (.00) .44 (.14) .31 (.02) .50 (.05) .33 (-.04) .26 (-.08) .45 .00
O6: Liberalism -.10 (-.03) -.08 (-.03) .06 (-.04) -.02 (.02) .07 (-.02) -.10 (.03) .13 (-.07) -.06 (-.02) -.04 (.02) .60 -.01
A1: Trust .35 (.05) .36 (.01) -.38 (.00) .59 (.14) .13 (-.06) .44 (.05) .37 (.01) .39 (.03) .45 (.07) .43 .03
A2: Morality .05 (-.02) .09 (-.07) -.20 (-.06) .21 (-.03) .08 (.11) .22 (.07) .18 (-.02) .28 (.06) .15 (.07) .31 .02
A3: Altruism .26 (-.01) .41 (.07) -.24 (.07) .52 (.03) .15 (.04) .37 (-.02) .47 (.07) .45 (.02) .35 (.02) .24 .02
A4: Cooperation .07 (.03) .08 (.00) -.22 (-.03) .17 (-.07) -.06 (-.19) .13 (-.04) .14 (-.06) .17 (-.01) .11 (.00) .33 -.03
A5: Modesty -.32 (-.11) -.22 (.01) .14 (-.03) -.17 (-.11) -.15 (.09) -.26 (-.01) -.15 (-.04) -.24 (-.09) -.38 (-.16) .43 -.04
A6: Sympathy .09 (.07) .14 (.00) -.01 (.09) .33 (.04) .03 (.02) .05 (-.07) .36 (.06) .20 (.01) .11 (.04) .32 .01
C1: Self-efficacy .49 (.02) .51 (-.05) -.52 (.01) .44 (.07) .53 (.14) .70 (.05) .56 (.11) .70 (.07) .64 (.06) .26 .05
C2: Orderliness .07 (-.09) .22 (.01) -.17 (-.02) .02 (-.05) .06 (-.10) .25 (-.09) .11 (-.05) .30 (-.12) .13 (-.06) .36 -.06
C3: Dutifulness .14 (-.09) .25 (-.07) -.26 (.01) .22 (-.05) .23 (.11) .37 (-.01) .28 (-.02) .41 (-.06) .26 (-.01) .35 -.02
J. Anglim, S. Grant
 Table 5 continued
SWL PA NA PR AU EM PG PL SA Uniquea Meanb

C4: Achievement striving .42 (.13) .54 (.16) -.31 (.07) .30 (.01) .33 (.04) .57 (.04) .47 (.12) .72 (.21) .48 (.08) .29 .08
C5: Self-discipline .41 (.05) .49 (.08) -.44 (.00) .27 (-.04) .30 (-.12) .62 (.05) .34 (-.06) .60 (-.02) .47 (-.03) .25 -.01
C6: Cautiousness .06 (-.01) .03 (-.15) -.22 (-.05) .05 (.08) .17 (.01) .23 (-.01) .11 (-.06) .26 (-.04) .12 (-.01) .35 -.02

Significant semi-partial correlations (p \ .001) are bolded

SWL satisfaction with life, PA positive affect, NA negative affect, PR positive relations, AU autonomy, EM emotional mastery, PG personal growth, PL purpose in life, SA self-acceptance
a
Proportion of personality variance not predicted by Big 5 factors of personality
b
Average semi-partial correlation for facet across well-being variables, where negative affect has been reversed
Incremental Prediction from 30 Facets Over the Big 5

123
 J. Anglim, S. Grant

Table 6 Standardized regression coefficients predicting well-being scales from Big 5 personality
DV Standardized beta

N E O A C

Satisfaction with life -.37 .31 .02 -.01 .11

Positive affect -.14 .43 .08 .03 .31
Negative affect .71 -.07 .03 -.07 .00
Positive relations -.18 .49 .03 .33 -.01
Autonomy -.51 -.07 .29 -.18 .13
Environmental mastery -.42 .31 .00 .01 .32
Personal growth -.13 .21 .45 .08 .25
Purpose in life -.13 .33 .13 .03 .52
Self-acceptance -.52 .29 .09 .00 .14

