| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Georgia Institute of Technology (C.A.H.), Atlanta, Georgia; and the University of Fribourg (P.L.K.), Fribourg, Switzerland.
Address correspondence and reprint requests to Christiane A. Hoppmann, PhD, Georgia Institute of Technology, School of Psychology, 654 Cherry Street, Atlanta, GA 30332. E-mail: ch295{at}mail.gatech.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
Methods: A total of 53 dual-earner couples with preschool children participated in a 1-week interval-sampling study. At the beginning, participants reported their personal work and family goals. During the time-sampling phase, both partners reported the goal relevance of their daily activities, current mood, and provided saliva samples for cortisol estimation every 3 hours.
Results: Hierarchical linear models show that the performance of goal-furthering activities is associated with more positive mood and decreased secretion of cortisol. The relationship between the goal relevance of daily activities and cortisol was partially mediated by affect quality.
Conclusions: These findings speak to a person-centered approach in research on stress by showing that knowledge of individual goals is important for an understanding of affective and neuroendocrine stress reactions in employed parents with preschool children.
Key Words: psychological stress • cortisol • affect • personal goals • work–family balance
Abbreviations: AUC = area under the curve; HPA system = hypothalamus–pituitary–adrenal system.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
The cognitive theory of stress by Lazarus (5,6) emphasizes the role of cognitive appraisals on individual stress responses. In this perspective, personal goals define what is at stake for a specific individual in a given situation. Personal goals are cognitive representations of the self that are future-oriented and express what an individual wants to achieve during a certain phase in life. As such, personal goals guide behavior over time and serve as an individual reference frame against which success in important life domains is evaluated (7,8). Research on young parents shows that in middle adulthood, work and family goals prevail and that the combination of work and family is very challenging (9–11). When examining individual differences in affective and neuroendocrine stress responses in the daily lives of middle-aged adults who juggle the demands posed by employment and young parenthood, it therefore seems important to attend to the personal goals people set in these domains. This study investigates the extent to which appraisals of the relevance of daily activities with regard to self-set work and family goals are related to affective and neuroendocrine stress responses in employed parents with preschool children.
Two lines of research suggest that goal relevance appraisals are a promising arena for research on the individual differences in daily stress responses. First, theories of self-regulation (12,13) suggest that action- and emotion-regulation systems are highly intertwined. They propose that individual perceptions of the goal relevance of activities feed back to the emotion system and that experiences of goal hindrance or goal blockage give rise to affective stress reactions. In line with these models, daily events interfering with personal goals have been shown to be associated with negative affect qualities (14). Daily goal attainments in contrast have been found to be associated with positive affect qualities (15,16). Within the framework of the present study, both activities hindering personal goals as well as activities furthering them are investigated simultaneously in the daily lives of employed parents using interval-sampling methods. It is expected that an involvement in activities hindering personal goals and a failure to work on these goals is related to high levels of negative affect, whereas performance of activities furthering them is associated with high levels of positive affect.
Second, biopsychosocial models of stress (17,18) propose a close connection between the performance of activities and physiological reactivity. Therein, it is assumed that the secretion of stress hormones such as cortisol is highly task-specific (4,19). For an activation of the cortisol system to occur, the performance situation must be appraised as both goal-relevant and evaluative. Cortisol secretion is assumed to be highest if situations provide a threat to or impede personal goals (4,6,17,18,20). It is therefore assumed that the performance of activities that do not serve work and family goals, or even hinder them, is associated with increased cortisol secretion in employed parents. Activities furthering work and family goals, in contrast, are expected to be related to decreased cortisol secretion. The concept of allostatic load (21) differentiates between acute stress responses that help the individual adapt to current demands at a specific point in time and chronic changes in the cortisol regulation that indicate an imbalance in the activity–recovery cycle, resulting in a heightened risk for negative tertiary effects such as hypertension (22,23). To capture acute stress responses, we examined intraindividual associations between in situ reports of goal-relevant activities and in situ changes in mood and free cortisol in saliva. To capture chronic stress responses, we also considered total cortisol secretion as a between-person variable and examined whether it contributes to the prediction of in situ changes in cortisol over and above the influence of goal pursuit and mood. To provide a meaningful interpretation of salivary cortisol, a number of factors need to be considered. Specifically, our models will account for the influence of wakeup time, sleep duration, physical activity, smoking, hormonal contraceptive use, and body mass index.
