Vous êtes sur la page 1sur 26

EMOTIONAL INTELLIGENCE AND AGE

DOES EMOTIONAL INTELLIGENCE CHANGE WITH AGE?


Paul W.B. Atkins
Australian National University
National Graduate School of Management
Sir Roland Wilson Building, McCoy Circuit
Canberra, ACT 0200
AUSTRALIA
(e-mail: paul.atkins@anu.edu.au)
(tel: +61 2 6125 8979)
Con Stough
Professor in Cognitive Neuroscience
Director, Swinburne Centre for Neuropsychology
Swinburne University of Technology
PO Box 218 Hawthorn Vic 3122
AUSTRALIA
(e-mail: CStough@groupwise.swin.edu.au)
(tel: + 61 3 9214 8167)

July 2005
A draft of this paper was presented at the Society for Research in Adult Development
Annual Conference, Atlanta, US April 6-7 2005.

Acknowledgements: We wish to thank Ben Palmer for providing access to data used in this
analysis.

EMOTIONAL INTELLIGENCE AND AGE

ABSTRACT
What is the relationship between emotional intelligence and age? Studies of changes in
positive and negative effect across the adult lifespan suggest that, despite the additional
emotional challenges that older adults face, they experience more subjective well-being. In
addition, studies of affective processing across the lifespan suggest that older adults may be
more adept at regulating their emotions than younger adults. Taken together, these findings
suggest that particular aspects of emotional intelligence may increase with age. In this
study we examine age effects in three samples using two measures of emotional ability: the
self-report SUEIT and the abilities-based MSCEIT. For the self-report measure effect sizes
were small or non-existent. At the subscale level, it was found that the extent to which
emotions are used in problem solving increases across the lifespan, particularly for females.
In addition, in the larger sample, there was evidence that the ability to control strong
emotions may decline across the lifespan. There were no significant age effects for the
MSCEIT. The implications of these results are explored for emotional intelligence and
developmental theory.

EMOTIONAL INTELLIGENCE AND AGE

What is the relationship between emotional intelligence and age? This question has
profound implications for organisations as the workforce gets older and increasing
emotional demands are made upon workers. Older adults face a variety of challenges that
might be expected to induce negative affect such as the loss of loved ones, diminished
health and unrealised expectations. Yet, paradoxically, the evidence suggests that older
people are more able to maintain and even increase subjective well-being than younger
people. In general, negative affect declines with age while positive affect appears to
increase (Carstensen, Fung, & Charles, 2003; Mroczek, 2001; Mroczek & Kolarz, 1998;
Ryan & Deci, 2001; Staudinger, Marsiske, & Baltes, 1995).
This paradox has been interpreted as showing that older adults are more adept at selfregulating their emotions through either rearranging their environments or acquiring
strategies and capacities to manage their emotions (Carstensen, Pasupathi, Mayr, &
Nesselroade, 2000; Labouvie-Vief & Medler, 2002; Staudinger et al., 1995). It might be
expected that such strategies and capabilities would be reflected in measures of emotional
intelligence that have been proposed to assess how we understand, use and manage
emotions. This study investigates this possibility.
In this study, I first review theories and evidence suggesting changes in emotional
experience across the adult lifespan and then present the construct of emotional intelligence
as measured by the Swinburne University Emotional Intelligence Test (SUEIT) and the
Mayer, Salovey-Caruso Emotional Intelligence Test (MSCEIT). Results are then presented
from three cross-sectional samples of adults: an executive sample and two general samples.
Finally the implications of this work both for theories of affective change across the lifespan
and for the use of the SUEIT as a tool for measuring emotional intelligence are considered.
One of the most systematic studies of emotional experience across the lifespan was
conducted by Carstensen et al. (2000). They sampled 184 adults aged from 18 to 94 years
across one week using an experience sampling method. Older adults experienced less
negative affect than younger adults. Older adults were able to maintain positive moods for
longer and negative moods were terminated more quickly than for younger adults. Older
adults also experienced more mixed and complex emotions, more frequently experiencing

EMOTIONAL INTELLIGENCE AND AGE

both positive and negative emotions simultaneously. This apparent capacity to more
effectively regulate emotions parallels evidence from studies suggesting neuroticism
decreases with age (Costa, Herbst, McCrae, & Siegler, 2000).
Socioemotional Selectivity Theory suggests that these changes result from a growing
awareness of the limited time that the person has left in their lives (Carstensen et al., 2003;
Carstensen et al., 2000). As people realise they have little time left, they are more inclined
to focus on immediate positive emotional experience derived from social interactions than
to invest in acquiring knowledge and experience to better manage the future. Other studies
have demonstrated other ways in which the salience of emotion increases with age. For
example, as people grow older they remember emotional material more easily (Carstensen
& Turk-Charles, 1994), use more emotional content in their representations of other people
(Carstensen & Frederickson, 1998), and make more use of emotions in social problem
solving (Blanchard-Fields, 1997)
Labouvie-Vief and her colleagues (Labouvie-Vief, DeVoe, & Bulka, 1989; Labouvie-Vief &
Medler, 2002) have proposed that there are two orthogonal aspects of emotional experience
during aging: affect optimisation and cognitive-affective complexity. Affect optimisation
involves the capacity to dampen negative affect and enhance and sustain positive affect.
Cognitive-affective complexity involves the capacity to coordinate positive and negative
emotions in more flexible and differentiated cognitive structures. These cognitive structures
allow older adults to differentiate more complex emotions, integrate positive and negative
aspects of affective experience as well as deliberately inhibit, evaluate or analyse aspects of
their emotional experience and its relationship to their identity.
This approach to cognitive-affective complexity refers to the cognitive-affective processes
through which emotions are experienced. Increased cognitive-affective complexity allows
people to differentiate emotions more effectively, perceive themselves and others in a more
flexible, open and tolerant way, recognising and accepting both negative and positive
aspects of themselves and others (Kegan, 1982; Labouvie-Vief et al., 1989; Labouvie-Vief
& Medler, 2002; Loevinger, 1987). Complexity is not simply indexed by the presence or
absence of conflicting emotions (c.f. Carstensen et al., 2000), but rather by the richness of

