why not everyone gets their fair share of stress: adolescent's perceived relationship affection...

European Journal of Personality, Eur. J. Pers. 29: 125–137 (2015)Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/per.1989

Why Not Everyone Gets Their Fair Share of Stress: Adolescent’s PerceivedRelationship Affection Mediates Associations Between Temperament andSubsequent Stressful Social Events

O. M. LACEULLE1,2*, B. F. JERONIMUS2, M. A. G. VAN AKEN1 and J. ORMEL2

1University of Groningen, University Medical Center Groningen, Groningen, The Netherlands2Utrecht University, Developmental Psychology, Utrecht, The Netherlands

*CorrGroniE-mai

Copy

Abstract: Temperamental differences are associated with subsequent stressful life events, a phenomenon that has inpart been attributed to evocation. However, we remain ignorant about the mechanisms that mediate this process. Inthe current paper, we test whether differences in ‘perceived relationship affection’ accounted for part of the prospec-tive association between temperament and stressful social event evocation in three social domains, viz. parents, peersand romantic partners. Data were derived from the Tracking Adolescents’ Individual Lives Survey, a large populationcohort of Dutch adolescents (n=1158). Parent-reported adolescent temperament and adolescent’s perceived affec-tion were assessed at age 11 years. Stressful social events that occurred between age 11 and 16 years were capturedusing the event history calendar. Results indicate that adolescents evoke subsequent stressful social events based ontheir temperament, and that this association is partially mediated by perceived affection. Importantly, we found evi-dence for both generic and domain-specific associations, which indicates that social domains are related yet distinct.Taken together, the findings suggest that a search for mediating variables may be a promising way to increase ourunderstanding of the mechanisms that underlie the social stress selection principle, and that perceived relationshipaffection is one of the candidates. Copyright © 2015 European Association of Personality Psychology

Key words: temperament; perceived affection; stressful events; adolescents; prospective study

Individuals may evoke life events based on individual char-acteristics and are therefore active agents of their own devel-opment (Caspi & Shiner, 2011; Neyer, Mund, Zimmermann& Wrzus, 2013). These social selection and evocation pro-cesses have strong empirical support (Kendler & Baker,2007; Vinkhuyzen, Sluis, Geus, Boomsma & Posthuma,2010), but the underlying mechanisms remain poorly under-stood. Part of the prospective association between tempera-ment and stressful life events may be mediated by thirdvariables, but hitherto, it remains unknown what kind of fac-tors we are looking for, how they manifest themselves, howthey are constituted or where they are to be found. In thisstudy, we test whether individual differences in perceived re-lationship affection mediate the prospective association be-tween temperament and the evocation of stressful socialevents in adolescents. A demonstration of mediation by per-ceived relationship affection could propel the explorationand understanding of the mechanisms that drive the social se-lection principle.

Temperament and stressful social events

Individual differences are thought to be particularly salientduring adolescence, because this period is characterised by

espondence to: Odilia M. Laceulle, ICPE, University Medical Centerngen, University of Groningen, 9713 GZ Groningen, The Netherlands.l: [email protected]

right © 2015 European Association of Personality Psychology

environmental changes and shift of focus and attachmentfrom parents to peers (Caspi & Moffitt, 1993; Cyranowski,Frank, Young, & Shear, 2000). Hallmark of adolescence issocial change, including the selection of a rapidly expandingpeer network, which, in contrast to family and early child-hood (dyadic) peer relations, is not shaped by parentalsocialisation (Back et al., 2011; Neyer et al., 2013; Selfhoutet al., 2010; Wrzus, Hänel, Wagner, & Neyer, 2013). Addi-tionally, romantic relationships emerge. These romancescan be rather intense and function as a socialising agent thatcan affect subsequent development and identity formation,for example, via ‘consensual validation’ or reality confirma-tion via comparison of perceptions (Collins, Welsh, &Furman, 2009; Simon, Aikins, & Prinstein, 2008). Adoles-cents start to perform social roles at multiple stages, in inter-action with either parents, peers or romantic partners. Each ofthese stages can enable stressful social events, for example,conflicts, fights and relationship termination (Furman &Buhrmester, 1992; Furman & Shomaker, 2008; Lempers &Clark, 1992).

Not all adolescents seem equally prone to evoke suchstressful social events. These differences are partly accountedfor by individual differences in temperament or personality(Harris, 2009; Kandler, Bleidorn, Riemann, Angleitner, &Spinath, 2012; Kendler & Baker, 2007; Vinkhuyzen et al.,2010; Scarr & McCartney, 1983). Prospective twin studiesshowed that emotionally instable (versus stable) individualsare more often exposed to subsequent stressful events and,

Received 31 July 2014Revised 2 February 2015, Accepted 9 February 2015

126 O. M. Laceulle et al.

additionally, are also more sensitive to the influence ofstressful events (Middeldorp, Cath, Beem, Willemsen, &Boomsma, 2008; Riese et al., 2014). These processes are of-ten referred to as the corresponsive principle (Jeronimus,Riese, Sanderman, & Ormel, 2014; Roberts, Caspi, &Moffitt, 2003). Other studies showed that high (versus low)extraversion and conscientiousness were related to fewerstressful social events (Lüdtke, Roberts, Trautwein, & Nagy,2011) and high neuroticism to smaller declines in familyconflict (Neyer & Asendorpf, 2001), low family support(Windle, 1992) and more romantic relationship conflict(Caspi, Bem, & Elder, 1989; Jeronimus, Ormel, Aleman,Penninx, & Riese, 2013; Robins, Caspi, & Moffitt, 2002).

