Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Optimism, positive affectivity, and salivary cortisol

Julian C. L. Lai1*, Phil D. Evans2, Sik Hung Ng1, Alice M. L. Chong1,Oswald T. Siu1, Cecilia L. W. Chan3, Samuel M. Y. Ho3,Rainbow T. H. Ho3, Plato Chan3 and Charles C. Chan4

1City University of Hong Kong, Hong Kong, People’s Republic of China2University of Westminster, UK3University of Hong Kong, Hong Kong, People’s Republic of China4Hong Kong Polytechnic University, Hong Kong, People’s Republic of China

Objectives. Research on stress and salivary cortisol has focused almost exclusivelyon the effects of negative psychological conditions or emotional states. Little attentionhas been drawn to the impact associated with positive psychological conditions, whichhave been shown recently to have significant influences on neuroendocrine regulation.The aim of this study is to examine the impact of optimism and positive affect on salivarycortisol with the effects of their negative counterparts controlled for.

Design. Optimism and pessimism, and positive and negative affectivity were studiedin relation to the diurnal rhythm of salivary cortisol in a group of 80 Hong KongChinese, who provided six saliva samples over the course of a day on two consecutivedays. The separate effects of optimism and positive affect on two dynamic componentsof cortisol secretion, awakening response, and diurnal decline were examined.

Methods. Optimism and pessimism were measured using the Chinese version of therevised Life Orientation Test while generalized affects and mood states were assessedby the Chinese Affect Scale. An enzyme-linked immunoabsorbent assay kit (EIA)developed for use in saliva was adopted for the biochemical analysis of cortisol. Testingof major group differences associated with positive psychological conditions wascarried out using two-way (group by saliva collection time) ANOVAs for repeatedmeasures with negative psychological conditions and mood states as covariates.

Results. Participants having higher optimism scores exhibited less cortisol secretionin the awakening period when the effect of pessimism and mood were controlled.This effect was more apparent in men than in women who had higher cortisol levels inthe awakening period. Optimism did not have similar effect on cortisol levels duringthe underlying period of diurnal decline. On the other hand, higher generalized positiveaffect was associated with lower cortisol levels during the underlying period of diurnaldecline after the effects of negative affect and mood states had been controlled.

* Correspondence should be addressed to Dr Julian C. L . Lai, Department of Applied Social Studies, City University of HongKong, 83 Tat Chee Avenue, Kowloon, Hong Kong, People’s Republic of China (e-mail: ssjulwin@cityu.edu.hk).

TheBritishPsychologicalSociety

467

British Journal of Health Psychology (2005), 10, 467–484

q 2005 The British Psychological Society

www.bpsjournals.co.uk

DOI:10.1348/135910705X26083

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Generalized positive affect did not significantly influence cortisol secretion during theawakening period.

Conclusions. These findings suggest that positive psychological resources includingoptimism and generalized positive affect had higher impact on cortisol secretion thantheir negative counterparts, and point to the need for increased attention to thepotential contribution of positive mental states to well-being.

The detrimental role that psychosocial stress can have on our physical and mental health

is increasingly being recognized. In particular, the glucocorticoid steroid, cortisol, is often

referred to as a key ‘stress hormone’. This is partially supported by scientific data which

broadly links cortisol dysfunction to both stress and ill-health (e.g. Rozlog, Kiecolt-Glaser,

Marucha, Sheridan, & Glaser, 1999; Sapse, 1997; Seeman & McEwen, 1996). The use ofsalivary cortisol has become common in stress and psychophysiological research to index

the activity of the hypothalamic-pituitary-adrenocortical (HPA) axis, which is a major

component of neuroendocrine regulation. This can be attributed to the convenience and

noninvasiveness associated with saliva collection and the high correlation of

physiologically active cortisol with salivary cortisol. Although the available evidence

points to the association between negative emotional experiences or stress and levels of

salivary cortisol, several issues remain inadequately addressed in the literature.

Firstly, it is unclear whether elevated or attenuated secretion is more indicative of thepresence of negative experiences or stress. For instance, chronic burnout has been

shown to be associated with either elevated (Melamed et al., 1999) or attenuated

(Pruessner, Hellhammer, & Kirschbaum, 1999) cortisol secretion. In addition, different

indices of distress have been found to relate differentially with cortisol ( Vedhara et al.,

2003). The pronounced diurnal cycle (e.g. the awakening response and the diurnal

decline) of cortisol secretion may explain this mixed pattern of findings. Other indices

derived from components of the diurnal cortisol cycle, such as the rise in response to

awakening and the negative slope characterizing diurnal decline, have been consideredas equally important indices (Edwards, Clow, Evans, & Hucklebridge, 2001). In another

study, Vedhara et al. (2003) have successfully demonstrated that ‘rate of change in

cortisol levels’ instead of the conventional index of absolute level is associated with

different levels of stress, anxiety and depression.

The second issue is conceptual in nature and relates to the prevalent use of indices of

negative mental states or functioning in research on stress and cortisol. The

predominance of ‘negativity’ in this line of research may stem from the disease-oriented

view of health and well-being prevalent in the Western world (Ryff, 1989). Thisperspective places disease and health on two opposite ends of a continuum and defines

the two states as presence and absence of negative conditions, respectively. Consequently,

the health implications of positive states or conditions have been downplayed.

Recently formulated theories (Diener, Suh, Lucas, & Smith, 1999; Ryff & Singer, 1998)

and growing evidence suggest that positive and negative mental states are independent

and thus have different implications for health or well-being. This has been illustrated by

recent findings concerning the relatively independent functioning of positive versus

negative affect (e.g. Clark & Watson, 1991), positive versus negative well-being (Huppert& Whittington, 2003), and optimism versus pessimism (e.g. Carver & Scheier, 2001; Lai,

1994, 1997). Furthermore, the actual health implications of negative traits or mental

states have also been called into question. It has been pointed out that although negative

affectivity (trait or state) is strongly correlated with self-reported health complaints, it is

largely unrelated to objective indicators of health status including pathology and

Julian C. L . Lai et al.468

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

mortality rates (Costa & McCrae, 1987; Watson, 1988; Watson & Pennebaker, 1989).

Therefore, the correlation between negative affect and health complaints could possibly

be attributed to the sensitization of symptom perception in people high in trait or state

negative affect. This idea has been further substantiated in a recent study examining the

relation between negative affect and symptoms reports in individuals exposed to

experimentally induced cold (Cohen, Doyle, Skoner, Gwaltney, & Newson, 1995).The increasing attention drawn to positive emotions or conditions is indicated by the

recent formulation of a theory of positive health by Ryff and Singer (1998), a shifting

research focus from ‘problematic’ to ‘positive’ ageing (e.g. Ng, Weatherall, Liu, & Loong,

1998; Rowe & Kahn, 1998) and the emergence of positive psychology (Seligman, 2000)

as a framework for scientific research. Empirical support for the health implications of

positive emotions or well-being is also mounting. For instance, findings from a recent

study have shown conclusively that absence of positive well being is more predictive of

mortality rate over a period of 7 years than the presence of psychological symptoms or

negative well-being (Whittington & Huppert, 1998). Rescoring the GHQ by computingtwo separate scores for the positively and negatively worded items, the two

investigators found that negative responses to positively worded items (rather than

endorsement of negatively worded items) predicted the probability of death over 7 years

when the effects of health, lifestyle, and demographic variables had been controlled.