Significant coefficients (p \ .01) are bolded

Table 7 Variance explained in well-being scales by Big 5 personality and 30 facets of personality
DV 2ðfactorsÞ
Radj
2ðfacetsÞ
Radj Dq^2 (95 % CI)

Satisfaction with life .40 .52 .12 [.06, .18]

Positive affect .47 .51 .04 [.00, .09]
Negative affect .58 .59 .01 [.00, .05]
Positive relations .50 .57 .07 [.02, .13]
Autonomy .36 .50 .14 [.07, .21]
Environmental mastery .68 .73 .04 [.01, .08]
Personal growth .51 .59 .07 [.02, .13]
Purpose in life .62 .72 .11 [.07, .15]
Self-acceptance .63 .78 .15 [.10, .20]
2ðfacetsÞ 2ðfactorsÞ
Dq^2 ¼ Radj  Radj

3.4 Prediction of Well-Being from Personality Factors and Facets

A set of linear multiple regressions (direct entry) was conducted predicting each well-being
variable using the Big 5 as predictors. Table 6 reports the obtained standardized regression
coefficients. Table 7 reports the corresponding estimate of population variance explained
2ðfactorsÞ
(i.e., Radj . In general, the Big 5 explained substantial variance. In all cases, the Olkin-
Pratt formula for adjusted r2 was used as it aligns with the assumption that the predictor
2ðfactorsÞ
variables are a random sample from a population. Average Radj was .55 for PWB
variables and .48 for SWB variables.
Table 7 reports the adjusted r2 predicting each well-being scale, first with the five
personality factors as predictors, and then with the 30 personality facets as predictors.
Estimates of incremental population variance explained, Dq2 , were obtained by subtracting
R2adj for factors from R2adj for facets. Confidence intervals were obtained using a double-
adjusted-r squared bootstrap method. This method involves first sampling with replace-
ment from the data to generate K bootstrap samples of equal size as the raw data. For each

123
Incremental Prediction from 30 Facets Over the Big 5

bootstrap sample, R2ðfactorsÞ and R2ðfacetsÞ is obtained. Then for both R2 values, the formula for
adjusted r2 is applied twice
  
R~2adj ¼ f f R2

where f ð:Þ is the formula for adjusted r2. The adjustment formula is applied twice, first to
adjust for the bias associated with the bootstrap treating the sample as the population, and
second to adjust for the standard bias in estimating q2 from sample data. Finally the
estimate is obtained for the particular bootstrap sample as Dq^2 ¼ R~2adjðfacetsÞ  R~2adjðfactorsÞ .To
obtain 95 % confidence intervals, the .025 and .975 sample quantiles are obtained from the
K bootstrap estimates.
The mean ratio of facets to factors adjusted r2 was 1.17. The mean Dq^2 was .08. In
general, Dq^2 was larger for PWB (mean Dq^2 = .10) than for SWB (mean Dq^2 = .06). In
particular, negative affect, which correlated very highly with neuroticism, showed minimal
incremental facet prediction. Positive affect and environmental mastery showed small
amounts of incremental facet prediction. Autonomy and self-acceptance showed the largest
amounts of incremental facet prediction.

4 Discussion

This study aimed to examine the relationship between personality and well-being. In
particular, it examined the incremental prediction of personality facets over Big 5 factors.
In general, personality and well-being showed substantial correlation. Facets accounted for
additional population variance in well-being but the increase was often modest, ranging
from almost no additional variance explained to around a third more variance explained.
The subsequent discussion elaborates, first, on factor-level relationships, then on facet-
level relationships, and finally on broader theoretical and methodological issues.