In our Tracking Adolescents’ Individual Lives Survey(TRAILS) sample high (versus low) frustration, low effortfulcontrol and high intensity pleasure and affiliation and lowlevels of shyness were related to more stressful life eventsover adolescence—both social and other events such ashouse moves and illnesses (Jeronimus, Riese, Oldehinkel,& Ormel, 2015, Laceulle, van Aken, Ormel, & Nederhof2014). Evidence thus suggests that emotional instabilityand to some extent also conscientiousness and extraversionpredisposes individuals for more stressful social events.

Mechanisms underlying temperamental effects onsubsequent social stressful events

Although support for evocation of stressful social events in-creases gradually, the mechanisms that underlie this processremain largely speculative. Possibly, temperament shapesthe way adults interact with children and the activities inwhich children choose to participate (Harris, 2009; Shanahan& Flaherty, 2001). This, in turn, may affect stressful socialevents children experience, such as peer rejection (Hay,Payne, & Chadwick, 2004). Negative affective temperamentsalso predict increases in problems and exacerbate the effectsof other risk factors (Moran, Lengua, & Zalewski, 2013;Rothbart & Bates, 2007). When children navigate into ado-lescence, their ability to influence their environments in-creases. Consequently, it seems plausible that temperamentbecomes more predictive of stressful social events in variousdomains, for example, in interaction with parents, peers andromantic partners.

Temperament has been suggested to modulate cognitiveand affective-emotional processes (Chan, Goodwin, &Harmer, 2007; Furman & Buhrmester, 1992) that colourhow adolescents perceive their world, a process called ‘envi-ronmental construal’ (Caspi & Shiner, 2011; Rothbart,2011). Additionally, the way we look at ourselves and theworld around us is believed to develop from early childhoodonwards based upon our interpersonal interactions and men-tal imaginations of how we believe others perceive us (i.e. the‘looking glass’ self; Cooley, 1902; Fraley, 2002; Harris,1995, 2009). These belief systems, in turn, affect how ado-lescents perceive and experience their relationships withothers and navigate in their social worlds (Caspi & Shiner,2011; Hartup & Laursen, 1999; Rothbart, 2011; Soto, John,Gosling, & Potter, 2008). It is possible that an individual’sperception of his or her relationships becomes a self-fulfilling

Copyright © 2015 European Association of Personality Psychology

prophecy when perceptions become internalised and influ-ence subsequent transactions with the social environment inwhich the individual lives and grows (Beam & Turkheimer,2013) and thus leads them to construct relationship realities.Differences in temperament may therefore explain both theactive process of self-construction (a process ofinternalisation of perceptions) as well as how adolescent’screate trajectories for themselves as a response, which mayshape how they are seen by others (an externalisationprocess).

Indeed, previous research provided some support for theidea that temperament modulates adolescents’ perceptionsof their affective relationships with others (Branje, VanLieshout, & Gerris, 2007; Hagemeyer, Neyer, Neberich, &Asendorpf, 2013; Rothbart, 2011). In addition, evidencehas been found that people’s perceptions of their relation-ships with others can influence the course and functioningof their relationships (see for a review Bradbury & Fincham,1990). For example, Sroufe (1990) suggested that individualswho expected to be rejected also behaved in ways that maderejection from others more likely. Also, anxious expected re-jection predicted social anxiety and withdrawal, while angryexpected rejection—an established predictor of aggression—antedated decreased social anxiety (London, Downey,Bonica, & Paltin, 2007). Both anxious and angry expecta-tions predicted increased loneliness (London et al., 2007),and individuals who expected to be rejected, those whoscored high (versus low) on rejection sensitivity, seemedmore likely to break up (Downey, Freitas, Michaelis, &Khouri, 1998). Perhaps, subjective expectations and percep-tions form a process chain that underlies part of the prospec-tive association between temperament and stressful lifeevents. The proof of the pudding is in the eating, however,and to our knowledge, ‘environmental construal’ (or ‘internalworking models’ more general) as a mechanism underlyingthe association between temperament and subsequent stress-ful social events have not been tested by means of sophisti-cated mediation models. We hypothesise thattemperamental differences lead to differences in perceivedaffect, which in turn result in the evocation of stressful socialevents. Specifically, in the current study, we tested whetherperceived relationship affection—that is, adolescents subjec-tive experience of care, protection, comfort and approval pro-vided by significant others (i.e. parents, peers)—mediates theprospective association between temperament and stressfulsocial events. For example, we tested whether individualshigh (versus low) on temperamental frustration are morelikely to perceive low relationship affection, because highfrustration may be related to more anger, frustration andwithdrawal, resulting in more conflicts with significantothers. We feel such evidence can bolster future aims to dis-entangle the extant factors that converge into the social selec-tion principle.

Current study

To recapitulate, in the current study, we verify whether ado-lescent’s perceived relationship affection mediates the asso-ciation between adolescent temperament and evocation of

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per

Temperament and stressful social events 127

stressful social events (see Figure 1 for a conceptual model).Perceived relationship affection is not introduced as a noveltheoretical construct but as a measure for subjective interpre-tations of social interactions in functional terms (e.g. instru-mental and emotional support), alike perceived socialsupport, perceived rejection and felt (in)security.

We studied temperamental facets that often are consideredto be part of the broader personality domains of neuroticism,extraversion and conscientiousness, because these traits havebeen found to showmost consistent associations with psycho-pathology (Kotov, Gamez, Schmidt, & Watson, 2010). Foremotional instability, we included two traits (fear and frustra-tion), for extraversion three traits (affiliation, shyness andhigh intensity pleasure) and one trait related to conscientious-ness (effortful control). However, because these higher-orderdomains are not clearly structured over adolescence, we onlyreport upon the facet traits and refrain from usage of thehigher-order dimensions themselves.