With regard to the physiological correlates of positive emotions and psychosocial

resources, a recent review supported the notion that supportive social relationships are

associated with attenuated HPA reactivity (Seeman & McEwen, 1996). In line with this, a

significant negative correlation between perceived quality of social support and salivary

cortisol has been found in women with breast cancer (Turner-Cobb, Sephton,Koopman, Blake-Mortimer, & Spiegel, 2000). In another study examining susceptibility

to common cold, a lower level of salivary cortisol indexed by the area under the curve

has also been reported among individuals who tend to experience more positive

emotions in general (e.g. Cohen, Doyle, Turner, Alper, & Skoner, 2003). Taken together,

these data point clearly to the significant role of psychosocial resources that confer

resilience to stressful situations in determining cortisol secretion.

Despite the significance of the aforementioned findings, an important issue has not

been given due attention in the literature. Specifically, there has been no attempt to look

systematically at the association between cortisol and resilience-related factors such asoptimism, positive affectivity, or self-esteem, which are most closely related to the

construct of positive well-being. Independent of the effect of social support, these

dispositional factors have been shown to correlate negatively with a wide range of poor

health outcomes including psychiatric morbidity (Lai, Cheung, Lee, & Yu, 1998; Lai &

Yue, 2000) and health complaints (Scheier & Carver, 1985; Watson, 1988). Positive affect

has been found to be predictive of well-being over a period of 30 years in a longitudinal

study (e.g. Harker & Keltner, 2001). Optimism has been shown specifically to relate to

better surgical recovery (e.g. Chamberlain, Petrie, & Azariah, 1992; Scheier et al., 1989)

and long-term adjustment to breast cancer (Gall, Miguez-de-Renart, & Boonstra, 2000).More importantly, the aforementioned variables tend to moderate the impact of

stressors on health outcomes. For instance, self-esteem has been shown to attenuate the

positive relation between daily hassles and reports of physical and psychological

symptoms (DeLongis, Folkman, & Lazarus, 1988). Similarly, optimism has been reported

to lower the negative impact of unemployment on psychological health (Lai & Wong,

1998) and reduce the negative health consequences associated with daily hassles (Lai,

1995). Positive emotions are able to mitigate the negative physiological sequelae of

Positive psychological resources and salivary cortisol 469

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

negative emotions (e.g. Fedrickson & Levenson, 1998). Recent evidence also suggests

that optimism is associated with positive changes in psychological functioning in

response to adversity or trauma (e.g. reviewed in Carver, 1998). In other words, these

psychological factors provide resilience to stressful situations and are likely to act on the

same pathway through which stressors express their neuroendocrine effects.

In spite of the potential contributions of positive psychological resources to healthoutcomes, there have been only a few isolated studies examining the link between

positive and negative affect and salivary cortisol (e.g. Cohen et al., 2003; Smyth et al.,

1998; van Eck, Berkhof, Nicolson, & Sulon, 1996; van Eck & Nicolson, 1994). Although

no clear relationship has been reported, Smyth et al. (1998) have shown that a positive

mood state was associated with lower cortisol levels while negative mood state was

associated with higher levels. One study has also examined the relationship between

optimism and cortisol following a natural disaster, but no significant relation was

reported (Steffen, 1999). The important health implications of such resilience-relatedfactors warrants a systematic examination of the link between positive psychological

resources and cortisol secretion.

We designed the current study to examine the relative contribution of positive

versus negative dispositions to predict different components of the diurnal cycle of

salivary cortisol. We define positive psychological resources as those dispositional

factors, such as optimism and generalized positive affect, that potentially confer

resilience to stressful situations, whereas pessimism and generalized negative affect are

negative dispositions that may be associated with increased vulnerability. We focused onthe two major components of the diurnal cortisol cycle, namely the awakening response

(an acute and substantial rise) and underlying diurnal decline over the entire day, which

recently have been shown to be correlated with different psychological parameters

(e.g. Edwards, Hucklebridge, Clow, & Evans, 2003). Special attention will be given to the

issue of the independence between positive and negative dispositions, and whether this

independence can be demonstrated by studying the associations between salivary

cortisol and measures of these psychological factors.

We selectively studied optimism/pessimism and positive/negative affectivity becausethese two sets of constructs have been shown to be conceptually distinctive (e.g. Chang,

Maydeu-Olivares, & D’Zurrilla, 1997; Lai, 1997; Marshall, Wortman, Kusulas, Hervig, &

Vickers, 1992) and function independently in health-related contexts. The relation

between positive psychological resources and cortisol was conceptualized in

accordance with the framework developed by Segerstrom (2000) to explain the effects

of personality factors on physiological systems. The impact of positive and negative

dispositional factors on cortisol status was assumed to be tonic in nature and potentially

mediated by mood states. Therefore, we were particularly interested in the effects ofpositive psychological resources on cortisol secretion after the effects of mood and

negative dispositions were controlled. We predicted that positive resources would exert

a significant effect on cortisol secretion and that this would be independent of the

effects of their negative counterparts.

Method

Participants and procedureA total of 80 healthy, active Hong Kong Chinese took part in the study. The ages of the

participants ranged from 19 to 55 years (Mean ¼ 28:3 years, SD ¼ 8:6 years). The mean

ages for men and women were 28.6 years (SD ¼ 7:7) and 27.9 years (SD ¼ 9:5),

Julian C. L . Lai et al.470

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

respectively. Men comprised 47.7% of the participants. Participants were recruited from

three universities in Hong Kong and consisted of both students and university staff. All of

the participants took part voluntarily without any financial inducement. They were free

from any psychiatric or neuroendocrine disorder, and were not on medication that

would potentially affect cortisol concentrations during the study period. All participants

were required to attend a briefing session in which they were given a study packcontaining full standardized written instructions, questionnaires, saliva sampling tubes

(Salivettes), and a comprehensive description of procedure. They were also given an

opportunity for questioning.

Participants took part in the experiment for two consecutive days within 1 week.