4.1 Well-Being and the Big 5

There were several general patterns in the cross-correlations between Big 5 personality and
well-being. First, neuroticism was clearly the largest and most consistent correlate of well-
being; then came extraversion, closely followed by conscientiousness. These findings are
generally consistent with DeNeve and Cooper (1998) and Steel et al. (2008), whose meta-
analytic studies focused on SWB, and are consistent with Grant et al. (2009) and Keyes
et al. (2002). While agreeableness and openness still had meaningful correlations, these
were less consistent and generally smaller. Second, PWB dimensions showed a slightly
stronger relationship with the Big 5 than did SWB dimensions. Butkovic et al. (2012)
likewise reported that personality explained more variance in PWB than SWB. Third,
consistent with Schmutte and Ryff (1997), PWB showed a more diverse relationship with
personality than did SWB. In broad terms, SWB dimensions were often well predicted by
neuroticism and extraversion, whereas agreeableness, openness, and conscientiousness
were important correlates of several PWB dimensions (c.f. Grant et al. 2009). In addition,
residual cross-correlations and standardized betas highlighted several relationships
between PWB and the Big 5 that shed light on the nature of the PWB construct contrib-
uting to broader discussion regarding the meaning of PWB (e.g., Ryff and Singer 1998,
2006). These points are elaborated in more detail below. Taken together, the results

123
 J. Anglim, S. Grant

reinforce the notion that the key dispositional influences on well-being vary across well-
being dimensions (Grant et al. 2009).
There were no significant residual cross-correlations for SWL, indicating that there were
no personality variables that correlated more strongly with SWB than expected based on
their correlations with other well-being variables. Neuroticism had the strongest stan-
dardized beta for SWL, which was also predicted by extraversion and to a lesser extent
conscientiousness. In contrast, extraversion was the strongest predictor of PA, which was
also predicted by conscientiousness and neuroticism. There were no significant residual
cross-correlations for PA, but neuroticism and openness showed significant residual cross-
correlations for NA. Neuroticism was the only significant predictor of NA. The finding that
SWB dimensions were well predicted by neuroticism and extraversion is consistent with
meta-analytic studies (DeNeve and Cooper 1998; Steel et al. 2008).
Personal growth had a positive residual cross-correlation with openness and a negative
residual cross-correlation with neuroticism. Bardi and Ryff (2007) similarly reported that
individuals who were higher on openness and lower on neuroticism reported higher per-
sonal growth. Standardized betas showed that personal growth was predicted by all five
traits, with openness emerging as the strongest predictor. This strong relationship between
personal growth and openness is consistent with Schmutte and Ryff (1997). Personal
growth items include the perception that the individual is growing, a belief that change is
possible, and valuing of change (Ryff 1989). Thus, beyond the pure well-being elements,
the measure of personal growth also captures a disposition to the concept that growth and
change is positive, which helps to explain the relationship with openness. Arguably, these
more attitudinal elements go beyond pure well-being and actually suggest a humanistic
value system regarding what is the good life.
Autonomy was one of the least well-predicted well-being dimensions. There were no
significant residual cross-correlations for this dimension, although standardized betas
indicated that the dimension was predicted by greater openness and conscientiousness and
less neuroticism and agreeableness, with neuroticism being the strongest predictor. Pre-
vious studies have also primarily identified an association between autonomy and agree-
ableness or neuroticism (Grant et al. 2009; Schmutte and Ryff 1997), perhaps reflecting the
focus of autonomy items on a lack of care for what others think or low self-consciousness.
However, there is also arguably an implicit assumption that autonomy involves some
degree of independent thinking. Items capture self-confidence and as well as a spectrum of
not being excessively influenced by others to more extreme independence of thought.
Emotional stability (the inverse of neuroticism) and antagonism (the inverse of agree-
ableness) capturinquie elements of self-confidence and independent thinking respectively.
A readiness to not conform can go against being agreeable. The Ryff scale measures a
relatively social conception of autonomy. While much of the autonomy construct captures
positive aspects, there is an aspect that might actually result in less well-being. For
instance, not listening to the views of others, never sacrificing one’s needs for the needs of
others, or an inability to accept the rituals and values of a society could have a range of
negative consequences. Similarly, some individuals may place less value on independence
of thought thereby further reducing the relationship between autonomy and well-being.
Positive relations had a positive residual cross-correlation with agreeableness and a
negative residual cross-correlation with conscientiousness and it was predicted particularly
by extraversion, agreeableness and neuroticism. Items for positive relations capture not
only whether a person has good friends, but also whether the person values interactions
with others and sees him or herself as capable of being a good friend. In this sense,
extraversion relates to both social engagement and a desire to be social, and agreeableness