The major changes in social environments during adoles-cence render it important to differentiate between the domainsin which stressful social events take place (Furman &Buhrmester, 1992; Harris, 2009). Moreover, the interpretationof meaning of affectionate behaviours may be relatively stablewithin social groups across social contexts but differ across so-cial groups, which we therefore might compare. In this study,we distinguish between the parental domain (e.g. conflictwith parents, running away and being thrown out of theparental home), the peer domain (e.g. friendship terminationafter a fight or argument, being bullied) and the romanticrelationships domain (e.g. breaking up after a relationshipor being dumped). Subsequently, we examined whetherperceived affection mediates the prospective associationbetween temperament and stressful social events. To testfor spill-over effects between different social domains, weexamined mediation effects of both perceived parentalaffection and perceived peer affection in the associations

Figure 1. A theoretical model of how the prospective effect of temperament on strparental and peer affection.


between temperament and stressful social events in theparental, peer and romantic partner domain.

In line with previous studies, we hypothesised that ado-lescents low (versus high) on effortful control (H1a), high(versus low) on frustration (H1b) and high on affiliation orintensity pleasure and low on shyness (H1c) experience(evoke) more subsequent stressful social events. Based on re-cent findings on the data used in the current study, it ishypothesise that fear does not predict subsequent stressfulevents (Laceulle et al., 2014). Additionally, we hypothesisedthat prospective associations between temperament and sub-sequent stressful social events are partially mediated by per-ceived relationship affection (H2). More specific, wehypothesised domain-specific associations, viz. perceivedparental affection as the primary mediator of stressful socialevent evocation effects in the parental domain (H3a) and per-ceived peer affection as the mediator of stressful social eventevocation effects in the peer domain (H3b). Lastly, we testedwhether associations between temperament and subsequentstressful social events in the romantic partner domain weremediated by perceived either parental or peer affection.

METHODS

Sample

The TRAILS is a large prospective cohort study of Dutchadolescents, who are followed biennially or triennially from11 to at least 25 years of age. The present study involves datafrom the first and third assessment wave. The study wasapproved by the Dutch Central Committee on ResearchInvolving Human Subjects. Written informed consent wascollected from the parents at Wave 1, whereas for Wave 3,written informed consent was obtained from both parents andadolescent. At Wave 1, 2230 preadolescents (50.8% girls)

essful social events is divided over direct effects and mediation via perceived

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


enrolled in the study (response rate, 76.0%) of whom, 1816(response rate 81.4%, 45.3% girls) participated in Wave 3.At Wave 1, the mean age of the adolescents enrolled in thestudy was 11.09 years (SD=0.56). At Wave 3, the meanage was 16.13 years (SD=0.59). Prerequisites to be includedin the current study were that, at Wave 1, parents had filledout the temperament questionnaire and adolescent the per-ceived affection list, and that, at Wave 3, the adolescentswere interviewed with regard to stressful social events. Thisresulted in a total number of 1158 adolescents participatingin the current study. No differences were found betweenresponders and nonresponders with respect to teacher ratingsof problem behaviours and in the associations betweensocio-demographic variables and mental health indicators.We examined whether individuals who were interviewedabout exposure to life events differed from those who werenot interviewed on the temperament and affiliation scales atage 11 years. To facilitate comparisons, partial η2 measuresof effects were computed. The effect sizes for beinginterviewed were all smaller than 0.01, which can beinterpreted as negligible effects (Cohen, 1992). Althoughwe observed slightly higher attrition (p< .05) for childrenwith low scores on effortful control and affiliation, effectsizes were negligible (partial η2 = 0.002 and 0.003), and ourresults seem not seriously biassed. A detailed description ofthe sample selection, procedures and methods can be foundin De Winter, Oldehinkel, Veenstra, Brunnekreef, Verhulstand Ormel (2005).

Procedures

At baseline, well-trained interviewers visited one of the parentsor guardians (preferably the mother, 95.6%) at their homes.Parents were asked to fill out a written questionnaire, includingquestions about the child’s temperament. Children wereasked to fill out questionnaires on perceived affiliation inschool. When adolescents were 16 years old, children wereinterviewed at a central facility in the child’s home area bywell-trained interviewers to collect life-event data.

Measures

TemperamentChild temperament was assessed at age 11 with the shortform of the parent version of the Early Adolescent Tempera-ment Questionnaire–Revised (EATQ-R; Hartman, 2000;Putnam, Ellis, & Rothbart, 2001). The following six scaleswere distinguished: (i) fear (negative affect related to anti-cipated pain or distress, five items, Cronbachs’ α= .63);(ii) frustration (negative affect related to interruption of ongo-ing tasks or goal blocking, five items, α= .74); (iii) shyness(slow or inhibited approach and/or discomfort in social situa-tions, four items, α= .84); (iv) effortful control (capacity tocontrol attention, activation and inhibition, 11 items,α= .86); (v) affiliation (desire for and pleasure in warmth andcloseness with others, six items, α= .66); and (vi) high inten-sity pleasure (pleasure or enjoyment related to high stimulusintensity or novelty, six items, α= .77). Answers were ratedon a 5-point Likert-type scale (1= ‘almost always untrue’ to


5= ‘almost always true’). Higher values indicated a higherpresence of the temperamental trait concerned. Eight-weektest–retest stability of the parent-reported EATQ-R scaleshas been found to be moderate to good, ranging from 0.69for high intensity pleasure to 0.85 for frustration (Muris &Meesters, 2009).

Stressful social eventsStressful social events were captured at age 16 years usingthe event history calendar (EHC), a data collection methodfor obtaining retrospective data about life events and activi-ties developed by Caspi et al. (1996) and colleagues. Forthe present study we adapted the calendar into an interviewon several life domains that lasted about 45minutes. Partici-pants were asked about events that occurred since baseline(i.e. between ages 11–16). Detailed and accurate data about theevents could be collected by proceeding serially from one lifedomain to another and using a month-by-month horizontaltimeline. For example, with regard to school, adolescentswere asked by the interviewer, respectively, about the datesof changing school, changing class, repeating class, as wellas about their educational levels for the subsequent years.Test–retest reliability has generally been found to be reason-able to good [respectively, 72–87% in a sample of youngadults (Freedman, Arland Thornton, Camburn, Alwin, &Young-DeMarco, 1988) and >90% in a sample of adoles-cents (Caspi et al., 1996)]. Construct validity of the EHCwas investigated in a comparative study by Belli and col-leagues (2001), showing reasonable correlation coefficientsbetween a written questionnaire and the EHC (ranging from0.63 to 0.79).