EMOTIONAL INTELLIGENCE AND AGE

schemata and cognitive processes for integrating experience with the self and managing
emotional experience. With increasing maturity:
the individual is able to acknowledge conflicting feelings within self and other.
Overall, the language of self-regulation becomes more vivid and specific, and
less stereotypical. Mature cognitive-emotional complexity thus is evidenced by
a language that is complex, nonstereotypical and nondualistic; that tolerates
intra- and interindividual conflict; and that appreciates the uniqueness of
individual experience." (Labouvie-Vief et al., 1989: 426)
Studies of ego development (Loevinger, 1976), which is related to the level of cognitiveaffective complexity (Labouvie-Vief & Medler, 2002), have suggested that cognitiveaffective complexity is likely to be unrelated to the capacity to optimise positive affect. For
example, stage of ego development has consistently been found to be unrelated to emotional
adjustment as measured by measures of life satisfaction and positive affect (Helson & Wink,
1987; Labouvie-Vief & Medler, 2002; McCrae & Costa, 1983; Vaillant & McCullough).
The literature on ego development is relevant in other ways to the question of emotional
change with aging. Loevingers theory of ego development emphasises four inter-related
domains of development: character development, cognitive style, interpersonal style, and
conscious preoccupations (Loevinger, 1976; Manners & Durkin, 2001). Character
development refers to changes in the focus of moral concerns and moral behaviours as well
as impulse control. Cognitive style refers most particularly to changes in conceptual
complexity but is reflected in increasing tolerance of ambiguity and more reflective and
objective processing of experience. Interpersonal style refers to the orientation and
attitudes a person has towards others and relationships with others, particularly with regard
to belonging, independence and interdependence. It includes preferences in the type of
relationship and understanding of relationships. Conscious preoccupations refers to the
focus of attention of the persons thoughts and behaviour and includes issues such as the
extent to which they conform to social rules, take responsibility, and seek and realize
independence from others.

EMOTIONAL INTELLIGENCE AND AGE

In short, and emphasizing the affective domain of most interest here, individuals at later
stages of ego development have better impulse control, have more complex and integrated
cognitive-affective processes, are more likely to respect, tolerate and collaborate with
others, and are more likely to be able to differentiate, communicate and use emotions in
their thinking and problem solving. Manners and Durkin (2001: 543) summarise this
developmental trajectory as follows: "Each sequential stage represents a restructuring of
the self-system toward greater self and interpersonal awareness, conceptual complexity,
flexibility, personal autonomy, and responsibility." The development of a positive identity,
or core self, appears to be critical for integrating negative affect and maintaining resiliency.
Labouvie-Vief and Medler (2002) distinguished between 4 possible combinations of affect
optimisation and cognitive-affective complexity: integrated (high optimisation, high
complexity), complex (low optimisation, high complexity), defended (high optimisation,
low complexity) and dysregulated (low optimisation, low complexity). They found evidence
that older adults tended to be more integrated or defended than younger adults. In other
words, aging was consistently associated with increasing affect optimisation while
cognitive-affective complexity peaked in mid-life and then, for some people at least,
declined in later life perhaps as a result of declining cognitive resources (Labouvie-Vief et
al., 1989; Labouvie-Vief & Medler, 2002). This result reflects the finding that older adults
exhibit decreasing openness to experience (Costa et al., 2000).
However, it is worthwhile noting that age effects in the studies mentioned above were
usually very small. For example, Labouvie-Vief et al. (1989) reported a significant effect of
age on levels of emotional understanding after controlling for ego development1. However,
this effect disappeared when the 10-18 year old groups were excluded from the analysis and
only adults were considered. The authors concluded that affective development in
adulthood is not best indexed by age. Labouvie-Vief and Medler (2002) also showed no
effect of age on cognitive-affective complexity after controlling for education, SES and
1 Note that, since age was highly correlated with ego development (r = .61), this is an
extremely stringent test of an age effect however the authors did not report the effect of age
uncontrolled for ego development.

EMOTIONAL INTELLIGENCE AND AGE

relationship quality, although they did demonstrate a significant relationship between age
and affective optimisation. Carstensen et al. (2000) reported a nonsignificant correlation of .
07 between age and positive emotion. This was comparable to the significant correlation of .
10 in a much larger sample reported by Mroczek and Kolarz (1998). Thus the relationship
between age and emotional variables are has been modest even in studies that relied upon
the collection of much more extensive and detailed data than the self-report measure
described below.
The distinction between optimisation and complexity may be characterised in terms of
differing goals for affective regulation. Other authors have proposed a similar distinction
between a hedonic approach to emotion regulation that focuses on increasing positive
affect and decreasing negative affect, and a eudaimonic approach that focuses on
representing the world more accurately in complex, multivalenced ways, thereby increasing
individuation and meaning in life (Ryan & Deci, 2001).
Mayer & Saloveys (1997) notion of emotional intelligence is related to this idea of
cognitive-affective complexity, recognising as it does that development is accompanied by
an increasing ability to perceive, understand and manage emotions. However, it may also
incorporate aspects of affective optimisation. Specifically the capacity to facilitate positive
emotional experiences in self and others seems to reflect more of a hedonic than eudaimonic
emphasis. We might therefore, expect that subscales within a measure of emotional
intelligence might show different patterns of change across the lifespan.
In summary, there is evidence that older adults are more adept at regulating their emotional
experience and optimising positive affect. Further there is evidence that adults, in mid life at
least, are more likely than younger adults to be able to represent and manage more complex
blended emotions and integrate their emotions with cognitions about the world in more
complex ways. However, where changes in cognitive-affective complexity across the
lifespan have been demonstrated the effect sizes have been small.