Two different mechanisms linking performance situations with physiological reactivity are proposed in the literature. According to Blascovitch and Mendes (17), cognitive appraisals and affective processes operate in parallel and provide relatively independent contributions to physiological reactivity. Frankenhaeuser (18,24) assumes, in contrast, that cognitive threat appraisals evoke negative emotions, which in turn trigger physiological stress responses. To subject the two proposed mechanisms to empiric testing, the present study further examines whether affect quality mediates the proposed relationship between goal relevance appraisals and cortisol secretion.
To examine the role of personal goals for affective and neuroendocrine stress responses in employed parents, we used a 1-week interval-sampling design, including repeated appraisals of the goal relevance of daily activities, mood ratings, and saliva sampling with an interval length of approximately 3 hours. This approach allows for an in-depth investigation of the proposed associations between a) goal-relevant activities and affect quality and b) goal-relevant activities and cortisol secretion. Furthermore, we can test whether c) affect quality mediates the relationship between goal relevance appraisals and cortisol secretion. Multilevel modeling techniques are used to examine the influence of situation-specific and person-specific characteristics on in situ changes in cortisol and mood states.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
Participants had an average age of 37 years (standard deviation [SD] = 4.9) and one to four children (mean = 1.7). The majority of participants were married (84%); the remainder was cohabiting. Ninety percent held a university degree; 10% had received other vocational training after 13 years of schooling. For both men and women, average weekly work hours ranged from 20 to 40 hours according to their work contracts. Fifty-eight percent of the men and 32% of the women earned more than 2000 Euro (approximately 2,550 U.S. dollars) of gross monthly income.
Instead of a monetary reimbursement, employed parents were offered feedback about their time management. Additionally, all participating couples took part in a lottery for a wellness weekend.
Procedure
Data collection took place between January and November 2003 and was in accordance with the ethical guidelines of the Social/Behavioral and Biomedical Sciences. Participants responding to advertisements were informed about the study design by telephone and screened for demographic variables as well as for variables that could interfere with the interpretation of cortisol results. Then they were asked to complete an Internet questionnaire, including sociodemographic characteristics and personal goals.
On an individually scheduled evening, a research assistant went to the family home and introduced the partners to the use of the study device: a Psion series 3a pocket computer and the saliva sampling device (Salivette; Sarstedt, Rommelsdorf, Germany). The research assistant went through the questionnaire in great detail and answered any questions before participants completed a practice questionnaire. Both partners also received a paper summary containing all instructions and a cell phone number for further questions during the study. Finally, the research assistant set the alarms for the interval-sampling phase, which started on the following morning. The interval-sampling phase covered 6 consecutive days (always including 4 weekdays and both weekend days). The first questionnaire was completed right after waking up. During the day (between 9:30 am and 9:30 pm), both partners simultaneously received five different beeps separated by approximately 3 hours. At each of these measurement points, study participants provided ratings of current affect, reported their activities since the previous beep (on the first measurement point, these were activities performed the night before after the last signal), and rated the goal relevance of those activities. The questionnaires took 5 to 7 minutes to complete. Participants provided a saliva sample for estimation of cortisol at the same time as completing each questionnaire and reported the respective sample number on their pocket computer. At the end of the interval-sampling phase, a research assistant collected the study materials from the family homes and asked both partners to complete a reactivity questionnaire. This questionnaire addressed questions regarding the acceptance of the interval-sampling methodology, representativeness, and accuracy of answers.
Eighty-five percent of the participants said that the week under study represented a typical week of their daily lives. Of the 3816 daily questionnaires issued (106 participants x 6 days x 6 measurement occasions), only 3.5% were missing. Two thirds of the questionnaires were answered within 15 minutes after the beep. The remaining third were delayed by up to 2 hours. Within the context of this study, delayed answers were treated as indicators of low commitment but remained in the data set because no predictions were made with respect to specific time points and the recency of the goal relevance appraisals was always given. Of the saliva samples taken parallel to questionnaire completion, only 5.5% were missing, indicating a high level of compliance.