For the current study, we selected all stressful socialevents assessed in the parental, peer and romantic domains.Stressful social events were defined as time-discrete eventslikely to bring about a major change in social or relationshipstatus (cf. Holmes & Rahe, 1967; Luhmann, Orth, Specht,Kandler, & Lucas, 2014). Stressful events in the parentaldomain included being thrown out of the parental home(n=20), having a serious fight (n=92) and running awayfrom home (n=52). Stressful events in the peer domain in-cluded losing a good friend because of a fight or argument(n=128) and being bullied (n=256). Stressful events in theromantic partner domain included being dumped (n=204)and breaking up (self) after a relationship (n=538). Intercor-relations between the various events were rather low, rangingfrom r= .001 for the correlation between losing a good friendbecause of a fight or argument and being dumped to r= .215for the correlation between having a fight serious fight withfamily members and running away from home. For thisstudy, three event variables were constructed indicating thenumber of events the adolescents experienced in the respec-tive domains. With regard to the conflicts with parents, beingexposed to two (n=20) and three (n=3) events were mergedand recoded as ‘≥2 events’.

Perceived affectionChild-rated perceived parental and peer affection were mea-sured at age 11 using two scales based on the Social Produc-tion Function (SPF) theory (Nieboer, Lindenberg, Boomsma,

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


& Van Bruggen, 2005). The SPF asserts that well-being canbe measured in terms of universal goals, viz. affection, be-havioural confirmation, status, comfort and stimulation(Ormel, Lindenberg, Steverink, & Vonkorff, 1997). Per-ceived affection from parents (i.e. an aggregated measureof perceived paternal and maternal affection, each four items,e.g. ‘he/she likes being with me’, α= .84 for paternal affec-tion and α= .78 for maternal affection) and perceived affec-tion from classmates (four items, e.g. ‘my classmates enjoybeing with me’, α= .84 ) were measured with five-pointscales, with answer categories ranging from 1 (never) to 5(always). No test–retest data of the SPF list are available.

Table 1. Descriptive statistics for the unstandardized variables

Temperament N Min Max Mean SD

Fear 1196 1.00 4.60 2.41 0.71Frustration 1196 1.00 4.80 2.77 0.65Affiliation 1196 1.50 5.00 3.89 0.55Shyness 1197 1.00 5.00 2.51 0.86High intensity pleasure 1194 1.00 5.00 3.30 0.93Effortful control 1197 1.09 5.00 3.28 0.69Perceived parental affection 1162 3.00 10.00 8.64 1.27Perceived peer affection 1183 2.00 10.00 7.41 1.40Events parents 1197 0 2 0.13 0.390 events 10591 event 1152 events 23

Events peers 1197 0 2 0.32 0.530 events 8491 event 3122 events 36

Events romantic relations 1197 0 2 0.62 0.660 events 5741 event 5042 events 119

Gender 1197Female 660Male 537

Statistical analyses

Variables were transformed into z-scores for both the correla-tion and mediation analyses. Subsequently, we examined thedirect effects of temperament on subsequent stressful socialevents, as well as, the possible mediating role of perceivedparental affection and perceived peer affection in these asso-ciations, as outlined in Figure 1. Three mediation analyseswere performed for each of the six temperament traits, onefor each of the stressful social event domains (i.e. parents,peers and romantic partners). Perceived parental affectionand perceived peer affection were entered simultaneously inthe analyses, resulting in a total of 18 analyses. All mediationanalyses were controlled for gender using the single multiplemediation method proposed by Preacher and Hayes (2008).

The theoretical model of the prospective association be-tween temperament and stressful social events in Figure 1shows a direct effect on stressful social events (path c) and thedirect effect when the indirect path is controlled for (path c′).The relationships between temperament and the two media-tors are figured through path a1 and path a2. The effects ofthe two mediators on stressful social events are figuredthrough path b1 and b2. The total indirect paths fromtemperament to stressful social events are the sum of thetwo mediators. Testing a single multiple mediation model(rather than separate simple mediation models) has the advan-tage of allowing intercorrelations between the respectivemediation variables (Preacher & Hayes, 2008). Additionally,a single multiple mediation model enables us to quantify theextent to which perceived parental affection mediates theeffect of temperament on stressful social events, conditionalon the presence of perceived peer affection (and vice versa).The linear regression technique is known to remain valid whenthe dependent variable violates the ‘normality assumption’ in asample of our size (Li, Wong, Lamoureux, & Wong, 2012).However, to ensure the robustness of our results, webootstrapped all linear regression analyses (k=1000 with biascorrected confidence intervals) to obtain asymptotic 95%confidence intervals around the indirect effects using the SPSS(IBM Inc., Armonk, NY, USA) macro developed by Preacherand Hayes (2008). Confidence intervals not including zeroreflect significant indirect effects. To enable comparison withother literature, we converted some results to Cohen’s d(standardised effect sizes), based on formulas derived fromBorenstein (2009) and Peterson (2005). To reduce family-wise


alpha inflation, we only interpreted correlations that weresignificant at p< .01.