EMOTIONAL INTELLIGENCE AND AGE

EMOTIONAL INTELLIGENCE
The construct of emotional intelligence refers to the extent to which individuals deal
effectively with their own and others emotions. There has been a tremendous amount of
popular and academic interest in emotional intelligence, probably reflecting the intuitively
appealing idea that emotions are important determinants of successful performance in
various domains including leadership, workplace performance and life satisfaction (Gardner
& Stough, 2002; Palmer, Gardner, & Stough, 2003b).
At present, there are two broad categories of measures of emotional intelligence; ability
models (e.g. Mayer & Salovey, 1997) and self-report, mixed-models that assess a variety of
individual characteristics including mental abilities, motivational, dispositional and
personality-like elements (e.g. zeal, persistence, warmth and optimism). Bar-Ons
noncognitive EQ:i measure (Bar-On, 1997) and Golemans (e.g. Goleman, 1998)
competency-based model are the two best-known examples of self-report measures of
emotional intelligence.
Emotional intelligence and age
To date there appear to be few studies examining the relationships between age and
emotional intelligence. One study of 3821 US and Canadian citizens reportedly found a
significant increase in emotional intelligence with age, at least up until the early 50s using
the BarOn EQ:i (Multi-Health Systems Inc., 1997, March 3, cited in Rippeth, 2003: 26).
However, this relationship was limited to the subscales, independence, social responsibility,
optimism, reality testing and problem solving (Rippeth, 2003)2. With the exception of
problem solving, these subscales are not usually included within models of emotional
intelligence emphasising emotional abilities.
Mayer and Saloveys (1997) hierarchical model of emotional abilities explicitly includes a
developmental progression. The four branches of this model, emotion perception,
emotional facilitation, emotional understanding and emotional regulation and control are
2 To date, the authors have been unable to obtain this report to examine the details of the reported
relationships between age and emotional intelligence.

EMOTIONAL INTELLIGENCE AND AGE

hypothesized to unfold progressively and incrementally through development. Emotional


perception arises in early childhood. Emotional facilitation is the ability to bring emotional
experiences into general awareness. Emotional understanding includes reasoning about
emotions including understanding how and why they arise. The final, and highest, ability is
emotional management and regulation.
Mayer, Caruso and Salovey (2000) compared 229 adolescents with approximately 250
adults using the MEIS scale and found significant increases by age in the Faces, Synesthesia
and Blends tasks as well as in the overall combined test scores. These tasks are categorised
under the branches of perceiving, using and understanding emotions, not managing
emotions. However, this study involved a comparison of adolescents with adults and
therefore says nothing about changes in emotional abilities across the adult lifespan.
Furthermore, it is notable that only 3 of the 12 subscales of the MEIS showed significant
age effects even though adolescents were included.
A quite different pattern of results was obtained in the normative sample of 5000 people
drawn from the general population for the MSCEIT reported in the manual (Mayer,
Salovey, & Caruso, 2002). There Perceiving Emotions showed no increase across the
lifespan, while Facilitating/Using, Understanding Emotions and Managing Emotions
all showed reliable differences between the 18-24 year old age bracket and later ages,
although not between ages later than 25. Age explained the most variance in Understanding
Emotions (8.7%) particularly for the Changes subtask (10.4%). It explained very small
portions of variance in Managing Emotions (5.9%) and Facilitating/Using emotions (3.1%).
In short, effect sizes of age appear to be very small however the strongest evidence for age
effects is in the more cognitive branch of Understanding the causes of emotions and being
able to differentiate and label emotions.
In the present study, we primarily examined the relationship between emotional intelligence
and age using a new measure known as the Swinburne University Emotional Intelligence
Test (SUEIT). This model was derived from an extensive factor analytic study of existing
measures to derive a brief self-report test of five main emotional intelligence factors that are

EMOTIONAL INTELLIGENCE AND AGE

reliably measured by the Bar-On EQi, MSCEIT, and four other measures of emotional
intelligence. The SUEIT conceptualises emotional intelligence in terms of five factors:
1. Emotional Recognition and Expression;
2. Understanding Emotions;
3. Emotions Direct Cognition;
4. Emotional Management;
5. Emotional Control.
The workplace SUEIT is specifically designed for measuring emotional intelligence in a
work context. Thus the items are more specifically targeted on work than is the case for
other tests of emotional intelligence. This is one respect in which this study differs from
previous studies and we consider the implications of this in the discussion. Although the
focus of this study was not on the MSCEIT, we also conducted analyses by age of this
measure as this is not a self-report but an abilities based measure.
The factor Emotional Recognition and Expression reportedly assesses individuals' capacity
to perceive and express their own emotions. However, of the 10 items used to measure this
factor, 8 are concerned exclusively with emotional expression, not an ability suggested by
developmental theory to be necessarily associated with aging. The remaining two items
refer to detecting and distinguishing emotions at work. Given the context specificity of
these items, and the fact that only two items are specifically devoted to emotion perception.
We did not expect this factor to be correlated with age.
The factor Understanding Emotions measures individuals' capacity to perceive and
understand the emotions of others. This factor is assessed by nearly a third of the test items
in the SUEIT. As discussed above, there is considerable evidence that emotion perception
increases during childhood and perhaps adolescence. However it is unclear whether this
ability continues to improve throughout adulthood. Mayer and Saloveys (1997) theory of
emotional abilities would suggest that emotion perception is one of the earliest acquired
emotional capabilities. On the other hand, Mayer et al (2000) reported improvements on the
faces subtask of the MEIS between adolescence and adulthood, indicating that at least some