Measures
Personal Goals
A two-step procedure was used to assess personal goals. First, participants were briefly introduced to the concept of personal goals and asked to list all the goals they actively planned to pursue within the upcoming weeks. As a result of a focus on work and family goals, we secondly asked about the most important goals with respect to four content categories, namely 1) one partnership-related goal, 2) one child-related goal, 3) one goal that was related to professional development, and 4) one goal that referred to current work projects and tasks. These content categories were derived from qualitative analyses of the open goals that 20 randomly chosen parents (10 women and 10 men) with small children had named in a study by Salmela-Aro and colleagues (11). These content categories are in accordance with theoretical assumptions about the developmental tasks of this age group (25).
Goal Relevance of Daily Activities
We asked participants to report chronologically, in 15-minute intervals, the activities in which they had been engaged since the previous beep. Next, separately for each of the four work and family goals, we asked which of the listed activities had been goal-relevant. When participants rated an activity as goal-relevant, they were further asked in which respect the goal-relevant activity affected goal progress using a scale ranging from –3 for "very much hindered" to +3 for "very much furthered" that goal (26). These goal relevance ratings were then recoded into two separate variables: goal furtherance (values equal to or above "1") and goal hindrance (values equal to or below "–1"). Summing both variables across all activities for a particular interval indicates the engagement in goal-furthering (mean = 2.54; SD = 2.09) and goal-hindering activities (mean = 0.47; SD = 1.44).
Affect Quality
We used six items from the Multidimensional Affect Scale (27) to assess participants' affect quality. At each measurement point, participants indicated their current positive affective state (good, alert, relaxed) and negative affective state (bad, tired, fidgety) on a 5-point rating scale ranging from not at all (1) to very much (5). Scores on the positive and negative affect scales were unit-weighted composites of the respective items (positive affect: mean = 3.47; SD = 0.40; negative affect: mean = 1.89; SD = 0.49). Alphas, calculated based on intraperson averages of each of the three items per scale, were satisfactory (positive affect 
= 0.84; negative affect = 0.84).
Salivary Cortisol
For estimation of cortisol, participants provided saliva samples, using the Salivette (Sarstedt, Rommelsdorf, Germany), on every measurement occasion during the interval-sampling period. Saliva samples were kept in participants' home freezers during the sampling period and then stored at the Technical University of Berlin at –20°C until analysis. The free cortisol in saliva was determined using a time-resolved immunoassay with fluorometric end point determination with an interassay variation of less than 9% (28). Samples from the same couple were always analyzed together to reduce the effects of interassay variation. Cortisol is measured in nanomole per liter.
Two measures of cortisol were of primary interest within this study. The first measure addresses in situ changes in cortisol. Because cortisol undergoes a diurnal cycle, we included a time variable that models the hour of the day when the saliva sample was taken. Hence, this cortisol measure reflects situation-specific deviations from a person's mean cortisol that are adjusted for time of day. The second measure reflects the overall secretion of cortisol during an observation period and is called "area under the curve" (AUC (29)). Following Pruessner and colleagues (29), the AUC was computed by the trapezoid formula that takes into account changes in cortisol between the different measurement points and the level at which the changes over time occur. The AUC were computed for each day in study, then averaged over the 6 days, and treated as a between-person variable.
Other Measures
Smoking habits, hormonal contraceptive use, height, and weight were assessed in the Internet questionnaire before the interval-sampling phase. The measure of physical exercise is based on the time use reports. Sleep duration and wakeup time were elicited on a daily basis in the morning questionnaire. Eighty-two percent of participants were nonsmokers, 13% indicated that they smoke occasionally, and 5% said that they smoked regularly. Twenty-six percent of the women used hormonal contraceptives. Body mass indices ranged from 18 to 33 kg/m2 with a mean of 24 kg/m2. Time spent in physical activities during the interval-sampling period ranged from 0 to 9 hours with a mean of 1 hour. Average sleep duration during time in study was 7 hours and 5 minutes (5 hours–8 hours 40 minutes). The average wakeup time was 7:06 am (5:29–9:28 am).