RESULTS

Descriptive statistics

Descriptive statistics for the unstandardized variables are re-ported in Table 1. Table 2 presents correlations between thesix temperament traits, perceived affection by parents andpeers and stressful social events in the three domains. We re-frain from using higher-order dimensions (e.g. neuroticismcomposed from fear and frustration) because the higher-orderdomains could not be clearly distinguished in adolescents(Table 2, cf. overlap between frustration and effortfulcontrol).

Direct effects and (partial) mediation

Most temperamental traits showed a direct prospective path-way to stressful social events. Adolescents high on frustra-tion and low on effortful control were more likely toexperience stressful social events in the parental and peerdomain but not with romantic partners. Adolescents high(versus low) on intensity pleasure, low (versus high) onshyness and low (versus high) on affiliation were more likelyto experience stressful social events in the romantic partnerdomain but not with parents or peers. Only fear was unre-lated to subsequent stressful events in all three domains.

Lower parental affection was observed for adolescentshigh (versus low) on frustration and low (versus high) effort-ful control and affiliation. Lower perceived peer affectionwas reported for adolescents high (versus low) on frustrationand shyness or low on effortful control and affiliation.

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per

Table 2. Correlations between the study variables

1. Fear 2. 3. 4. 5. 6. 7. 8. 9. 10.

2. Frustration .30***3. Affiliation .08** �.19***4. Shyness .14*** .12*** �.29***5. High intensity pleasure �.25*** �.06 .15*** �.29***6. Effortful control �.23*** �.37*** .12*** �.01 .09**7. Perceived parental affection .04 �.13*** .14*** �.00 �.01 .12***8. Perceived peer affection .03 �.13*** .16*** �.06 .00 .10*** .46***9. Events parents .05 .12*** .01 �.02 .04 �.13*** �.06 .0010. Events peers .07* .07* .05 .02 �.03 �.07* �.06 �.06* .09**11. Events romantic partners �.01 .02 .07* �.11*** .14*** �.02 �.01 .08** .14*** �.00

Note: *p< .05; **p< .01; ***p< .001.


Perceived affection, in turn, predicted stressful social events.More perceived parental affection predicted less subsequentstressful events in the parental domain and in the romanticpartner domain and more perceived peer affection predictedless subsequent stressful events in the peer domain but moreevents in the romantic partner domain. Although path coeffi-cients varied slightly across the temperament models (i.e. as-sociations between perceived affection and stressful eventswere estimated in each of the univariate temperamentmodels), they did not differ in any meaningful matter depen-dent on which temperament trait was included in the model(Figure 2 shows the path coefficients for events in the roman-tic partner domain). Furthermore, high levels of perceived af-fection predicted fewer subsequent social stressful events. Insum, the observed associations varied across temperamenttraits, stressful social event domains and types of perceivedaffect (mediators).

Our bootstrapped regression models (Table 3) showedthat several of the associations between temperament andstressful social events were mediated by perceived relation-ship affection. Perceived parental affection mediated the as-sociations between frustration, affiliation and effortful

Figure 2. Standardised path coefficients for the direct paths. Path coefficients ofacross models.


control and later stressful social events in the romantic part-ner domain. That is, higher levels of frustration, lower levelsof affiliation and lower levels of effortful control were all re-lated to less perceived parental affection, which in turn pre-dicted more stressful social events in the romantic partnerdomain. Also, perceived parental affection mediated the as-sociations between affiliation and stressful social events inthe parental domain. So, lower affiliation was related to lessperceived parental affection, which, in turn, predicted morestressful social events in the romantic partner domain.

Perceived peer affection mediated the associations be-tween, respectively, frustration, affiliation, shyness and ef-fortful control and stressful social events in both the peerand romantic partner domain. That is, higher levels of frus-tration and shyness and lower levels of affiliation and effort-ful control were related to less perceived peer affection,which, in turn predicted more stressful social events in thepeer domain, but less in the romantic partner domain. No me-diation effects were found for the associations between, re-spectively, fear and high intensity pleasure and any of thesocial domains. All significant mediation effects had a rathersmall-effect size (between d=0.10 and 0.15).

the associations between affection and stressful social events varied slightly

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per

Table 3. Bootstrap results for indirect relationships (bias-correctedand accelerated confidence intervals (CIs))

Boot SELower95% CI

Upper95% CI

Fear →Events PaParental affection �.002 .003 �.010 .001Peer affection .000 .001 �.002 .004Fear → Events PeParental affection �.001 .002 �.009 .001Peer �.001 .003 �.009 .003Fear → Events RoPaParental affection �.002 .003 �.010 .001Peer .001 .003 �.005 .010Frustration → Events PaParental affection .009 .005 �.001 .021Peer �.003 .005 �.013 .006Frustration → Events PeParental affection .004 .005 �.005 .015Peer .009 .005 .000 .022Frustration→Events RoPaParental affection .009 .005 .001 .023Peer �.013 .005 �.028 �.005Affiliation → Events PaParental affection �.009 .005 �.023 �.001Peer .001 .005 �.009 .011Affiliation → Events PeParental affection �.005 .005 �.017 .004Peer �.011 .006 �.026 �.001Affiliation →Events RopaParental affection �.010 .005 �.022 �.002Peer .013 .005 .005 .025Shyness → Events PaParental affection .001 .002 �.003 .007Peer �.001 .003 �.007 .005Shyness → Events PeParental affection .000 .002 �.001 .007Peer .006 .004 .001 .017Shyness →Events RoPaParental affection .001 .002 �.003 .007Peer �.007 .004 �.017 �.001High Int Pl → Events PaParental affection .000 .002 �.005 .005Peer .000 .001 �.002 .004High Int Pl → Events PeParental affection .000 .002 �.003 .003Peer �.001 .003 �.009 .002High Int Pl→Events RoPaParental affection .000 .002 �.005 .005Peer .002 .003 �.004 .010Eff Contr → Events PaParental affection .006 .005 �.17 .001Peer .001 .003 �.004 .008Eff Contr → Events PeParental affection .003 .004 �.013 .004Peer �.005 .003 �.014 �.000Eff Contr→Events RoPaParental affection �.007 .004 �.018 �.001Peer .008 .004 .001 .017

Note: Events Pa, events in the parental domain; Events Pe, events in the peerdomain; Events RoPa, events in the romantic partners domain.Bold emphases show significant associations. SE = Standard Error.