10

EMOTIONAL INTELLIGENCE AND AGE

improvement in the ability to discern emotions in others may continue through at least early
adulthood. On balance, we did not expect this ability to be strongly correlated with age
within the adult sample analysed herein.
The factor Emotions Direct Cognition assesses the extent to which individuals utilise
emotions and emotional information in reasoning and decision-making. Given the
developmental evidence presented above, we would expect this ability to be positively
related to age and, in particular, the development of cognitive-affective complexity.
Socioemotional selectivity theory (Carstensen et al., 2003; Carstensen et al., 2000) suggests
that with age, adults make more use of emotions in goal-setting and in evaluating courses of
action. Aging is also associated with an improvement in memory for the emotional aspects
of experiences (Carstensen & Turk-Charles, 1994). The studies by Labouvie-Vief and
Loevingers ego development theory both suggest that, with higher stages of ego
development, affective and cognitive processes should become more complex and
integrated. To the extent that ego development is correlated with age, we might expect
Emotions Direct Cognition to increase. There is also quite direct evidence that older people
make more use of emotion in problem solving about interpersonal matters (BlanchardFields, 1997).
The factor Emotional Management assesses individuals' capacity to effectively regulate and
manage both their own emotions and the emotions of others. Half of the items measuring
this factor refer to management of ones own emotions while the other half refer to
management of others emotions. Given the developmental studies and theories presented
above, we expected that self-management of emotions at least should be related positively
with age.
The final factor, Emotional Control, measures individuals' capacity to effectively control
strong emotions such that they do not affect their performance at work. Given the evidence
that older adults appear to be more able to sustain positive emotions and limit the
persistence of negative emotions, we hypothesised that this ability should be positively
related to age. Older adults should be more adept at ensuring strong emotions dont interfere
with their work.

11

EMOTIONAL INTELLIGENCE AND AGE

Thus, in summary, we hypothesised the following relationships between the factors of the
SUEIT and age:
H1: Age should be positively related to Emotions Direct Cognition
H2: Age should be positively related to the self-management items from Emotional
Management.
H3: Age should be positively related to Emotional Control.
METHOD
Two samples were analysed for this study. The first sample included senior executives
defined as earning over AUD$ 200K per annum and with 250 staff reporting to them. They
were recruited as part of a larger study in which 1000 top-level Australian executives were
administered the SUEIT in order to develop an executive database. The first three hundred
of these executives were also assessed for leadership competencies with the MLQ 5x (Bass
& Avolio, 2000) and measures of intuition (See Gardner & Stough, 2002, Downey,
Papageorgiou & Stough, submitted).
The second sample was a convenience sample drawn from the general population and
represented employees from a wide range of Australian companies and
workplaces/industries. This sample could be differentiated from the first sample in that they
were drawn from lower levels of management or were not involved in any management/and
or leadership responsibilities. Participants in both samples did not receive feedback and
were not paid to participate in the study. Participants in the second sample only completed
the Workplace SUEIT.
The third sample consisted of 365 people drawn from the general population not necessarily
at work. This sample was drawn for a large factor analytic study to determine the most
definitive and common elements of the construct emotional intelligence (Palmer, Gardner,
& Stough, 2003a). Participants were given six of the predominant measures of emotional
intelligence available at the time including the MSCEIT (Mayer, Salovey, & Caruso, 1999),
the Bar-On EQ:i (Bar-On, 1997), the Trait Meta-Mood Scale (Salovey, Mayer, Goldman,
Turvey, & Palfai, 1995), the Toronto Alexithymia Scale-II (Bagby, Taylor, & Parker, 1994),

12

EMOTIONAL INTELLIGENCE AND AGE

the scale by Scutte et al. (Schutte et al., 1998) and the scale by Tett et al. (Tett, Wang,
Thomas, Griebler, & Linkovich, 1997). Only the results from the MSCEIT are presented
here as the main point of this comparison was to consider whether a non-self-report measure
of emotional intelligence revealed effects of age.
Age
18-29
30-39
40-49
50-59
60-72
Total
22-29
30-39
40-49
50-59
Total
18-29
30-39
40-49
50-59
60-69
70-78
Total

Female

Male
General Sample
183
44
241
140
267
164
167
71
20
14
878
433
Executive Sample
16
8
47
67
37
75
20
33
120
183
General MSCEIT sample
81
46
52
14
62
19
46
20
11
6
4
1
256
106

Total
227
381
431
238
34
1311
24
114
112
53
303
129
66
81
67
17
5
365

Table 0: Age distribution of the three samples.


RESULTS
Executive Sample
It is possible that patterns of development of emotional capabilities during adulthood may
be linear, quadratic (rising during middle adulthood and then declining in later life), or
logarithmic (rising initially and then asymptoting during later life). Inspection of
scatterplots of the data suggested that, where relationships may be present, they were most

13

EMOTIONAL INTELLIGENCE AND AGE

likely to be linear or logarithmic. This was confirmed by conducting multiple regressions


using either linear, logarithmic or quadratic transformations of age.
R
Fem Male

Linear Logarithmic

Quadratic

(age)

(age^2)