123
Incremental Prediction from 30 Facets Over the Big 5

captures many aspects related to being friendly and accommodating. Consistent with this,
previous studies have primarily linked positive relations to agreeableness and extraversion
(Grant et al. 2009; Schmutte and Ryff 1997). However, the Ryff measure of positive
relations goes beyond measuring presence of or satisfaction with interpersonal relation-
ships, also measuring evaluative judgments about the importance of friendship and skills in
friendship formation, suggesting that other personality dimensions are also important.
Indeed, Siegler and Brummett (2000) linked positive relations with select facets of all Big
5 traits.
Purpose in life had a strong positive residual cross-correlation with conscientiousness.
Conscientiousness also had the strongest standardized beta, followed by extraversion while
neuroticism and openness showed a weaker relationship with purpose in life. The strong
association between purpose in life and conscientiousness is consistent with previous work
(Grant et al. 2009; Schmutte and Ryff 1997), and others have also documented the asso-
ciations between this dimension and extraversion and neuroticism (Schmutte and Ryff
1997; Siegler and Brummett 2000). Purpose in life items focus on having longer term
projects, getting pleasure from moving towards goals, and aspects of life satisfaction.
Self-acceptance and environmental mastery tended to have similar patterns to satis-
faction with life, with significant betas for neuroticism, extraversion and conscientiousness
(self-acceptance was also predicted by openness, though to a lesser extent). Environmental
mastery had a significant residual cross-correlation with openness; there were no signifi-
cant residual cross-correlations for self-acceptance. Both of these PWB dimensions have
been flagged (Bouchard and Loehlin 2001) as more reflective of SWB than PWB. Self-
acceptance items largely focus on self-esteem, positive comparison of self versus others,
and elements of life satisfaction. Environmental mastery focuses on a sense control, with
elements of life satisfaction.
It is noteworthy that the Big Five predicted self-acceptance and environmental mastery
more strongly than they predicted satisfaction with life. Once again, this is consistent with
previous work supporting a stronger relationship between personality and well-being for
PWB than SWB (Butkovic et al. 2012; Grant et al. 2009) and reinforces the distinctiveness
of these dimensions.
At present, the Ryff scales seem to incorporate more than just whether the well-being
aspects are present; they also embody a range of assumptions about what constitutes the
good life. Of course, psychological theory underpins the importance of such dimensions,
but each dimension captures a unique flavor of the concept of PWB and also seems to
measure the degree to which that dimension is characteristically valued by the individual.
Thus, open people may search for personal growth. Disagreeable people may be more
willing to assert their opinion in defiance of what a group thinks. And conscientious people
may value purpose in life and seek to achieve projects and plans. While any measure of
well-being will have a particular orientation, there is a risk of imposing a humanistic value
system on to people by labeling such dimensions as well-being rather than using the more
theoretically neutral SWB dimensions.