To avoid the potential difference in cortisol levels between workdays and weekend days

(e.g. van Eck & Nicolson, 1994), participants were instructed to provide saliva samples

on two consecutive workdays (weekdays for students). They were instructed to collect

one saliva sample using the Salivette (Sarstedt Lt., Leicester, UK) on each day at each ofthe six collection times: immediately after awakening; 20 minutes and 40 minutes post-

awakening; at 11:00, 16:00, and 21:00 hours. To reduce burden on the subjects, cortisol

responses were only monitored three times (rather than four times) at 0, 15, 30 and 45

minutes. Although the latter sampling schedule may be a more sensitive measure for the

peak of the awakening response, we expected that our approach would also be

adequate given that a number of recent studies (e.g. Edwards et al., 2001; Pruessner

et al., 1999) observe that the increase in cortisol levels between 15 and 30 minutes post-

awakening is limited.After collecting the last saliva sample on each day, participants were also required

to indicate how they had been feeling over the day using a mood checklist comprised

10 positive and 10 negative affect adjectives. Additional psychological data were

collected via the administration of a set of questionnaires included in the study pack.

Participants had been instructed to fill them in at least 1 day before they started to

collect saliva samples. Saliva samples were stored in the participants’ home freezers

until they returned the saliva samples together with the completed questionnaires to the

experimenter within 1 week. The saliva samples were then stored at 2208C until theywere thawed for biochemical analysis. Data collection was completed within 6 months,

which was also the longest refrigerated time for some of the saliva samples.

Measures

Optimism and pessimismThese were measured by the Chinese version of the revised Life Orientation Test(C-RLOT) adapted previously to assess optimism in Hong Kong Chinese (Lai, 2002; Lai &

Yue, 2000). The C-RLOT consists of three positively worded (e.g. ‘I am always optimistic

about my future’) and three negatively worded items (e.g. ‘I hardly expect things to go

my way’). It was an adaptation of the English version of the revised Life Orientation Test

(Scheier, Carver, & Bridges, 1994), which measures generalized positive outcome

expectancies or optimism. The scale and its original version (Scheier & Carver, 1985)

have been the most commonly used instruments to measure optimism in psychological

research.The C-RLOT has been shown to be a reliable and valid index of optimism in Hong

Kong Chinese (e.g. Lai, 2002; Lai & Yue, 2000). However, there is also evidence showing

that the positively and negatively toned items may measure two different dimensions of

optimism and pessimism, respectively, (e.g. Lai, 1994, 1997; Chang & Bridewell, 1998).

Positive psychological resources and salivary cortisol 471

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Lai and Yue (2000) have pointed out that the correlation between the optimism and

pessimism index of the Life Orientation Test varies across studies, and ranges only from

weak to moderate. In addition, the one-factor model of optimism was not consistently

replicated in different samples of Chinese (Lai & Yue, 2000). Therefore, the current

study scored the positively and negatively toned items separately to form two indices of

optimism and pessimism.To complete the scale, respondents were asked to indicate on a 5-point scale

(1 ¼ strongly disagree; 3 ¼ neutral; 5 ¼ strongly agree) the extent to which they

agree or disagree with each of the six items. An optimism score was also computed by

adding rating of the positive and reversed negative items. The C-RLOT exhibited

acceptable reliability when used as a single measure for optimism (a ¼ :72). Optimism

(positive outcome expectancy) and pessimism (negative outcome expectancy) scores

were computed by adding ratings on the positive and negative items respectively.

Optimism and pessimism turned out to be weakly correlated, r ¼ 2:38 ( p , :01).The two subscales also had an acceptable degree of internal consistency, with as ¼ :68

and .66 for the positive and negative subscale, respectively.

Median split was used to form two groups of high versus low in optimism, and two

groups of high versus low in pessimism for subsequent analysis. Participants with scores

equal to, or higher than, the median of either optimism or pessimism were classified

accordingly as high in optimism or pessimism. Those with scores lower than the median

were classified as low in either optimism or pessimism.

Positive and negative affectThis was measured by the Chinese Affect Scale (CAS; Hamid & Cheng, 1996) based upon

affective terms indigenous to the Hong Kong Chinese culture. The scale consists of

10 negative affect and 10 positive affect adjectives, which are loaded separately on two

factors of negative versus positive affect. There is evidence demonstrating the reliability

and validity of the scale for measuring both generalized and state affect in Hong KongChinese. Lai, Hamid, and Chow (1996) have also shown that scores of the CAS correlate

with levels of stress in the predicted direction in Hong Kong Chinese.

To complete the scale, participants used a 6-point scale (1 ¼ not at all; 6 ¼ very

often) to describe how they had been feeling during the past month (generalized affect)

or over the day (state affect) with the 20 affect terms. A 30-day response frame was used

for both affect measures to capture more generalized affect. This response frame is seen

to be sensitive to both contextual and personality factors, and most appropriate for

assessing generalized affect (Mroczek & Kolarz, 1998). Ratings on the positive andnegative items were scored separately. A positive affect score (generalized or state) was

computed by adding ratings on the 10 positive mood items. Similarly, a negative affect

(generalized or state) was derived by adding ratings on the negative mood items.

The CAS should be a reliable measure for both affects. Cronbach as for the trait positive

and negative affect scale were .87 and .89, respectively. The correlation between these

two scales was modest, r ¼ 2:51, p , :01. As with positive expectancy scores,

two groups with high versus low trait positive affect were formed by using the median

split. Participants were also classified as either low or high in trait negative affect byusing the corresponding median split.

Participants were required to complete both the generalized and state version of the

CAS. The latter was completed at the end of each of the two consecutive days

immediately after the collection of the last saliva sample. The CAS turned out to be a

Julian C. L . Lai et al.472

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

reliable measure for mood state. Cronbach as for the positive mood scale on day 1 and

day 2 were .88 and .90, respectively. For the negative mood scale, these were .85 and

.90, respectively. Reports of mood were stable across the 2 days: positive mood, r ¼ :78,

p , :01; negative mood, r ¼ :61, p , :01. To derive an index of mood state across the 2

days, scores of positive and negative mood score on day 1 and day 2 were averaged

separately. The two indices of mood state correlated moderately, r ¼ 2:61, p , :01.Participants were divided into two groups with high and low positive/negative mood

using the corresponding median score.

Statistical analysesAnalysis of cortisol secretion focused on the awakening response and diurnal decline

respectively. The former comprised the three morning samples whereas the latter was

defined by the first, and the last three samples. The testing of major group differences

associated with optimism, pessimism and generalized affect was carried out using two-

way (group by saliva collection time) ANOVA for repeated measures (Statistica 6.0).

Positive and negative mood scores were used as covariates by adopting the means model

of ANCOVA. Degrees of freedom were adjusted using the Greenhouse-Geisser Epsilon.

Bonferroni procedure (Rosenthal & Rosnow, 1991, p. 329) was used to control potentialinflation of experimental error Type I in post hoc comparisons, in which one-tailed tests

were adopted.

Cortisol analysisSaliva samples were thawed and centrifuged at 3,000 rpm for 15 minutes at room

temperature. Clear supernatant was used for analysis. Cortisol levels were determined

by using an enzyme-linked immunoabsorbent assay kit (EIA) developed for use in Saliva

(Salimetrics, Inc., State College, PA, USA). The assay sensitivity for the kit was 0.2 nmol/l.