Emotional
Recognition/Expression .03

.02

-.01

-.01

-.23

.23

Understanding Emotions .17

-.04

.01

.01

.32

-.32

.16**

.17**

.58

-.42

Emotions Direct
Cognition

.29** .12

Emotional Management .18

.03

.07

.08

.60

-.54

Emotional Control

-.08

-.05

-.05

.44

-.49

.00

Table 0: Bivariate correlations between age and various aspects of EI. (Note that these
correlations are not corrected for unreliability in the SUEIT). Standardised beta
weights and F values for regressions of linear, logarithmic and quadratic transformed
age on the emotional intelligence variables.
Overall, only the subscale Emotions Direct Cognition was significantly related to age and
was best modelled by either a linear or logarithmic function. Given the strong theoretical
and empirical reasons for supposing that changes in emotional abilities might be nonlinear
across the lifespan, only the log-transformed analyses are presented below although both the
transformed and untransformed data produced an identical pattern of results.
A multivariate linear model was fitted with the five subscale scores as dependent variables.
Both age (F(5,295) = 3.29, p < .01) and sex (F(5,295) = 3.23, p < .01) were significantly related to
EI with older adults scoring more highly and females scoring more highly than males. At
the subscale level, only Emotions Direct Cognition was significantly related to age (F(1,299) =
11.3, p < .001) while Emotional Recognition and Expression (F(1,299) = 9.1, p = .003),
Understanding Emotions (F(1,299) = 10.2, p = .002) and Emotions Direct Cognition (F(1,299) =
6.7, p =.01) were all related to sex with females scoring higher than males.

14

EMOTIONAL INTELLIGENCE AND AGE

SD

Female Male

All

Female Male

All

Emotional Recognition/Expression

41.5

39.7

40.4

4.5

5.1

5.0

Understanding Emotions

79.9

77.0

78.2

7.6

7.9

7.9

Emotions Direct Cognition

37.9

36.2

36.9

6.8

7.6

7.3

Emotional Management

44.7

44.2

44.4

5.1

5.0

5.0

Emotional Control

33.8

33.7

33.7

4.2

4.4

4.3

237.7 230.8 233.5

20.3

Total Score

21.0 21.0

Table 0: Means and standard deviations for males and females.


A second analysis was conducted using age, sex and level of education as predictors. The
power of this analysis was reduced as only 235 individuals had recorded level of education.
Overall age (F(5,227) = 4.91, p < .001), sex (F(5,227) = 2.72, p = .02) and level levels of
education producing higher EI scores3. At the subscale level, the pattern of results for age
and sex was the same. Level of Education was related to Understanding Emotions (F(1,231) =
4.4, p = .036) and Emotional Control (F(1,231) = 4.0, p = .048) with more educated
participants gaining higher scores.
Although the relationship to age was significant for females but not males in separate
analyses, a multiple regression with age, sex and the interaction term fitted revealed that
there was no significant interaction between age and sex for Emotions Direct Cognition ( =
-0.06, t = 1.05, p > .05)
General Population Sample
As before linear, logarithmic and quadratic regressions were applied to age to examine
curve fit (Table 4).

3 Interestingly the effect size for age was larger for this analysis than for the analysis
without level of education. However, because this was on a different sample, it was
impossible to tell whether this was due to a masking effect of level of education.

15

EMOTIONAL INTELLIGENCE AND AGE

R
Fem Male

Linear Logarithmic

Quadratic

(age)

(age^2)

Emotional
Recognition/Expression -.03

-.06

-.04

-.04

-.03

-.01

Understanding Emotions -.04

-.05

-.05

-.05

-.10

.05

.09**

.47*

-.39*

.02

-.09

.11

-.08**

-.33

.26

Emotions Direct
Cognition

.09** .13** .08**

Emotional Management .00


Emotional Control

.06

.03

-.08* -.11* -.07*

Table 0: Bivariate correlations between age and various aspects of EI. (Note that these
correlations are not corrected for unreliability in the SUEIT). Standardised beta
weights and F values for regressions of linear, logarithmic and quadratic transformed
age on the emotional intelligence variables.
Consistent with the executive sample, in the general population sample, Emotions Direct
Cognition was related to age. In addition Emotional Control was also related to age but
negatively not positively. Again, the relationships were best modelled using a logarithmic
function.
For the multivariate analysis, both age (F(5,1302) = 9.48, p < .001) and sex (F(5,1302) = 12.1, p < .
001) were significantly related to EI with older adults scoring more highly and females
scoring more highly than males. At the subscale level, both Emotions Direct Cognition
(F(1,1306) = 15.1, p < .001) and Emotional Control (F(1,1306) = 9.7, p < .002) were significantly
related to age while all of the subscales except Emotional Management were significantly
related to sex with females scoring marginally higher than males for Emotional Recognition
and Expression, Understanding Emotions and Emotions Direct Cognition but less than
males for Emotional Control.

16

EMOTIONAL INTELLIGENCE AND AGE

SD

Female Male

All

Female Male

All

Emotional Recognition/Expression

39.0

38.4

38.8

5.4

5.1

5.3

Understanding Emotions

77.0

76.0

76.7

7.6

7.8

7.7

Emotions Direct Cognition

36.0

33.7

35.2

5.9

6.3

6.2

Emotional Management

42.2

42.5

42.3

5.7

5.7

5.7

Emotional Control

32.5

33.3

32.8

4.6

4.5

4.6

226.6 223.9 225.7

19.9

Total Score

19.7 19.9

Table 0: Means and standard deviations for males and females.

Item and inter-scale analyses


At the level of individual items, four items from the Emotions Direct Cognition scale were
significantly related to age using an alpha level controlling for Type I error with a
Bonferroni adjustment. Three of these assessed the extent to which emotions were used and
relied upon in decision making while the fourth referred to the extent to which emotional
reactions to situations at work helped the respondent to remember them. Older adults were
more likely to use emotions in decision making, problem solving, and remembering than
were younger adults
Three items from the Emotional Control subscale were significantly, and negatively, related
to age. One of these assessed the extent to which stressful situations at work cause irritation,
another assessed the extent to which difficult situations elicit emotions at work that are hard
to overcome while the third referred to the extent to which the participant was able to
control their anger at work. Contrary, to expectations, older adults report they have more
difficulty controlling irritation and anger at work and overcoming strong emotions.
The pattern of rising scores for Emotions Direct Cognition and falling scores for Emotional
Control suggests the possibility that these two subscales may be negatively related to one
another. Although the SUEIT technical manual reports positive associations between all
subscales in this test, the correlation between EDC and EC is the only nonsignificant one.