Statistical Analysis: Multilevel Modeling
Multilevel analysis is a methodology for the analysis of data with a hierarchically nested structure (30). The present study assumes a three-level structure in the data. The first level concerns repeated assessments in daily life, which are nested within individuals (second level) who in turn are nested within couples (third level). At each hierarchical level, explanatory variables can be added. Multilevel analysis allows for missing observations and unevenly spaced time intervals.
First, mood was modeled as a function of a characteristic of daily activities, namely whether they had furthered or hindered goal pursuit, leading to the following level 1 model:
Positive [negative] affect sijk = ß0jk + ß1jk (sum of goal-furthering activities) + ß2jk (sum of goal-hindering activities) + rijk
To test if average mood differed between women and men, gender was added as an individual-level variable leading to the following level 2 model:
ß0jk = 
00k + 01k (gender) + u0jk
ß1jk = 10k ß2jk = 20k
As a result of the fact that this sample consists of couples, a third level of analysis was added. Because no specific predictions were made at the couple level, analyses were based on an empty level 3 model:
00k = 
000 + v00k
01k = 010 02k = 020
Level 2 and level 3 slopes are treated as fixed for reasons of parsimony in all models (31). Because analyses are based on repeated observations with a short spacing in the intervals, we always checked for the effect of autocorrelative error structures (32).
Second, salivary cortisol was modeled as a function of a characteristic of daily activities, namely whether they had furthered or hindered goal pursuit, and the time of day when the saliva sample was taken. This yields the following level 1 model:
Salivary cortisol sijk = ß0jk + ß1jk (sum of goal-furthering activities) + ß2jk (sum of goal-hindering activities) + ß3jk (time of day in h) + rijk
To control for the influence of relevant variables and to test if salivary cortisol was influenced by gender and/or total cortisol secretion, the following variables were added at level 2:
ß0jk = 00k + 01k (gender) + 02k (smoking) + 03k (contraceptive use) + 04k (body mass) + 05k (physical activity) + 06k (sleep duration) + 07k (wakeup time) + 07k (AUC) + u0jk
ß1jk = 10k ß2jk = 20k ß3jk = 30k
Again, no specific predictions were made at the couple level. Hence, analyses were based on an empty level 3 model.
According to Baron and Kenny (33), a mediation requires that a) the independent variable (goal relevance appraisals) is significantly associated with the dependent variable (cortisol secretion), b) the independent variable is significantly associated with the mediator (mood), and c) the mediator is significantly associated with the dependent variable reducing or eliminating the effect of the independent variable on the dependent variable when entered simultaneously. The significance of a mediational effect can be estimated using the Sobel test (34).
Third, salivary cortisol was modeled as a function of mood and of a characteristic of daily activities, namely whether they had furthered or hindered goal pursuit, plus the previously described control variables. Hence, the level 1 model reads as follows:
Salivary cortisol sijk = ß0jk + ß1jk (positive affect) + ß2jk (negative affect) + ß3jk (sum of goal-furthering activities) + ß4jk (sum of goal-hindering activities) + ß5jk (time of day in h) + rijk
The person-level variables were included at the second level leading to the following level 2 model:
ß0jk = 00k + 01k (gender) + 02k (smoking) + 03k (contraceptive use) + 04k (body mass) + 05k (physical activity) + 06k (sleep duration) + 07k (wakeup time) + 07k (AUC) + u0jk
ß1jk = 10k ß2jk = 20k ß3jk = 30k ß4jk = 40k ß5jk = 50k
Although no specific couple-level predictions were made, an empty third level was added to account for the fact that participants are nested in couples.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
Table 1 displays the means and standard deviations for the central variables in the male and female partners of this study as well as their intercorrelations. Overall, participants reported higher levels of positive as compared with negative affect and more goal-furthering than goal-hindering activities. Women and men did not differ on any of the central study variables with the exception of body mass, which was significantly higher in men than women, sleep duration, which was significantly higher in women than men, and the use of hormonal contraceptives.
|
Content analyses of the personal goals revealed that work and family goals dominated the personal goals in the present sample. Among the employed mothers, 85% named at least one family goal, 76% named at least one work goal, and 51% named hobby goals. Among the employed fathers, 81% generated at least one work goal, 79% reported at least one family goal, and 73% named hobby goals. These results justify the current study's focus on work and family goals in the interval-sampling phase.