Intra-assay and inter-assay coefficients of variation was 3% and 10%, respectively.

Results

Preparation of cortisol dataAs with data reported in prior studies, the cortisol data of the current sample are also

highly skewed. Extreme values were winsorized in order to reduce the impact of

outliers. Logarithm transformation, which has been one of the most commonly used

methods, was adopted to normalize the distribution of the data before they were subjectto further analysis. To avoid singularities with near zero values in cortisol concentration,

a more appropriate transformation (logX þ 1) was used (Ferguson & Takane, 1989,

p. 265). This transformation procedure substantially reduced skewness in the data.

All analyses carried out on cortisol concentrations were based on the log-transformed

data. Missing data was minimal in the current study and was not considered to be

a potential threat to the validity of the data. Of the 960 samples that were possible

(80 participants £ 2 days £ 6 collection points per day), only three were missing

(,0.3%). With respect to psychological data, 10 participants failed to fill in thequestionnaires. In addition, five participants were forced out of analysis because of a

large amount of missing data in their questionnaires meaning that a total of 65

participants had sufficient data for inclusion. The proportion of two sexes in

participants whose data were not included in analysis, was not significantly different

Positive psychological resources and salivary cortisol 473

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

from that among participants who provided sufficient data, x2ð1Þ ¼ :005, p ¼ :94.

Neither did age differ between these two groups, tð78Þ ¼ 1:2, p ¼ :24. The mean

cortisol levels across the two days were not significantly different between these two

groups, Fð1; 76Þ ¼ 1:6, p . :20. There was no significant interaction effect between

group and collection time, Fð4:1; 313:6Þ , 1, p ¼ :76. Results of these analyses suggest

that the attrition was not likely to create any biases in the data.

Diurnal rhythm of cortisol secretionThe mean cortisol concentrations observed in the current study are comparable to those

reported previously in studies with western samples (Table 1; e.g. Nicolson, Storms,

Ponds, & Sulon, 1997; Nicolson & van Diest, 2000). The correlations of cortisol

concentration at each of the six sampling times across the two days were significant:

Sample 1, r ¼ :29, p , :01; Sample 2, r ¼ :51, p , :01; Sample 3, r ¼ :36, p , :01;

Sample 4, r ¼ :23, p , :05; Sample 5, r ¼ :40, p , :01; Sample 6, r ¼ :55, p , :01.

These relatively weak correlations imply moderate intra-individual stability across the 2days, but are somewhat smaller than that reported in prior studies with western samples

(e.g. Edwards et al., 2001; Pruessner et al., 1997).

On the other hand, the correlation between several composite measures of cortisol

secretion including the mean of the morning samples, the mean of the afternoon

samples, the area under the cortisol curve with reference to zero (AUC; Edwards et al.,

2001) and diurnal decline as defined by the difference between the first and the last

sample (Edwards et al., 2001) were substantial: rs ¼ :50 ( p , :01), .54 ( p , :01), .57

(p , :01), .31 (p , :01), respectively. In addition, the diurnal rhythm of cortisol

secretion was not significantly different across the two days as indicated by the results of

a two-way ANOVA with repeated measures: Fð1; 77Þ ¼ 1:8, p . :05. In view of the

stability of these indices and the pattern of secretion across the two consecutive days,

the cortisol profiles of the two days were averaged to give a more parsimonious

presentation. Subsequent analyses were based on this averaged cortisol profile.

As shown in Fig. 1, the typical diurnal pattern of cortisol secretion was replicated in

the current sample. A one-way ANOVA with repeated measures was run to examine the

Table 1. Mean cortisol levels across the two consecutive days

Samples Mean (nmol/l) Standard deviation

Day 1 – Sample 1 7.97 5.4Day 1 – Sample 2 11.17 6.5Day 1 – Sample 3 9.85 6.3Day 1 – Sample 4 4.22 3.4Day 1 – Sample 5 2.98 3.1Day 1 – Sample 6 1.20 1.2Day 2 – Sample 1 7.70 5.1Day 2 – Sample 2 10.51 5.7Day 2 – Sample 3 9.57 5.0Day 2 – Sample 4 4.59 3.7Day 2 – Sample 5 3.71 4.6Day 2 – Sample 6 2.14 3.0

Sample 1 ¼ immediately after waking, Sample 2 ¼ 20 minutes after waking, Sample 3 ¼ 40 minutesafter waking, Sample 4 ¼ 11:00 hours, Sample 5 ¼ 16:00 hours, Sample 6 ¼ 21:00 hours.

Julian C. L . Lai et al.474

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

effects of sampling time on cortisol secretion. Results showed that there was asignificant awakening response in terms of the change in concentrations across the first

three morning samples, Fð1:7; 131:5Þ ¼ 21:4, p , :001. Cortisol secretion increased

significantly 20 minutes after awakening, tð77Þ ¼ 26:7, p , :025, and remained

significantly higher after 40 minutes, tð78Þ ¼ 23:5, p , :025. Diurnal decline in cortisol

secretion was analysed by looking at the change across the first, fourth, fifth and sixth

sample. Results of a trend analysis indicated that cortisol secretion decreased linearly

from awakening to late evening, Fð1; 78Þ ¼ 369:5, p , :001. Cortisol secretion dropped

significantly from the first to the fourth sample, tð78Þ ¼ 6:6, p , :025, and from thefourth to the last sample tð79Þ ¼ 8:8, p , :025. Cortisol level in the morning (mean of

the 3 morning cortisol levels) was significantly higher than that in later part of the day

(mean of the last 3 samples), tð77Þ ¼ 22:7, p , :001. Health behaviour including

physical exercise level, smoking status, alcohol intake and quality of sleep had no

significant relationship with cortisol level at each of the six collection times.

Correlations between cortisol levels and psychological variablesThe correlations between the mean cortisol levels and expectancy and affect are

summarized in Table 2. Gender and optimism correlated significantly with cortisol

secretion at various collection times. The values of Cronbach alphas indicate that

positive and negative psychological conditions were assessed adequately by the

corresponding scales.Because prior data suggest that females have a higher awakening response compared

with males (e.g. Pruessner et al., 1997; Pruessner et al., 1999), potential gender

differences in the awakening cortisol response were examined in the current study

using a two-way ANOVA with repeated measures. Figure 2 shows that female

participants tend to have higher cortisol response in the morning than their male peers.

Cortisol concentrations changed significantly within the first 40 minutes after

awakening, Fð1:7; 101:9Þ ¼ 14:4, p , :001. A significant gender effect was apparent,

Fð1; 61Þ ¼ 5:0, p , :05. Cortisol levels in females were significantly higher than that inmales only at 40 minutes post-awakening, tð63Þ ¼ 22:8, p , :016. There was also a

trend for females having higher cortisol secretion 20 minutes post-awakening,

tð61Þ ¼ 22:1, p ¼ :02. The interaction between gender and time of sampling was not

significant, Fð1:7; 101:9Þ ¼ 1:8, p ¼ :18.