4.2 Incremental Prediction of Well-Being by Facets

Overall, personality facets provided a meaningful increase in the variance explained in

well-being over and above the personality factors. However, there was no doubling of
explained variance as eluded to in the extant literature (see Quevedo and Abella 2011).
Instead, increases ranged from almost nothing to around a third more. Well-being
dimensions varied substantially in the size of this increase. Positive and negative affect

123
 J. Anglim, S. Grant

both showed minimal increases, which is inconsistent with the very large incremental
prediction achieved for facets over factors for some traits in meta-analytic research (Steel
et al. 2008). In contrast, self-acceptance, autonomy, satisfaction with life, and purpose in
life showed fairly large increases. However, this is the first study to estimate the incre-
mental prediction of facets over factors for PWB and the results await replication. Fur-
thermore, the cause of the variation in incremental variance is not entirely clear and
warrants exploration in future research.
Examination of the semi-partial correlations between facets and well-being helped to
explain the incremental prediction by facets. For example, autonomy was associated with
more anger and assertiveness, and less with self-consciousness, cooperation, and gregar-
iousness, reinforcing the above notion that this dimension reflects the degree of importance
placed on what others think and independent thought. Purpose in life had a strong link with
achievement striving, reinforcing the goal-directed emphasis of this dimension. More
generally, depression and cheerfulness emerged as incremental correlates for many well-
being variables. In many cases, these correlates seemed to be related to overlap in the
conceptual nature of the constructs (for further discussion of construct overlap in this
context, see Schmutte and Ryff 1997).
Overall, the bootstrapping and the semi-partial correlations helped to explain the
incremental contribution of facets. First, bootstrapping highlighted the uncertainty around
estimates of incremental variance explained. While the size of the confidence intervals
varied, the sample size of approximately 300 was sufficient for 95 % confidence intervals
to provide a good understanding of the ‘ball park’ of the effect size. Also, the semi-partial
correlations helped to yield a more parsimonious view of the incremental role of facets.
Compared to zero-order correlations, semi-partial correlations flagged only a select few
facets, taking factor correlations as a starting point and presenting a more parsimonious
view. Compared to stepwise regression, the results were less binary in terms of inclusion of
predictors.

4.3 Incremental Facet Prediction

Conclusions about incremental facet prediction in the present study are based on the
inclusion of nested facets. As Quevedo and Abella (2011) found, inclusions of non-nested
facets can substantially increase the incremental prediction of facets. There are several
reasons for this. First, by construction, factors capture some of the variance of nested
facets. So for instance, when comparing facets to the Big 5 from a given test, incremental
prediction should be greater when facets come from a different test. However, by taking
facets from a different test, some of the incremental variance would be obtained by the
slightly different measurement of the Big 5. Second, the selection of facets in a personality
test may be partially constrained by the need to fit within a Big 5 theoretical framework.
Thus, personality traits not captured by the Big 5 might be omitted. However, alternatively,
there is the potential to include variables that are not typically considered personality traits,
or that get even closer to well-being related constructs.
This raises questions about what is a natural or useful way of framing incremental
prediction of well-being from personality facets. It also relates to issues of how personality
tests should be constructed in order to both reliably measure the Big 5 but also capture
diverse facets that assist with incremental prediction. At the very least, it is necessary to be
clear when describing estimates of incremental prediction as to what class of facets is being
included.

123
Incremental Prediction from 30 Facets Over the Big 5

Overall, the results of this study support the value of a facet-level analysis, but suggest
that the contribution is more modest than some previous studies have suggested. The
increases in estimated population prediction seen in this study are of a magnitude that
justifies the increased complexity. Furthermore, in contrast to the complexity of zero-order
correlation matrices, the semi-partial correlation analysis helps to provide a parsimonious
picture of the relevant facets that support the incremental prediction.