Goal Relevance of Daily Activities and Mood States
The first hypothesis stated that goal relevance appraisals of daily activities are associated with mood. Fully unconditional models showed that for positive affect 74% of the variance originated at the occasion level, 15% at the individual level, and 11% at the couple level. For negative affect, 59% of the variance originated at the occasion level, 25% at the individual level, and 16% at the couple level.
In line with our assumptions, the sum of goal-furthering activities had a reliable positive effect on both positive and negative affect reports in the present sample (see Table 2). That means that when employed parents engaged in activities furthering their work and family goals, they reported more positive affect as compared with intervals in which goal-hindering or goal-irrelevant activities were performed. The sum of goal-hindering activities did not add to the prediction of either positive or negative affect over and above the sum of goal-furthering activities (see Table 2). Model 1 explained 3% of the variance in positive affect and model 2 explained 2% of the variance in negative affect. The reduction in deviance, which is a measure of model fit, was significant in both models (model 1 deviance = 74.57, df = 3, p < .01; model 2 deviance = 60.32, df = 3, p < .01). Replications of these results allowing for autocorrelated error structures did not change the results.
|
Goal Relevance of Daily Activities and Cortisol Secretion
According to the second hypothesis, goal-furthering and goal-hindering activities were assumed to affect cortisol secretion. The fully unconditional model revealed 100% of the variance in in situ changes in cortisol to originate at the occasion level.
In examining the conditional models, first, we estimated models including only the control variables. Because of well-documented gender differences in affective and hormonal stress reactivity, we estimated the influence of participants' gender in every model (18). As can be seen in Table 3, no differences emerged between women and men. Also, smoking, hormonal contraceptive use, physical activity, and sleep duration did not significantly predict in situ changes in cortisol. As expected, the time of the day when the saliva sample was taken had an effect on in situ changes in cortisol and so did wakeup time and body mass index.
|
In a next step, the influence of the goal-relevance appraisals on in situ changes in cortisol was estimated. In support of our hypotheses, both goal-furthering and goal-hindering activities were reliable predictors of cortisol secretion. The average increase in cortisol levels for intervals with goal-hindering activities was 0.32 nmol/L per hour (4%). The average decrease in cortisol levels for intervals with goal-furthering activities was 0.36 nmol/L per hour (4%). In other words, at moments in which participants engaged in activities hindering their work and family goals, they displayed higher levels of cortisol than when they performed either no goal-hindering activities or goal-furthering ones. The relationships remained stable when only considering the last hour before the beep and when contextual variables (presence of others, being at work) were included. The AUC was associated with higher mean levels of cortisol. Replications of these results allowing for autocorrelated error structures did not change the results. This model explained 1% of the variance in cortisol (deviance = 12.61, df = 2, p < .01).
Testing the mediational effect of mood states on the relationship between a characteristic of daily activities, namely whether they had furthered or hindered goal pursuit, and cortisol secretion, the third set of hypotheses was based on the assumption that mood mediates the relationship between goal relevance appraisals and cortisol secretion. Because earlier analyses had already shown a significant relationship between goal relevance appraisals and mood and between goal relevance appraisals and cortisol secretion, the next step was to investigate whether mood is significantly related to cortisol secretion and whether this effect reduces the association between goal relevance appraisals and cortisol secretion when considered simultaneously.