Figure 1. Salivary cortisol by sample time.

Positive psychological resources and salivary cortisol 475

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

The observed gender effect was further analysed in relation to two indices of

awakening response: the mean increase from awakening (MnInc; Wuest, Federenko,

Hellhammer, & Kirschbaum, 2000) and the area under the cortisol curve with reference

to zero (AUC; Edwards et al., 2001). A gender difference was found only in AUC,

tð60Þ ¼ 22:2, p , :025. Mean awakening time across the 2 days was not significantly

related to either of these two indices, nor to any of the three morning cortisolconcentrations.1

Gender, optimism and cortisol secretionA three-way ANOVA with repeated measures (gender by group by sampling time) was

used to examine the effect of gender and optimism on the awakening samples (1, 2, and

3) with the effect of pessimism and mean mood scores across the 2 days controlled.

Figure 2. Salivary cortisol in awakening period by gender.

Table 2. Correlations for mean cortisol levels, gender, age and psychological factors

Gender Age PA NA OP PE DPA DNA

Sample 1 .12 2 .09 2 .07 2 .07 2 .25* 2 .02 2 .01 .10Sample 2 .26* 2 .09 2 .05 2 .02 2 .23 2 .12 2 .10 .15Sample 3 .33** 2 .06 .05 2 .19 2 .30** 2 .09 .12 .05Sample 4 2 .07 2 .02 2 .13 2 .05 2 .18 .02 2 .19 .10Sample 5 2 .07 2 .05 2 .17 2 .00 2 .32** .16 2 .03 .13Sample 6 2 .08 2 .00 2 .24 2 .09 2 .23 .18 2 .10 .09Mean (SD) – 28.3 (8.6) 37.6 (6.9) 28.1 (8.0) 10.3 (2.4) 8.0 (2.5) 36.1 (7.0) 22.0 (6.4)Cronbach a – – .90 .89 .68 .66 – –N 65 65 65 65 65 65 64 64

�p , :05, ** p , :01. PA ¼ trait positive affect, NA ¼ trait negative affect, OP ¼ optimism,PE ¼ pessimism, dPA ¼ mean positive mood scores across 2 days, dNA ¼ mean negative moodscores across 2 days. dPA and dNa are composite measures and thus no Cronbach alpha can beprovided.

1 The mean awakening time across the 2 days was 08:29, which had no significant correlation with morning cortisol levelsand the two composite measures: Sample 1, r¼ 2.05; Sample 2, r¼ 2.05, Sample 3, r¼ 2.20, MnInc, r¼ 2.08;AUC, r¼ 2.10 (all ps . .05).

Julian C. L . Lai et al.476

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

A two-way ANOVA with repeated measures (group by sampling time) was applied to

study the effect of optimism on the underlying period of diurnal decline (samples 1, 4, 5,

and 6) with pessimism and mood scores as covariates.

For the awakening period, optimism had a significant main effect on cortisol

secretion, Fð1; 52Þ ¼ 10:8, p , :01. The effect of gender was not significant,

Fð1; 52Þ ¼ 3:2, p ¼ :08. There was also a significant interaction effect between genderand optimism, Fð1; 52Þ ¼ 4:2, p , :05. As illustrated in Fig. 3, those with lower optimism

scores had overall more pronounced secretion in cortisol in the morning. This significant

difference was apparent at each of the three sampling times: immediately upon

awakening, tð57Þ ¼ 2:7, p , :005; 20 minutes post-awakening, tð57Þ ¼ 2:6, p , :01; 40

minutes post-awakening, tð57Þ ¼ 3:2, p , :0025. Consistent with this, higher optimism

scores were also associated with a lower AUC and mean morning secretion, tð61Þ ¼ 2:6,

p ¼ :011, and tð61Þ ¼ 2:8, p ¼ :006, respectively. In line with the literature, there was a

highly significant awakening response, Fð1:7; 93:6Þ ¼ 14:6, p , :001, across allparticipants but there was no significant interaction between optimism and sampling

time, Fð1:7; 93:6Þ , 1, p ¼ :81. The three-way interaction was not significant either,

Fð1:7; 93:6Þ ¼ 2:9, p ¼ :07.

The nature of the interaction between gender and optimism is illustrated in Fig. 4,

which is based on the mean cortisol levels of samples associated with the awakening

response. The effect of optimism on cortisol secretion appears to be observed only in

men. Men having higher optimism scores tended to have lower mean cortisol level in

the awakening period than those who had lower optimism scores, tð25Þ ¼ 3:8,p , :0005. This difference, however, did not apply to women, tð30Þ ¼ 0:9, p . :25.

With regard to the underlying profile of diurnal decline in cortisol secretion (samples

1, 4, 5, and 6), optimism had no significant effect, Fð1; 55Þ ¼ 3:1, p . :05, which was

independent of that associated with pessimism and mood scores. There was a significant

decline in cortisol secretion, Fð3; 174Þ ¼ 105:8, p , :001, but the interaction between

optimism and diurnal decline was not significant, Fð3; 174Þ ¼ 1:9, p ¼ :13.

Generalized positive affect and cortisol secretionSimilar analyses were carried out to examine the effect of generalized positive affect on

cortisol secretion with the effect of trait negative affect and mood scores controlled. As

with the analyses for the effect of optimism, gender was also included as one of the main

Figure 3. Salivary cortisol in awakening period by optimism.

Positive psychological resources and salivary cortisol 477

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

factors in a three-way ANOVA with repeated measures. Results indicated that neither

gender nor trait positive affect exerted any significant effects on the awakening

response, Fð1; 52Þ ¼ 3:7, p ¼ :06; Fð1; 52Þ ¼ 3:1, p . :05. There was no significant

interaction effect between any two of the three factors. With respect to the period of

diurnal decline, generalized positive affect had a significant effect on cortisol levels,

Fð1; 55Þ ¼ 6:7, p , :05. Data in Fig. 5 show that participants with higher generalized

positive affect had lower cortisol levels in the underlying period of diurnal decline than

their peers having lower positive affect. Results of post hoc comparisons showed thatthe difference between the these two group was apparent only at the last collection

time point, tð58Þ ¼ 3:3, p , :001. The interaction between positive affect and diurnal

decline was not significant, Fð3; 174Þ , 1, p ¼ :76.

Pessimism, generalized negative affect and cortisol secretionFinally, analyses looking at the effect of negative dispositions with optimism or

generalized positive affect as covariates did not yield any significant findings. This suggests

that positive psychological states or conditions are more important than negative ones in

Figure 4. Effect of gender and optimism on mean salivary cortisol in awakening period.

Note. Analysis based on mean cortisol levels across samples in the awakening period.

Figure 5. Diurnal salivary cortisol profile by positive affect.

Julian C. L . Lai et al.478

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

predicting diurnal pattern of cortisol secretion in the current sample. In addition, when

mean positive and negative mood scores were used as covariates in the analyses of

covariance, neither of these two state measures affected cortisol secretion significantly.