Post-hoc analyses

Multiple post-hoc analyses were performed to test the robust-ness of our results. First, the effects of temperament onstressful social events in the romantic partner domain were


repeated in the subgroup of adolescents who reportedly wereinvolved in at least one romantic relationship between age 11and 16. Second, we tested all temperamental effects for eachof the individual stressful events. Third, multivariate media-tion analyses were performed in which all temperament traitswere entered simultaneously to test for (the mediation of) theindependent effects of each temperamental facet adjusted forthe effect of all other facets. Finally, nonparametrical Spear-man partial rho tests were performed as an extra robustnesscheck (next to our bootstrap procedure) because our depen-dent variables were non-normally distributed.

Adolescents involved in a romantic relationshipAll significant associations between temperament and stress-ful events in the romantic partner domain disappeared in themuch smaller sample of adolescents who reported at leastone romantic relationship between age 11 and 16 (n=703;Supplementary material Table S1).

Single-event analysesAnalyses for all single events in Supplementary materialTable S2 showed that high (versus low) fear predicted runningaway from home (parental domain) and being bullied (peerdomain). High (versus low) frustration predicted fights withparents and running away from home (parental domain) andbeing bullied (peer domain). Adolescents high (versus low)on shyness were less likely to have a fight with their parents(parental domain) and less likely of being dumped or breakingup a relationship (partner domain). Adolescents high (versuslow) on surgency were more often thrown out of the parentalhome and were more often dumped or broke up theirromantic relationship (partner domain). Finally, adolescentshigh (versus low) on effortful control reported fewer seriousfights with parents and running away from home (parentaldomain) and were bullied less (peer domain). Most mediationpaths by perceived relationship affection remained significant(Supplementary material Table S2).

Multivariate analysesMultivariate analyses were performed in which all tempera-ment traits were entered simultaneously to examine the ef-fects of temperament traits adjusted for all other traits.Analyses showed that only part of the associations found inthe univariate analyses remained in the multivariate analyses.Most importantly, the indirect effects of frustration and affil-iation on stressful events via perceived affection remainedwhen adjusting for the other temperament traits. In contrast,the effects of shyness and effortful control disappeared, sug-gesting that these were not robust when adjusting for othertraits. Model statistics are reported in Supplementary mate-rial Table S3.

Nonparametric testsFinally, nonparametrical Spearman partial rho tests sup-ported the results of our univariate analyses and showedboth the direct effects of temperament on stressful eventsoccurrences and mediation of these associations by per-ceived parental and/or peer affection (Supplementary mate-rial Table S4).

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


DISCUSSION

In this paper, we used data from a large cohort of adolescentsand two waves to test whether adolescent’s perceived rela-tionship affection mediates part of the prospective associa-tion between adolescent temperament and stressful socialevent evocation. Our results support the hypothesis that tem-peramental differences are manifested in differences instressful social event evocation, in line with previous re-search. Our study innovated by the observation that per-ceived relationship affection mediates a modest part of thisassociation. In other words, temperaments colour the way ad-olescents perceive received affection, which, in turn, influ-ences the probability of subsequent stressful social events.Our distinction between three social domains (i.e. parents,peers and romantic partners) yielded support for both genericand domain-specific effects, which exemplifies the challengeof isolating mechanisms behind the stress selection principle.After having summarised our main findings, these will bediscussed in more detail subsequently.

Temperament and subsequent stressful social events

Low levels of effortful control (H1a) and high levels of frus-tration (H1b) were predictive of more subsequent socialstressful events, in line with our hypotheses. Interestingly,this held only true for the parental and peer domain; stressfulsocial events in the romantic partner domain were predictedby high levels of intensity pleasure and affiliation as wellas low levels of shyness (all traits related to the broader per-sonality domain of extraversion). Fear was not predictive ofstress in any of the domains.

The observation that patterns were different for peers andromantic partners might be somewhat surprising insofarthose relationships between peers and romantic partners areoften seen as more comparable than between parents and ro-mantic partners (Furman & Buhrmester, 1992; Furman &Shomaker, 2008; Hartup, 1989). However, it might be thatabove and beyond the similarities between peer and romanticpartner relations, peer relations have some resemblance withparental relations in the sense that they have both developedover years. Romantic relationships, in contrast, can develop(and finish) rather suddenly during adolescence. Conse-quently, temperament traits that are undesirable in social in-teractions (e.g. high frustration) may be more visible forparents and peers who know the adolescent for years,whereas adolescents may inhibit frustration-related behav-iours in the presence of their new romantic partner. Thismay explain why adolescents high on frustration may evokestressful social events in the parental and peer domain but notthe romantic partner domain.

Similarly, adolescents high on effortful control may bemore reliable in their friendships with peers and ‘easier’ to theirparents andmay therefore be less likely to evoke stressful socialevents in these contexts. In contrast, effortful control may notbe of much importance in the newly developing—and oftenonly short-term—romantic relationships adolescents have. Thisinterpretation aligns with observations by Furman (2002), whosuggested that parent–child relationship characteristics as


perceived by the adolescent were related to both (i) child–peerrelationship characteristics as perceived by the adolescent and(ii) child–romantic partner relationship characteristics as per-ceived by the adolescent, although the latter two were (at leastfor some relationship characteristics) unrelated.