17

EMOTIONAL INTELLIGENCE AND AGE

For this sample, all inter-correlations between subscales were significant at alpha = .05
however EDC and EC were negatively related.
Emotion
Emotional
s Direct Emotional Emotion
Recognition & Understandin Cognitio Manageme
al
Expression
g Emotions
n
nt
Control
Emotional Recognition
& Expression

1.00

Understanding
Emotions

.52

1.00

Emotions Direct
Cognition

.26

.26

1.00

Emotional Management

.38

.54

.07

1.00

Emotional Control

.20

.35

-.15

.70

1.00

Table 0: Correlations between subscales on the SUEIT for the larger samples. Note
that all correlations are significant at p < .01.
Given the negative relationship between EDC and EC in this sample, a posthoc analysis was
conducted to explore whether the relationship between these scales changed with age.
Contrary to expectations, the magnitude of the negative relationship decreased with age
suggesting that for younger adults making more use of emotions in decision making was
negatively associated with being able to control strong negative emotions. However, this
negative relationship diminished with age (Table 7).
Age

N r (EDC v EC)

18-29

44

-.37

30-39

140

-.30

40-49

163

-.18

50-59

71

-.13

60+

14

.03

Table 0: Correlations between Emotions Direct Cognition and Emotional Control


across 5 age bands.

18

EMOTIONAL INTELLIGENCE AND AGE

General sample with the MSCEIT


There were no significant relationships between age and any of the dimensions of the
MSCEIT. Indeed the bivariate correlations approached zero for nearly all of the scales
(Table 8). Examination of logarithmic and quadratic transformations of age also revealed no
significant effects.
Branch

Scale

Perceiving

Faces

-.04

Pictures

-.03

Using
Understanding
Managing

Facilitations

.00

Sensations

.01

Changes

.00

Blends

-.01

Emotional management

-.04

Social Management/
Relationship
Management

-.05

Table 0: Bivariate correlations between age and the various dimensions of the
MSCEIT.
A multivariate analysis including age, sex and level of education revealed no significant
effects in terms of either main effects or subscale effects.
DISCUSSION
In summary, a positive overall multivariate effect of age was obtained for the two samples
measured on the SUEIT but the effect sizes were extremely small. At the subscale level, age
was positively related to Emotions Direct Cognition for both samples and negatively related
to Emotional Control for the larger popular sample. There were no significant effects of age
on the MSCEIT even controlling for sex and level of education. It is necessary to consider
both why some subscales showed significant relationships to age and also why the majority
of subscales from both tests were not related to age. We consider each of these in turn.

19

EMOTIONAL INTELLIGENCE AND AGE

The data from the SUEIT provided weak support for the claim that age is monotonically
related to emotional intelligence. Logarithmic, rather than quadratic models provided a
better fit to the data both in a positive (Emotions Direct Cognition) and negative (Emotional
Control) direction.
Hypothesis 1 was supported: Older adults, particularly older female executives, reported
making more use of emotions in problem solving and decision making than their younger
counterparts. This may be indicative of greater cognitive-affective complexity in that
emotions and cognitions are more directly integrated in problem solving. It is particularly
interesting that the item regarding using emotions to remember events was significant
correlated with age in this study. Carstensen and Turk-Charles (Carstensen & Turk-Charles,
1994) found that older adults are more likely to remember the emotional content of events.
Therefore the presence of an apparent memory effect in this sample both replicates past
findings and provides some validation evidence that the SUEIT is measuring real changes in
emotional ability.
Hypothesis 2 was not supported: There were no significant age effects for Emotional
Management either for the items measuring management of ones own emotions or for the
management of others emotions. This was surprising as one of the most robust findings in
the adult development literature appears to be the capacity for older adults to regulate their
own emotions to minimise both the occurrence of, and the duration of negative affect (e.g.
Carstensen et al., 2000). The six self-management items all referred to the capacity to
generate positive emotions or the capacity to minimise either the intensity or the duration of
negative affect. This result is, in our view, problematic for the SUEIT.
Although a significant relationship between age and Emotional Control was obtained, it was
in the opposite direction to that predicted in Hypothesis 3. Younger adults reported being
better able to prevent strong emotions from interfering with their work than either middleage or older adults. This result contradicts the claim, on self-report measures, older adults
are more likely to claim that they successfully control their emotions because they feel that
they should be able to do so (Carstensen et al., 2000).

20

EMOTIONAL INTELLIGENCE AND AGE

One possible interpretation of this result given the developmental literature presented above
relies upon the fact that the SUEIT was a self-report measure and therefore asked
participants to remember and reflect upon global aspects of their emotional experience.
One intriguing possibility for explaining this pattern of results emphasises differences in the
way in which older adults construe their emotional experiences when answering a global
self-report measure such as the SUEIT. If older adults attend to emotional experiences more
closely and remember emotional aspects of their experience more than younger adults
(Carstensen et al., 2003), then it seems likely that they would simultaneously self-report
both using emotions more in decision making and problem solving as well as well as being
less able to prevent strong emotions from interfering with their work. Carstensen et al.
(2003) suggested that, during late adulthood, declines in attention and working memory
capacity may make it more difficult for older adults to inhibit emotions from flooding
output.
Unfortunately, posthoc analyses of the correlations between the Emotions Direct Cognition
and Emotional Control subscales argued against this interpretation. Although, these two
scales were the only scales that were negatively correlated overall (Table 6), this was
particularly true for the younger participants and the negative correlation regularly
diminished with age (Table 7). Therefore younger adults who report that they make more
use of emotions in decision-making do appear to find it more difficult to control strong
negative emotions and prevent them from interfering with their work. However, this
becomes progressively less true for older adults. It is difficult to explain this result in terms
of current developmental theory. Although it might seem to suggest that older adults are less
impulsive, the negative relationship between Emotional Control and age would argue
against this interpretation.
Perhaps this result arises from differences between age groups in the way the questions
were interpreted. Aside from being more able to remember their emotional experiences,
older adults may also apply different standards for evaluating when they have successfully
managed their emotions. The studies reported earlier showing that older adults are more
likely to seek to minimise negative affect and maximise positive affect are consistent with