4.4 Construct Overlap and Causal Pathways

Beyond identifying the correlations between personality and well-being, there is the
broader issue of the degree to which such relationships are based on construct overlap or
some form of causal relationship. Examination of item content strongly supports the idea
that construct overlap explains many of the observed correlations. Neuroticism measures
the tendency to experience a range of negative emotions, and clearly negative affect is
almost synonymous with this tendency. In the case of extraversion, there is a mixture of
items, some of which pertain directly to the experience of positive emotions whereas others
pertain more to experiences that often elicit positive emotions. However, personality traits
can be seen as more stable than well-being and thus as the cause of well-being. Arguments
can also be made for how personality traits influence the motivation, environment, and
interpretive lens of the individual, which in turn influences well-being. A recent study by
Soto (2014) of the longitudinal relationship between the Big 5 and the SWB dimensions
supported the notion that personality traits and well-being dimensions influence one
another reciprocally over time.
In some respects a facet-level analysis provides greater scope for both forms of pre-
diction, but perhaps greater construct overlap is particularly likely. The Big 5 is necessarily
broad, yet the chance that a well-being scale is going to overlap substantively with a
specific facet scale increases. For example, depression seems to be the aspect of neuroti-
cism that most directly relates to a wide range of well-being measures. Likewise, specific
facets like achievement striving overlap substantively with valuing an orientation to life
that emphasizes personal growth (Ryff and Singer 2006).
The research also raises issues regarding the position of PWB in the causal and defi-
nitional system that contains personality and SWB. For example, Diener et al. (2003)
proposed that there are a multiple pathways to well-being that may differ between people
and across cultures. Generally, environmental mastery, self-acceptance and, to some
extent, purpose in life substantially overlap with satisfaction with life. Satisfaction with life
seems to be the more ‘pure’ measure of well-being in that the individual is free to evaluate
their life on their own terms. Autonomy, positive relations, and personal growth seem to
capture important pathways to SWB. Even if they are viewed as an essential part of well-
being, care is needed when designing measures to ensure attitude to the dimension is not
confounded with status on the dimension.

4.5 Conclusion, Limitations, and Future Research

This study has provided a more complete picture of the relationship between personality,
SWB, and PWB. The results provide a balance between calls that only the Big 5 is
necessary and claims that facets substantially improve prediction. In addition, our meth-
odological approach provided a parsimonious explanation to the complex patterns of cross-
correlations. By making available all data and data analysis code, others are encouraged to
further explore the data to generate additional insights.

123
 J. Anglim, S. Grant

In terms of limitations, the research was conducted on a young adult sample, pre-
dominantly consisting of university students. Such a sample may have particular priorities
and values in life, which may have influenced the pattern of correlations observed. Clearly
more research is required to explore incremental facet prediction with different personality
tests, and different kinds of facets. Furthermore, while the Ryff scales have proven very
useful in advancing understanding of PWB, there may be a need to further refine measures
of PWB to minimize inappropriate measurement of values and unnecessary confounding
with life satisfaction and related measures.

Acknowledgments We thank Sue Carmen for her assistance with data collection.