Stressful social events in the romantic partner domain werepredicted by low shyness and affiliation and high intensitypleasure, traits that are all related to the broader personalitydimension of extraversion. Adolescents high on extraversionevoked more stressful events with romantic partners (in linewith hypothesis H1c) but not with parents and peers. Hence,extraversion (being out-going, sociable etc.) seems more in-fluential when engaging with romantic partners than in themore persistent relationships with parents and peers. This isconsistent with evidence that high (versus low) extravert ado-lescents engage more in romantic relationships (e.g. Lüdtkeet al., 2011; Neyer & Lehnart, 2007; Specht, Egloff, &Schmukle, 2011). It therefore seems plausible that low extra-vert adolescents evoke less social stressful events in the ro-mantic partner domain simply because they are not so muchinvolved yet in romantic relationships (a floor effect). Biserialcorrelations indeed showed that high affiliation and surgencyand low shyness predicted the presence of romantic relation-ships between age 11 and 16, while fear, frustration andeffortful control were unrelated (Supplementary materialTable S5).

However, it seems unlikely that the more frequent en-gagement of extraverted adolescents in romantic relation-ships explains all of the variance, because they probablyinteract more with romantic partners and with their peers.The cardinal features of high extraversion are social attentionand a larger impact on one’s social environments in general(Larsen & Buss, 2013; Nettle, 2007). Previous research in-deed showed both quantitative and qualitative differencesin social interactions between high-extraverted adolescents(Berry & Hansen, 1996). Given that we did not find an asso-ciation between traits related to extraversion and subsequentstressful social events in the peer domain, it seems plausiblethat extraverts have not just more but also other (i.e. more in-tense) interaction with romantic relationships than their lessextravert peers. It seems that adolescents high on extraver-sion had more frequent and more intense interactions with ro-mantic partners than adolescents whom were more reticent,resulting in both more positive and negative events. None-theless, it should be noted that the association between tem-perament (shyness, affiliation and high intensity pleasure)and stressful events disappeared in post-hoc analyses in thesubgroup of adolescents who reported at least one romanticrelationship between age 11 and 16. Although this seems tobe (partly) due to power issues, caution is needed wheninterpreting the effects of temperament on events in the ro-mantic partner domain.

Perceived affection: an intrapsychic characteristic withreal-world consequences

We proposed that the prospective association between tem-perament and subsequent stressful social events would bepartially mediated by perceived relationship affection (H2).

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


Indeed, perceived relationship affection mediated several ofthe associations between temperament and stressful socialevents. This finding may propel the exploration of other fac-tors that can account for part of the association between tem-perament and stressful event evocation, because we showedthat mediation studies have the potential to provide insightin the mechanisms underlying the stress selection principle.Such insights may enable clinicians to craft prevention strat-egies that alleviate stress-related psychopathology. Morespecifically, our results suggest that perceived relationshipaffection mediated part of the studied associations. More-over, although small, the observed effect sizes of the variouspaths are in the range of the average observed in psychology(Richard, Bond Jr., & Stokes-Zoota, 2003; Roberts, Kuncel,Shiner, Caspi, & Goldberg, 2007).

Our results align with the interpretation that individualsdevelop internal working models based on their temperamen-tal characteristics, which, in turn, modulates adolescents’perceptions of their affective relationships with others, asoutlined in the introduction. Hence, perceived relationshipaffection, a rather complex intrapsychic characteristic, canhave real-world consequences in terms of subsequent stress-ful social events. This is reminiscent of studies that showedthat preconceptions of a future identity (also an intrapsychicinfluence) can already change temperament (or personality)in anticipation of future social roles (Wood & Roberts,2006a, 2006b). Moreover, the findings confirm and extentprevious research showing that people’s perceptions of theirrelationships with others (e.g. expected rejection) can be-come a self-fulfilling prophecy when people start behavingin ways (e.g. withdrawal and aggression) that elicit stressfulsocial interaction (e.g. conflicts, rejection and breakup; seeBradbury & Fincham, 1990; Downey et al., 1998; Londonet al., 2007; Sroufe, 1990).

Domain specificity and spill-over effects

We hypothesised that mediation by perceived relation affec-tion would be largely domain-specific (H3). Indeed, parentalaffection was the primary mediator of temperamental stressfulsocial event evocation in the parental domain (H3a), whereasperceived peer affection mediated the evocation of stressfulevents in the peer domain (H3b). Perceived parental affectionmediated part of the association between affiliation and eventsin the parental domain, but surprisingly, no effect was foundfor the other temperamental traits. Although perceived affectmay be conceptually most akin to affiliation, the associationbetween affiliation and perceived parental affection was notmuch stronger than it was for frustration or effortful control.Moreover, affiliation was rated by the mother, whereas levelsof perceived affection were based on adolescents’ self-report,which may limit the overlap between both concepts.

With regard to perceived peer affection, several media-tion effects were found. Adolescents lower on frustration orshyness and/or higher on affiliation or effortful control re-ported more peer affection, which, in turn, predicted less sub-sequent stressful social events in the peer domain. Thissuggests that, as hypothesised, adolescent’s perceived affec-tion received from their peers is important in the association


between temperament traits and events in the peer domain.Notably, for both frustration and effortful control, no directeffects were found on stressful social events in the peer do-main. This indicates that perceived peer affection mediatespart of the stress-evocation effects of frustration and effortfulcontrol, whereas extraversion-driven evocation effects tendto be more independent of perceived affection.