21

EMOTIONAL INTELLIGENCE AND AGE

this interpretation. An experience of strong negative affect may be more aversive and
intrusive for an older adult than for a younger adult. If this were the case, then older adults
would be more likely to record more negative responses to the Emotional Control items
overall. However, this suggestion must be weighed against the claim by Carstensen et al.
(2003) that, with age, autobiographical memories become increasingly positive not negative
over time.
One might be tempted to wonder if the decline in Emotional Control has anything to do
with age-related declines in attention and working memory. Older adults may lack the
necessary cognitive resources to successfully prevent negative emotions from intruding on
their working lives. However, this is not a plausible explanation of these results because the
decline in Emotional Control did not occur exclusively in late adulthood. Adults in mid-life
are unlikely to be experiencing significant cognitive declines.
Looking at these results from a different perspective, one might ask why the effect sizes
were generally small. The only correlation to exceed r = .2 was for females on the Emotions
Direct Cognition scale. To some extent this is consistent with the small effect sizes reported
in the introductory section of this paper, even when much more intensive modes of data
collection such as interviews and experience sampling are used. Mayer et al. (2000) only
reported an age effect in a comparison of adolescents with adults and, even then, in only 3
of 12 subscales. It would appear that it is quite difficult to detect mean changes in emotional
capabilities across the adult lifespan without including adolescents. Indeed, this study
appears to be reasonably unusual in reporting any significant findings at all, albeit quite
small ones.
Age effects in the SUEIT might also have been attenuated somewhat by the self-report
nature of the scale. As discussed above, there is evidence that older adults pay more
attention to, and remember more emotional material. Carstensen et al. (2000) point out that
global self-report measures evaluating one's own performance are going to be highly
cognitive, involve comparisons with the past and present and involve idiosyncratic
standards. It is possible that older adults adopt higher standards for judging the success of
their attempts to control emotions or even for what is considered to be a difficult situation.

22

EMOTIONAL INTELLIGENCE AND AGE

If, for example, older adults have higher expectations of themselves with regard to
emotional control, then they might be more likely to report a higher frequency of being
unable to control their strong emotions adequately. But the MSCEIT is not a self-report
measure, it relies on emotional problem solving to measure emotional abilities and, if
anything, it was even less related to age than the SUEIT.
The self-report format of the SUEIT raises another interesting issue. To what extent can a
person adequately verbalise their emotional skills? Zeidner, Matthews, Roberts, and
McCann (2003: 71) note that ... much emotional behaviour, ranging from generating facial
expressions to responding to non-verbal social cues, appears to be implicit, depending on
procedural skills that are inaccessible to conscious awareness. However, it seems unlike
that the small effect sizes for age reported above resulted from an inability to verbalise selfknowledge of ones own emotional processes. Many of the items describe behaviours as a
result of ones emotional experience and people are certainly aware of the effectiveness of
their own behaviours. In addition, if this were a reasonable explanation, then we would not
expect such instruments to be predictive of objective behaviours such as leadership and life
satisfaction which they appear to be.
Of course, being a cross-sectional study it is possible that cohort effects influenced all of the
results described above. Even if a positive developmental trend were present in emotional
intelligence, it might have been obscured by such affects. For example, it is possible that
younger adults have learned more effective strategies for managing and controlling
emotions than older adults, perhaps because of changes in parenting or in the valuing of
emotional skills by society and in the workplace. Labouvie-Vief et al. (1989: 433) made a
similar point with regard to their finding of a levelling off of complexity of emotional
understanding in adulthood, patterns of coping and defense would seem to be strongly
influenced by sociocultural conditions, and it is possible that the oldest cohorts received
more rigid training about how to self-regulate emotions. Hence, what appears as a lack of
continued development could reflect enhanced development in more recent cohorts. In
some respects, this is the most interesting suggestion to arise from this research. Clearly

23

EMOTIONAL INTELLIGENCE AND AGE

longitudinal studies of changes in emotional intelligence are required to explore this


possibility.
In terms of further research, there is a need to explore the finding that older, female
executives were more likely than younger or male executive to self-report using emotions in
decision-making, problem-solving and memory. To our knowledge, this is the first
demonstration of this result and the implications for the leadership and management
literatures are substantial. Further research is also required regarding the significant decline
in emotional control found across the lifespan in the largest sample in this study. Given the
strength of the literature arguing for increases in emotional control, it seems most likely that
this result arose from memory and emotional salience effects associated with the way in
which items within a self-report measure of emotional intelligence are construed.