References

Bardi, A., & Ryff, C. D. (2007). Interactive effects of traits on adjustment to a life transition. Journal of
Personality, 75(5), 955–984. doi:10.1111/j.1467-6494.2007.00462.x.
Bouchard, T. J, Jr, & Loehlin, J. C. (2001). Genes, evolution, and personality. Behavior Genetics, 31(3),
243–273. doi:10.1023/A:1012294324713.
Brickman, P., & Campbell, D. T. (1971). Hedonic relativism and planning the good society. In M. H. Apley
(Ed.), Adaptation-level theory (pp. 287–305). New York: Academic Press.
Butkovic, A., Brkovic, I., & Bratko, D. (2012). Predicting well-being from personality in adolescents and
older adults. Journal of Happiness Studies, 13(3), 455–467. doi:10.1007/s10902-011-9273-7.
Compton, W. C. (1998). Measures of mental health and a five factor theory of personality. Psychological
Reports, 83(1), 371–381. doi:10.2466/pr0.1998.83.1.371.
Costa, P. T., & McCrae, R. R. (1992). Revised NEO personality inventory and NEO five factor inventory:
professional manual (Vol. 396). Odessa: Psychological Assessment Resources.
Costa, P. T., & McCrae, R. R. (2008). The revised NEO personality inventory (NEO-PI-R). The SAGE
handbook of personality theory and assessment, 2, 179–198.
Deci, E. L., & Ryan, R. M. (2008). Hedonia, eudaimonia, and well-being: An introduction. Journal of
Happiness Studies, 9(1), 1–11. doi:10.1007/s10902-006-9018-1.
DeNeve, K. M., & Cooper, H. (1998). The happy personality: a meta-analysis of 137 personality traits and
subjective well-being. Psychological Bulletin, 124(2), 197–229. doi:10.1037/0033-2909.124.2.197.
Diener, E. (1984). Subjective well-being. Psychological Bulletin, 95(3), 542–575. doi:10.1037/0033-2909.
95.3.542.
Diener, E. (Ed.). (2009). Culture and well-being: The collected works of Ed Diener (Vol. 38). New York:
Springer.
Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of
Personality Assessment, 49(1), 71–75. doi:10.1207/s15327752jpa4901_13.
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of
progress. Psychological Bulletin, 125(2), 276–302. doi:10.1037/0033-2909.125.2.276.
Diener, E., Oishi, S., & Lucas, R. E. (2003). Personality, culture, and subjective well-being: Emotional and
cognitive evaluations of life. Annual Review of Psychology, 54(1), 403–425. doi:10.1146/annurev.
psych.54.101601.145056.
Digman, J. M. (1997). Higher-order factors of the Big Five. Journal of Personality and Social Psychology,
73(6), 1246–1256. doi:10.1037/0022-3514.73.6.1246.
Fujita, F., & Diener, E. (2005). Life satisfaction set point: Stability and change. Journal of Personality and
Social Psychology, 88(1), 158–164. doi:10.1037/0022-3514.88.1.158.
Goldberg, L. R. (1992). The development of markers for the Big-Five factor structure. Psychological
Assessment, 4(1), 26–34. doi:10.1037/1040-3590.4.1.26.
Goldberg, L. R. (1993). The structure of phenotypic personality traits. American Psychologist, 48(1), 26–34.
doi:10.1037/0003-066X.48.1.26.
Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the lower-level
facets of several five-factor models. Personality Psychology in Europe, 7, 7–28.
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(1), 84–96. doi:10.1016/j.jrp.2005.08.007.