As can be seen in Table 4, positive affect mediated the relationship between goal-furthering activities and changes in cortisol secretion at the occasion level (positive affect: reliability: z = –4.80; p < .01; negative affect: reliability: z = –0.73; not significant). Goal-hindering activities, positive affect, and negative affect made independent contributions to the prediction of cortisol over and above this effect. This model explained 4% of the variance in daily cortisol secretion (deviance = 137.45, df = 4, p < .01).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
Despite the moderate scope of individuals' reported goals, the intensity of strain on individuals is considerable because the investigated goals are central ones in midlife. This is insofar important as the thwarting of personal goals seems to be one mechanism that has the potential to trigger long-term negative effects such as the development of mental and physical illness. Convergent evidence is reported by Mueller et al. (36) who discuss the role of goal achievement in the explanation of stress-related changes in well-being. The immediate irritation resulting from the experience of a discrepancy between current status and goals is seen as a precursor of long-term changes in mental and physical health.
With regard to the underlying processes, we found the effect of goal-furthering activities on cortisol secretion to be mediated by positive affect. This mediation yields support for the mechanisms posited by Frankenhaeuser (18). The processes triggered by goal-furthering events or activities can hence be conceived of as cascading from cognition through affect to hormone secretion. Goal-hindering activities, in contrast, were reliably associated with cortisol secretion but not with mood. The absence of a reliable relationship between goal hindrances and mood has to be interpreted with caution because, overall, participants reported goal hindrances with lower frequency than goal furtherances. Nevertheless, being engaged in goal-hindering activities seems to take its toll with respect to cortisol.
In line with expectations, we also showed that the AUC is positively associated with in situ cortisol over and above goal-relevant activities and mood. However, there was no cross-level interaction. This indicates that in our nonclinical sample, overall cortisol secretion did not influence the situation-specific relationships between goal-relevant activities and cortisol that were of central interest to this study.
Strengths and Limitations
The present study has a number of advantages. First, we used an interval-sampling technique to gather information on activities performed in the course of normal work and nonwork days and their relation to self-set goals. Second, in addition to affect quality, we used a physiological outcome. Third, we used differences between situations rather than between persons as the basis for our inferences. In that fashion, the effects cannot be attributed to interindividual differences in the habitual appraisal of situations (37), whereas at the same time, the subjective appraisal of the goal relevance as well as the interindividual differences therein can be preserved as part of the phenomenon of interest.
We also have to mention a number of shortcomings. First, the positive selectivity of the sample limits generalizability to other groups of parents. The consequence may have been an under- rather than an overestimation of the effects, however. Second, the time-sampling approach to everyday life processes confronts the researcher with a multitude of influences, including situational, individual, or couple characteristics (38) with two important consequences: a) Because we observed rather than manipulated independent variables, the links we found between appraisal, affective, and cortisol-related processes do not imply causality. For example, a working parent might feel good because he or she is performing a goal-furthering activity or he or she could be performing goal-furthering activities only when feeling good. b) Because of the multitude of influences, an individual variable often explains only small amounts of variance and its effect size may be rather small (39). We still believe that the processes we observed are of practical relevance because effects of this kind of daily hassles will accumulate over time. For example, the average increase in cortisol levels of 0.32 nmol/L per hour characterized by goal-hindering activities is much lower than the 2 to 7 nmol/L increases observed in laboratory studies using social stressors (40), but individuals may encounter these minor stressors rather frequently. Finally, we cannot rule out that part of the effects on cortisol were mediated by stress-related substance use (coffee, tobacco, alcohol; e.g., (41)), and possible inferences are further limited by the fact that our cortisol and mood samples did not refer to exactly the same time periods.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
The assistance of Sylvia Böhme, Cristina Cretulescu, Christine Hennen, and Kerstin Kaehlert in recruiting and briefing participants and establishing a research relationship with them was most valuable. The participants deserve special thanks for contributing their precious time to this demanding study. We thank Jacqui Smith for her advice and Thomas Rigotti for his comments to prior versions of the manuscript.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
The research reported here was funded by a Volkswagen-Foundation grant to Dr. Klumb. Preparation for this manuscript was supported by a research fellowship awarded by the German Research Foundation (DFG) to Dr. Hoppmann.
DOI:10.1097/01.psy.0000238232.46870.f1
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Ditzen, C. Hoppmann, and P. Klumb Positive Couple Interactions and Daily Cortisol: On the Stress-Protecting Role of Intimacy Psychosom Med, October 1, 2008; 70(8): 883 - 889. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