References

Bagby, R. M., Taylor, G. J., & Parker, J. D. A. (1994). The twenty-item Toronto alexithymia
scale - II: Convergent, discriminant, and concurrent validity. Journal of
Psychosomatic Research, 38, 33-40.
Bar-On, R. (1997). Emotional quotient inventory (EQ-i): Technical Manual. Toronto,
Canada: Multi-Health Systems.
Bass, B. M., & Avolio, B. J. (2000). Multifactor Leadership Questionnaire for Research.
Palo Alto, CA: Mind Garden.
Blanchard-Fields, F. (1997). The role of emotion in social cognition across the adult life
span. In K. W. Schaie & M. P. Lawton (Eds.), Annual review of gerontology and
geriatrics (Vol. 17, pp. 238-265). New York: Springer.
Carstensen, L. L., & Frederickson, B. (1998). Socioemotional selectivity in healthy older
people and ounger people living the the Human Immunodeficiency Virus: The
centrality of emotion when the future is constrained. Health Psychology, 17, 1-10.
Carstensen, L. L., Fung, H. H., & Charles, S. T. (2003). Socioemotional selectivity theory
and the regulation of emotion in the second half of life. Motivation and Emotion,
27(2), 103-123.
Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000). Emotional
experience in everyday life across the adult life span. Journal of Personality and
Social Psychology, 79(4), 644-655.
Carstensen, L. L., & Turk-Charles, S. (1994). The salience of emotion across the adult life
course. Psychology & Aging, 9, 259-264.

24

EMOTIONAL INTELLIGENCE AND AGE

Costa, P. T., Jr., Herbst, J. H., McCrae, R. R., & Siegler, I. C. (2000). Personality at midlife:
Stability, intrinsic maturation, and response to life events. Assessment, 7(4), 365378.
Gardner, L., & Stough, C. (2002). Examining the relationship between leadership and
emotional intelligence in senior level managers. Leadership & Organization
Development Journal, 23(2), 68-78.
Goleman, D. (1998). Working with emotional intelligence. London: Bloomsbury Publishing.
Helson, R., & Wink, P. (1987). Two conceptions of maturity examined in the findings of a
longitudinal study. Journal of Personality and Social Psychology, 53, 531-541.
Kegan, R. (1982). The Evolving Self. Cambridge, MA: Harvard University Press.
Labouvie-Vief, G., DeVoe, M., & Bulka, D. (1989). Speaking about feelings: Conceptions
of emotion across the life span. Psychology & Aging, 4(4), 425-437.
Labouvie-Vief, G., & Medler, M. (2002). Affect optimization and affect complexity: Modes
and styles of regulation in adulthood. Psychology and Aging, 17(4), 571-588.
Loevinger, J. (1976). Ego development: Conceptions and theories. San Francisco: Jossey
Bass.
Loevinger, J. (1987). Paradigms of personality. New York: Freeman.
Manners, J., & Durkin, K. (2001). A critical review of the validity of ego development
theory and its measurement. Journal of Personality Assessment, 77(3), 541-567.
Mayer, J. D., Caruso, D., & Salovey, P. (2000). Emotional intelligence meets traditional
standards for an intelligence. Intelligence, 27(4), 267-298.
Mayer, J. D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey & D.
Sluyter (Eds.), Emotional development and emotional intelligence: Implications for
educators (pp. 3-31). New York: Basic Books.
Mayer, J. D., Salovey, P., & Caruso, D. (1999). Instruction manual for the MSCEIT: Mayer
Solovey-Caruso Emotional Intelligence Test. Toronto, CA: Multi-Health Systems.
Mayer, J. D., Salovey, P., & Caruso, D. (2002). Mayer-Salovey-Caruso Emotional
Intelligence Test (MSCEIT): User's Manual. New York: Multi-Health Systems.
McCrae, R. R., & Costa, P. T. (1983). Psychological maturity and subjective well-being:
Toward a new synthesis. Developmental Psychology, 19(2), 243-248.
Mroczek, D. K. (2001). Age and emotion in adulthood. Current Directions in Psychological
Science, 10(3), 87-90.
Mroczek, D. K., & Kolarz, C. M. (1998). The effect of age on positive and negative affect:
A developmental perspective on happiness. Journal of Personality and Social
Psychology, 75, 1333-1349.
Palmer, B., Gardner, L., & Stough, C. (2003a). Measuring emotional intelligence in the
workplace with the Swinburne University Emotional Intelligence Test. Paper
presented at the First International conference on contemporary management
(ICCM): emotional intelligence in organisations, University of Adelaide.
Palmer, B., Gardner, L., & Stough, C. (2003b). The relationship between emotional
intelligence, personality and effective leadership. Australian Journal of Psychology,
55, 140.
Rippeth, R. A. M. (2003). An investigation of the relationship between emotional
intelligence and family environment, ego development and alexithymia.

25

EMOTIONAL INTELLIGENCE AND AGE

Dissertation Abstracts International Section A: Humanities & Social Sciences, 64(2A), 410.
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
2000, US http://www.
Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, t. P. (1995). Emotional
attention, clarity and repair: Exploring emotional intelligence using the trait metamood scale. In J. W. Pennebaker (Ed.), Emotion, disclosure and health (pp. 125154). Washington, DC: American Psychological Association.
Schutte, N. S., Malouff, J. M., Hall, L. E., Haggerty, D. J., Cooper, J. T., Golden, C. J., et al.
(1998). Development and validation of a measure of emotional intelligence.
Personality & Individual Differences, 25, 167-177.
Staudinger, U. M., Marsiske, M., & Baltes, P. B. (1995). Resilience and reserve capacity in
later adulthood: Potentials and limits of development. In D. Chicchetti & D. J.
Cohen (Eds.), Developmental Psychopathology: Risk, disorder and adaptation (Vol.
23, pp. 801-847). New York: Wiley.
Tett, R., Wang, A., Thomas, M., Griebler, J., & Linkovich, T. (1997, April 4 1997). Testing
a model of emotional intelligence. Paper presented at the 1997 Annual Convention
of the Southeastern Psychological Association, Atlanta, GA.
Vaillant, G. E., & McCullough, L. (1987). The Washington University Sentence Completion
Test compared with other measures of adult ego development. American Journal of
Psychiatry, 144, 1189-1194.
Zeidner, M., Matthews, G., Roberts, R. D., & MacCann, C. (2003). Development of
emotional intelligence: Towards a multi-level investment model. Human
Development, 46(2-3), 69-96.

26