123
Incremental Prediction from 30 Facets Over the Big 5

Grant, S., Langan-Fox, J., & Anglim, J. (2009). The Big Five traits as predictors of subjective and psy-
choloical well-being. Psychological Reports, 105(1), 205–231. doi:10.2466/PR0.105.1.205-231.
Headey, B., & Wearing, A. J. (1992). Understanding happiness: A theory of subjective well-being. Mel-
bourne: Longman Cheshire.
Inquisit 3.0 [Computer software]. (2012). Seattle: Millisecond Software.
John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical
perspectives. In L. A. Pervin & J. P. Oliver (Eds.), Handbook of personality: Theory and research (2nd
ed., pp. 102–138). New York: Guilford.
Johnson, J. A. (2000). Web-based personality assessment: Project description and rationale. Paper pre-
sented at the 71st annual meeting of the eastern psychological association, Baltimore, MD.
Keyes, C. L., Shmotkin, D., & Ryff, C. D. (2002). Optimizing well-being: The empirical encounter of two
traditions. Journal of Personality and Social Psychology, 82(6), 1007–1022. doi:10.1037/0022-3514.
82.6.1007.
Lucas, R. E., & Diener, E. (2008). Subjective well-being. In M. Lewis, J. M. Haviland-Jones, & L. Feldman
Barrett (Eds.), Handbook of emotions (pp. 471–484). New York: Guilford.
Lucas, R. E., Diener, E., & Suh, E. (1996). Discriminant validity of well-being measures. Journal of
Personality and Social Psychology, 71(3), 616–628. doi:10.1037/0022-3514.71.3.616.
Lykken, D., & Tellegen, A. (1996). Happiness is a stochastic phenomenon. Psychological Science, 7(3),
186–189. doi:10.1111/j.1467-9280.1996.tb00355.x.
McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal
of Personality, 60(2), 175–215. doi:10.1111/j.1467-6494.1992.tb00970.x.
McGregor, I., & Little, B. R. (1998). Personal projects, happiness, and meaning: On doing well and being
yourself. Journal of Personality and Social Psychology, 74(2), 494–512. doi:10.1037/0022-3514.74.2.
494.
Musek, J. (2007). A general factor of personality: Evidence for the Big One in the five-factor model. Journal
of Research in Personality, 41(6), 1213–1233. doi:10.1016/j.jrp.2007.02.003.
Paunonen, S. V., & Ashton, M. C. (2001). Big five factors and facets and the prediction of behavior. Journal
of Personality and Social Psychology, 81(3), 524–539. doi:10.1037/0022-3514.81.3.524.
Paunonen, S. V., & Jackson, D. N. (2000). What is beyond the big five? Plenty! Journal of Personality,
68(5), 821–835. doi:10.1111/1467-6494.00117.
Quevedo, R. J. M., & Abella, M. C. (2011). Well-being and personality: Facet-level analyses. Personality
and Individual Differences, 50(2), 206–211. doi:10.1016/j.paid.2010.09.030.
R Core Team. (2013). R: A language and environment for statistical computing. Vienna: R Foundation for
Statistical Computing.
Ryan, R. M., & Deci, E. L. (2001). On happiness and human potentials: A review of research on hedonic and
eudaimonic well-being. Annual Review of Psychology, 52(1), 141–166. doi:10.1146/annurev.psych.52.
1.141.
Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on the meaning of psychological well-
being. Journal of Personality and Social Psychology, 57(6), 1069–1081. doi:10.1037/0022-3514.57.6.
1069.
Ryff, C. D., & Essex, M. J. (1992). The interpretation of life experience and well-being: the sample case of
relocation. Psychology and Aging, 7(4), 507. doi:10.1037//0882-7974.7.4.507.
Ryff, C. D., & Keyes, C. L. M. (1995). The structure of psychological well-being revisited. Journal of
Personality and Social Psychology, 69(4), 719–727. doi:10.1037/0022-3514.69.4.719.
Ryff, C. D., & Singer, B. H. (1998). The contours of positive human health. Psychological Inquiry, 9(1),
1–28. doi:10.1159/000289026.
Ryff, C. D., & Singer, B. H. (2006). Best news yet on the six-factor model of well-being. Social Science
Research, 35(4), 1103–1119. doi:10.1016/j.ssresearch.2006.01.002.
Schimmack, U., Oishi, S., Furr, R. M., & Funder, D. C. (2004). Personality and life satisfaction: A facet-
level analysis. Personality and Social Psychology Bulletin, 30(8), 1062–1075. doi:10.1177/
0146167204264292.
Schmutte, P. S., & Ryff, C. D. (1997). Personality and well-being: Reexamining methods and meanings.
Journal of Personality and Social Psychology, 73(3), 549–559. doi:10.1037/0022-3514.73.3.549.
Siegler, I. C., & Brummett, B. H. (2000). Associations among NEO personality assessments and well-being
at midlife: Facet-level analyses. Psychology and Aging, 15(4), 710–714. doi:10.1037/0882-7974.15.4.
710.
Soto, C. J. (2014). Is happiness good for your personality? Concurrent and prospective relations of the Big
Five with subjective well-being. Journal of Personality. doi:10.1111/jopy.12081.
Steel, P., Schmidt, J., & Shultz, J. (2008). Refining the relationship between personality and subjective well-
being. Psychological Bulletin, 134(1), 138–161. doi:10.1037/0033-2909.134.1.138.

123
 J. Anglim, S. Grant

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(6),
1063–1070. doi:10.1037/0022-3514.54.6.1063.
Weiss, A., Bates, T. C., & Luciano, M. (2008). Happiness is a personal(ity) thing: The genetics of per-
sonality and well-being in a representative sample. Psychological Science, 19(3), 205–210. doi:10.
1111/j.1467-9280.2008.02068.x.

123