Discussion

In summary, our findings show that positive psychological resources including

optimism or positive outcome expectancy and generalized positive affect exert

significant effects on different components of cortisol diurnal rhythm. Specifically,

participants having higher optimism scores tended to have lower cortisol secretion in

the awakening period. On the other hand, participants having higher trait positive affect

exhibited attenuated secretion in the whole diurnal period omitting the superimposed

awakening peak. The significant effects of these positive psychological resources orconditions were demonstrated with the effects of their negative counterparts and mood

states controlled. This indicates that positive psychological resources may have tonic

influences on cortisol secretion and this is not affected by daily fluctuations in mood

states. Although both optimism and generalized positive affect had a significant impact

on cortisol secretion, the effect of optimism was more apparent in the morning whilst

that of generalized positive affect was stronger in the evening. This suggests that these

two positive resources are conceptually and functional independent of each other,

which is consistent with the modest correlation between these two variables (r ¼ :24,p ¼ :05) observed in the current study. Taken together, our data suggest the

independence of positive and negative psychological conditions and that the former are

more potent in influencing cortisol secretion. To the best of our knowledge, this has not

been reported previously. The implications of our findings warrant further discussion.

As mentioned earlier, female participants had higher cortisol levels than did men in

the awakening period, which is in line with prior research findings (e.g. Pruessner et al.,

1997). In addition, gender significantly interacted with optimism in determining cortisol

secretion during the same period. Higher optimism scores were associated with lesscortisol secretion in men but not in women. The effect of higher optimism on cortisol

may be counteracted by an opposite effect of gender. Although women who are in the

luteal phase of their menstrual cycle have been found to exhibit enhanced salivary

cortisol response to laboratory stressors (e.g. Kirschbaum et al., 1995; Kirschbaum,

Wust, & Hellhammer, 1992), these women do not exhibit increased cortisol secretion in

the awakening period when compared with women in other phases of the menstrual

cycle (Krischbaum, Kudielka, Gaab, Schommer, & Hellhammer, 1999). The latter finding

has been supported by a more recent study (Kudielka & Kirschbaum, 2003). Therefore,it is not clear whether or not the accentuated morning secretion of cortisol in female

participants of our study was caused by an estradiol-mediated increase in the sensitivity

of the adrenal cortex to adrenocorticotropic hormone. Kirschbaum, Klauer, Filipp, and

Hellhammer (1995) also reported a possible interaction between gender and positive

psychosocial resources, showing that social support attenuates the cortisol response

to acute laboratory stressor in men but not women. Further research is necessary

to explore interplay between gender and other psychological factors that contribute to

individual differences in HPA activity.The relative contribution of positive versus negative affect and optimism versus

pessimism to psychological functioning have been extensively researched and remains

keenly debated, but these controversies have been taking place largely on the

psychological level and focus primarily on the pathological effects of negative

Positive psychological resources and salivary cortisol 479

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

dispositions or conditions. Our findings demonstrate that the effects of positive versus

negative dispositions can be differentiated at the level of HPA activity, and thus lend

support to the notion that positive resources and their negative counterparts are

functionally independent. This will hopefully add a physiological dimension to the

ongoing debate.

Prior research on the psychological correlates of cortisol secretion has lookedprimarily at measures of negative psychological states or dispositions, which have been

shown to have significant relationships with different measures of cortisol secretion.

Among the few studies that looked at positive affect (e.g. Cohen et al., 2003; Smyth et al.,

1998; van Eck et al., 1996; van Eck & Nicolson, 1994), no consistent effect of these

positive measures on cortisol concentrations has been reported. This, however, could

possibly be attributed to different operationalization of positive psychological conditions

or mental states and the exclusive use of male participants in two of the four studies.

However, as mentioned earlier, Whittington and Huppert (1998) have conclusively

demonstrated the superiority of positive conditions in predicting health outcomes.Despite the significance of the findings reported here, their impact may be

undermined by a number of conceptual and methodological issues. For example, the

validity of the timing of saliva sample collection has not been examined (except for

awakening time) and relied solely on the compliance of participants. This may

undesirably increase the errors of the study and result in lower intra-individual stability

in cortisol secretion (Kudielka, Broderick, & Kirschbaum, 2003). However the principal

significant findings reported here concern the mean values of cortisol across various

sample points. While such results are not immune to measurement errors due to timing

inaccuracies, they are not as problematic as findings relating to dynamic measures,notably mean increase scores in the volatile 40 minutes post-awakening, or slope

measures of decline across the day.

Other issues remain unexplored. Although positive resources are associated with an

attenuated pattern of cortisol secretion, it remains unclear whether the decreased HPA

reactivity has any health-related implications or not. Since the relationship between

optimism and positive affectivity and health benefits have been well substantiated in the

psychological literature, it would be useful to examine the potential contribution of the

HPA axis in mediating this relationship.

Lastly, our findings may not be taken to imply that the supremacy of positivepsychological resources in predicting cortisol secretion could be extended to other

populations or age groups. There is evidence that optimism and pessimism become more

independent and each contributed unique variance to physical and psychological health

in older men (Mroczek, Spiro, Aldwin, Ozer, & Bosse, 1993). It is likely that both age and

gender could possibly moderate the relative contribution of positive versus negative

dispositions or psychological conditions to well being. Moreover, Schulz, Bookwala,

Knapp, Scheier, and Williamson (1996) have recently reported that pessimism, not

optimism, predicted mortality among middle-aged cancer patients. With respect to

positive versus negative affect, Charles, Reynolds, and Gatz (2001) have shown thatcompared with younger age groups, older people exhibit an increase in negative affect

and decrease in positive affect over time in a 23-year longitudinal study. In another study

using an experience-sampling procedure, Carstensen, Pasupathi, Mayr, and Nesselroade

(2000) have shown that the correlation between positive and negative affect decreases

with age, meaning that the experience of positive and negative emotion becomes more

independent of each other as people age. Moreover, age and gender have been reported to

interact in affecting the balance between positive and negative affect (Mroczek & Kolarz,

Julian C. L . Lai et al.480

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

1998). In view of these recent findings, it would be worthwhile for future research to look

further at the relative influences of positive versus negative psychological conditions or

dispositions upon health outcomes in relation to potential moderators.

Conclusion

The most important findings that distinguish the current study and prior research

concern the relationship between positive psychological conditions and cortisol

secretion. Our findings point to the importance of optimism and generalized positive

affect in influencing different components of the diurnal cortisol cycle when the effect of

negative psychological dispositions and mood states were statistically controlled. To our

knowledge, few previous studies (if any) have reported such a pattern of relationships.

Our findings lend support to the recent proposal stressing the need to ‘include measures

of positive well-being in studies of health outcomes and quality of life assessment’(Huppert & Whittington, 2003, p. 107). The findings in the current study may draw

increased attention to the potential impact of positive psychological dispositions or

conditions on cortisol secretion and thus initiate a shift of research focus to the

physiological substrates of positive states of minds, which according to Ryff and Singer

(1998) is indispensable for a better understanding of human functioning and well-being.

Acknowledgements

City University of Hong Kong Research Grant no. 9030938 provided support for this research.

Thanks are due to Dr Wing Sze Wong and Dr Wendy Wan for their assistance in data collection.

References

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,

644–655.

Carver, C. S. (1998). Resilience and thriving; Issues, models, and linkages. Journal of Social Issues,

54, 245–266.

Carver, C. S., & Scheier, M. F. (2001). Optimism, pessimism and self-regulation. In E. C. Chang

(Ed.), Optimism and pessimism: Implications for theory, research, and practice (pp. 31–51).

Washington DC: American Psychological Association.

Chamberlain, K., Petrie, K., & Azariah, R. (1992). The role of optimism and sense of coherence on

predicting recovery following surgery. Psychology and Health, 7, 301–310.

Chang, E. C., & Bridewell, W. B. (1998). Irrational beliefs, optimism, pessimism and psychological

distress: A preliminary examination of differential effects in a college population. Journal of

Clinical Psychology, 54, 137–142.

Chang, E. C., Maydeu-Olivares, A., & D’Zurilla, T. J. (1997). Optimism and pessimism as partially

independent constructs: Relations to positive and negative affectivity and psychological well

being. Personality and Individual Differences, 23, 433–440.

Charles, S. T., Reynolds, C. A., & Gatz, M. (2001). Age related differences and change in positive

and negative affect over 23 years. Journal of Personality and Social Psychology, 80, 136–151.

Clark, L. A., & Watson, D. (1991). General affective dispositions in physical and psychological

health. In C. R. Snyder & D. R. Forsyth (Eds.), Handbook of social and clinical psychology:

The health perspective (pp. 221–245). New York: Pergamon.

Positive psychological resources and salivary cortisol 481

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Cohen, S., Doyle, W. J., Skoner, D. P., Gwaltney, J. M., & Newson, J. T. (1995). State and trait

negative affect as predictors of objective and subjective symptoms of respiratory viral

infections. Journal of Personality and Social Psychology, 68, 159–169.

Cohen, S., Doyle, W. J., Turner, R. B., Alper, C. M., & Skoner, D. P. (2003). Emotional style and

susceptibility to the common cold. Psychosomatic Medicine, 65, 652–657.

Costa, P. T., Jr, & McCrae, R. R. (1987). Neuroticism, somatic complaints, and disease: Is the bark

worse than the bite? Journal of Personality, 55, 299–316.

DeLongis, A., Folkman, S., & Lazarus, R. S. (1988). The impact of daily stress on health and mood.

Journal of Personality and Social Psychology, 54, 486–495.

Diener, E., Suh, E., Lucas, R. E., & Smith, H. L. (1999). Subjective well being: Three decades of

progress. Psychological Bulletin, 125, 276–302.

Edwards, S., Clow, A., Evans, P., & Hucklebridge, F. (2001). Exploration of the awakening cortisol

response in relation to diurnal cortisol secretory activity. Life Sciences, 68, 2093–2103.

Edwards, S., Hucklebridge, F., Clow, A., & Evans, P. (2003). Components of the diurnal cortisol

cycle in relation to upper respiratory symptoms. Psychosomatic Medicine, 65, 320–327.

Fedrickson, B. L., & Levenson, R. W. (1998). Positive emotions speed recovery from the

cardiovascular sequelae of negative emotions. Cognition and Emotion, 12, 191–220.

Ferguson, G. A., & Takane, Y. (1989). Statistical analysis in psychology and education (6th ed.).

New York: McGraw-Hill.

Gall, T. L., Miguez-de-Renart, R. M., & Boonstra, B. (2000). Religious resources in long-term

adjustment to breast cancer. Journal of Psychosocial Oncology, 18, 21–37.

Hamid, P. N., & Cheng, S. T. (1996). The development and validation of an index of emotional

disposition and mood state: The Chinese affect scale. Educational and Psychological

Measurement, 56, 995–1014.

Harker, L. A., & Keltner, D. (2001). Expression of positive emotion in women’s college yearbook

pictures and their relationship to personality and life outcomes assessed across adulthood.

Journal of Personality and Social Psychology, 80, 112–124.

Huppert, F. A., & Whittington, J. E. (2003). Evidence for the independence of positive and negative

well-being: Implications for quality of life assessment. British Journal of Health Psychology, 8,

107–122.

Kirschbaum, C., Klauer, T., Filipp, S., & Hellhammer, D. H. (1995). Sex-specific effects of social

support on cortisol and subjective responses to acute psychological stress. Psychosomatic

Medicine, 57, 23–31.

Krischbaum, C., Kudielka, B. M., Gaab, J., Schommer, N. C., & Hellhammer, D. H. (1999). Impact of

gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamic-

pituitary-adrenal axis. Psychosomatic Medicine, 61, 154–162.

Kirschbaum, C., Pruessner, J., Stone, A. A., Federenko, I., Gaab, J., Lintz, D., Schommer, N., &

Hellhammer, D. H. (1995). Persistent high cortisol responses to repeated psychological stress

in a subpopulation of healthy men. Psychosomatic Medicine, 57, 468–474.

Kirschbaum, C., Wust, S., & Hellhammer, D. H. (1992). Consistent sex difference in cortisol

responses to psychological stress. Psychosomatic Medicine, 54, 648–657.

Kudielka, B. M., Broderick, J. E., & Kirschbaum, C. (2003). Compliance with saliva. Sampling

protocols: Electronic monitoring reveals invalid cortisol daytime profiles in noncompliant

subjects. Psychosomatic Medicine, 65, 313–319.

Kudielka, B. M., & Kirschbaum, C. (2003). Awakening cortisol responses are influenced by health

status and awakening time but not by menstrual cycle phases. Psychoneuroendocrinology,

28, 35–47.

Lai, J. C. L. (1994). Differential predictive power of the positively versus of the negatively worded

items of the Life Orientation Test. Psychological Reports, 75, 1507–1515.

Lai, J. C. L. (1995). The moderating effect of optimism on the relation between hassles and somatic

complaints. Psychological Reports, 76, 883–894.

Lai, J. C. L. (1997). Relative predictive power of the optimism versus the pessimism index of a

Chinese version of the Life Orientation Test. Psychological Record, 47, 399–410.

Julian C. L . Lai et al.482

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Lai, J. C. L (2002). Predictors of psychological well being and reemployment success in

unemployed Hong Kong Chinese. Manuscript submitted for publication.

Lai, J. C. L., Cheung, H., Lee, W. M., & Helen, Y. (1998). The utility of the revised Life Orientation

Test to measure optimism among Hong Kong Chinese. International Journal of Psychology,

33, 45–56.

Lai, J. C. L., Hamid, P. N., & Chow, P. (1996). Gender differences in hassles and symptom reporting

among Hong Kong adolescent. Journal of Social Behavior and Personality, 11, 149–164.

Lai, J. C. L., & Wong, W. S. (1998). Optimism and coping with unemployment among Hong Kong

Chinese women. Journal of Research in Personality, 32, 454–479.

Lai, J. C. L., & Yue, X. D. (2000). Measuring optimism in Hong Kong and mainland Chinese with the

revised Life Orientation Test. Personality and Individual Differences, 28, 781–796.

Marshall, G. N., Wortman, C. B., Kusulas, J. W., Hervig, L. K., & Vickers, R. R. (1992). Distinguishing

optimism from pessimism: Relations to fundamental dimensions of mood and personality.

Journal of Personality and Social Psychology, 62, 1067–1074.

Melamed, S., Ugarten, U., Shirom, A., Kahana, L., Lerman, Y., & Froom, P. (1999). Chronic burnout,

somatic arousal and elevated salivary cortisol levels. Journal of Psychosomatic Research, 46,

591–598.

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.

Mroczek, D. K., Spiro, A., Aldwin, C. M., Ozer, D. J., & Bosse, R. (1993). Construct validation of

optimism and pessimism in older men: Findings from the Normative Aging Study. Health

Psychology, 12, 406–409.

Ng, S. H., Weatherall, A., Liu, J. H. & Loong, C. S. F. (Eds.), (1998). Ages ahead: Promoting

intergenerational relationships. Wellington: Victoria University Press.

Nicolson, N., Storms, C., Ponds, R., & Sulon, J. (1997). Salivary cortisol levels and stress reactivity

in human aging. Journal of Gerontology: Medical Sciences, 54A, M68–M75.

Nicolson, N., & van Diest, R. (2000). Salivary cortisol patterns in vital exhaustion. Journal of

Psychosomatic Research, 49, 335–342.

Pruessner, J. C., Hellhammer, D. H., & Kirschbaum, C. (1999). Burnout, perceived stress and

cortisol response to awakening. Psychosomatic Medicine, 61, 197–204.

Pruessner, J. C., Wolf, O. T., Hellhammer, D. H., Buske-Kirschbaum, A., von Auer, K., Jobst, S.,

Kaspers, F., & Kirschbaum, C. (1997). Free cortisol levels after awakening: A reliable biological

marker for the assessment of adrenocortical activity. Life Sciences, 61, 2539–2549.

Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research (2nd ed.). New York:

McGraw Hill.

Rowe, H. W., & Kahn, R. L. (1998). Successful aging. New York: Pantheon.

Rozlog, L. A., Kiecolt-Glaser, J. K., Marucha, P. T., Sheridan, J. F., & Glaser, R. (1999). Stress and

immunity: Implication for viral disease and wound healing. Periodontology, 70, 786–792.

Ryff, C. D. (1989). Beyond Ponce de Leon and life satisfaction: New directions in quest for

successful ageing. International Journal of Behavioral Development, 12, 35–55.

Ryff, C. D., & Singer, B. (1998). The contours of positive human health. Psychological Inquiry, 9,

1–28.

Sapse, A. T. (1997). High cortisol diseases and anti-cortisol therapy. Psychoneuroendocrinology,

22, S3–S10.

Scheier, M. F., & Carver, C. S. (1985). Optimism, coping and health: Assessment and implications of

generalized outcome expectancies. Health Psychology, 4, 219–247.

Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing optimism from neuroticism

(and trait anxiety, self-mastery and self-esteem): A reevaluation of the Life Orientation Test.

Journal of Personality and Social Psychology, 67, 1063–1078.

Scheier, M. F., Matthews, K. A., Owens, J. F., Magovern, G. J., Sr, Lefebvre, R. C., Abbot, R. A., &

Carver, C. S. (1989). Dispositional optimism and recovery from Coronary artery bypass

Positive psychological resources and salivary cortisol 483

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

surgery: The beneficial effects on physical and psychological well-being. Journal of

Personality and Social Psychology, 57, 1024–1040.

Schulz, R., Bookwala, J., Knapp, J. E., Scheier, M., & Williamson, G. M. (1996). Pessimism, age, and

cancer mortality. Psychology and Aging, 11, 304–309.

Seeman, T. E., & McEwen, B. S. (1996). Impact of social environment characteristics on

neuroendocrine regulation. Psychosomatic Medicine, 58, 459–471.

Segerstrom, S. C. (2000). Personality and the immune system: Models, methods and mechanisms.

Annals of Behavioral Medicine, 22, 180–190.

Seligman, M. E. (2000). Positive psychology: An introduction. American Psychologist, 55, 5–14.

Smyth, J., Ockenfels, M. C., Porter, L., Kirschbaum, C., Hellhammer, D. H., & Stone, A. (1998).

Stressors and mood measured on a momentary basis are associated with salivary cortisol

secretion. Psychoneuroendocrinology, 23, 353–370.

Steffen, P. R. (1999). Optimism and psychological and physiological well-being following a natural

disaster. Dissertation Abstracts International, 60(6-B), 2963.

Turner-Cobb, J. M., Sephton, S. E., Koopman, C., Blake-Mortimer, J., & Spiegel, D. (2000). Social

support and salivary cortisol in women with metastatic breast cancer. Psychosomatic

Medicine, 62, 337–345.

van Eck, M., Berkhof, H., Nicolson, N., & Sulon, J. (1996). The effects of perceived stress, traits,

mood states, and stressful daily events on salivary cortisol. Psychosomatic Medicine, 58,

447–458.

van Eck, M., & Nicolson, N. (1994). Perceived stress and salivary cortisol in daily life. Annals of

Behavioral Medicine, 16, 221–227.

Vedhara, K., Miles, J., Bennett, P., Plummer, S., Tallon, D., Brooks, E., Gale, L., Munnoch, K.,

Schreiber-Kounine, C., Fowler, C., Lightman, S., Sammon, A., Rayter, Z., & Farndon, J. (2003).

An investigation into the relationship between salivary cortisol, stress, anxiety and depression.

Biological Psychology, 62, 89–96.

Watson, D. (1988). Intraindividual and interindividual analyses of positive and negative affect:

Their relation to health complaints, perceived stress, and daily activities. Journal of

Personality and Social Psychology, 54, 1020–1030.

Watson, D., & Pennebaker, J. W. (1989). Health complaints, stress, and diseases: Exploring the

central role of negative affectivity. Psychological Review, 96, 234–254.

Whittington, J. E., & Huppert, F. A. (1998). Creating invariant subscales of the GHQ-30. Social

Science and Medicine, 46, 1429–1440.

Wuest, S., Federenko, I., Hellhammer, D. H., & Kirschbaum, C. (2000). Genetic factors, perceived

chronic stress, and the free cortisol response to awakening. Psychoneuroendocrinology, 25,

707–720.

Received 23 June 2003; revised version received 9 July 2004

Julian C. L . Lai et al.484