Because perceived romantic partner affection was notmeasured in our study, we explored ‘spill-over’ effects of pa-rental and peer affection on stressful event selection in the ro-mantic partner domain. Findings suggested some spill-overeffects. Adolescents lower on frustration and/or higher onaffiliation or effortful control reported more affection fromparents and peers, which, in turn, predicted subsequent stress-ful social events in the romantic partner domain.Whereasmoreperceived parental affection predicted fewer events in the ro-mantic partner domain (in line with the negative associationbetween perceived parental affection and events in the paren-tal domain), more perceived peer affection predicted moreevents in the romantic partner domain (a positive association,diametrical to the negative association between perceivedpeer affection and events in the peer domain). Possibly ro-mantic partners and peers compete for the adolescents’ atten-tion (Furman & Shaffer, 2003; Laursen & Williams, 1997;Zimmer-Gembeck, 2002), which may explain part of thenegative association between peer affection and events inthe romantic relationship. Moreover, the spill-over effects ofperceived parental affection to the romantic domain mayreflect that young adolescents use their perceptions of theirparents to guide their behaviour in interaction with their (first)romantic partners (Linder, Crick, & Collins, 2002).

Yet, and as mentioned earlier, we have to interpret thementioned ‘spill-over’ effects with caution. Besides that theydid not hold in the post-hoc analyses where we examined theassociations only for those adolescents who reported at leastone romantic relationship between age 11 and 16, we wereunable to include perceived romantic partner affection, whilepart of the observed spill-over effects may reflect overlap be-tween the perceived parental, peer and romantic partner af-fection, which might disappear when a measure ofperceived romantic partner affection was included. Clearly,additional research is warranted elaborating on perceived pa-rental and peer affection with perceived romantic partner af-fection and can test our explanations of the alleged spill-overeffects we observed.

Finally, multivariate analyses were performed to test therobustness of the indirect effects when adjusting for all othertraits. Results showed that the effects of affiliation and frus-tration remained in these more conservative analyses, bol-stering the robustness of the findings. However, this wasnot true for the indirect effects of shyness and effortful con-trol disappeared, which suggests that these findings resultedfrom their co-occurrence with the other traits, and these re-sults should therefore be interpreted cautiously.

Strengths and limitations

Among the strengths of our study was our usage of a largesample of adolescents and data from different informants.

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


Parents rated the adolescent’s temperament. Perceived affec-tion was measured using adolescent’s self-report data. Thestressful social events were captured using a semi-structuredand sophisticated interview method, which provided infor-mation both on the nature and the timing of the event. Retro-spective self-reports of stressful events have inherentlimitations because of response components that may be in-fluenced by current mental state, such as cognition, appraisal,interpretation and recall. However, this was addressed in ourstudy by asking the participant to proceed serially from onelife domain to another using a month-by-month horizontaltimeline and under supervision of the interviewer. Thismethod is known for providing detailed and accurate dataabout the events (Caspi et al., 1996; Freedman et al., 1988).In addition, we were able to distinguish between three do-mains of stressful social events: parents, peers and romanticpartners. This allowed us to differentiate between genericand domain specific with regard to both direct and indirecteffects. Finally, to test the robustness of the findings, wetested all temperamental effects for each of the individualstressful events. The direct, as well as, most mediation pathsby perceived relationship affection remained significant,bolstering the robustness of the findings presented.

Despite these strengths, the study is limited in severalways. First and most important, we found evidence for medi-ation of several temperaments to stressful social event associ-ations by perceived relationship affection, but all effectsindicated only partial mediation. Moreover, the effects re-vealed were all very small. It would be interesting to examinewhether they would hold when including constructs relatedto perceived affection, such as social support, attachmentstyle or rejection sensitivity. Unfortunately, however, wedid not have this data available. Nevertheless, research inolder adolescents showed that felt insecurity mediates theassociations between personality (i.e. attachment style) andrelationships with romantic partners (Sadikaj, Moskowitz,& Zuroff, 2015). Additionally, Finn and colleagues (2013)showed that relationship-specific interpretation bias can ex-plain part of the association between personality (i.e. neurot-icism) and relationships with romantic partners.

A related issue is that convergent mechanisms may un-derlie the association between temperament and stressful so-cial events. For some of the paths, we found indirect but nototal effects. As discussed extensively by Preacher and Hayes(2008), total effects are no statistical prerequisite for the ex-istence of indirect effects. It might be, for example, that othermediation processes work in the opposite direction therebyleading to the lack of an overall effect (Küfner, Nestler, &Back, 2013). These observations thus stress the need for fu-ture research, and clearly, other extant variables may accountfor additional variance. Although beyond the scope of thecurrent study, future research may identify other and poten-tially stronger factors in the domain of information processingand add other-reported mediators and outcomes. For exam-ple, parent-reported parenting styles, cognitive characteristicsreported by the adolescent, parent or teacher might be impor-tant mediators to consider. Indeed, elsewhere, in the currentspecial issue, evidence has been provided bolstering theimportance of cognitive characteristics by showing that oral


fluency partially mediates the associations between extraver-sion and sociometric popularity (Ilmarinen, Vainikainen,Verkasalo, & Lönnqvist, 2015). Other mechanisms suggestedto underlie the association between personality and subse-quent peer relations are interpersonal motives and behaviours(Ackerman & Corretti, 2015).

Second, and as mentioned earlier, we included three do-mains of stressful social events in our study but had onlymeasures of perceived affection in the parental and peer do-main. As discussed previously, this makes interpretation ofthe current ‘spill-over’ effects difficult. Third, it might be thattemperament is not only related to an adolescent’s perceivedrelationship affection but also to the absolute amount of af-fection he or she receives. Future research including a mea-sure of perceived affection as well as a more objectivemeasure of actually received affection (maybe even a behav-ioural measure) may help to disentangle this issue. Fourth,our study may be limited by the timing of the perceived af-fection measures. In an optimal mediation design, the medi-ator is assessed in between the predictor and outcomevariable. In our study, perceived affection was measured si-multaneously with temperament. Consequently, we cannotbe conclusive about the direction of the association betweentemperament and perceived affection. However, our researchquestion did not fit very well to the classic mediation ap-proach. By measuring perceived affection in-between tem-perament and stressful life events, we would either includestressful social events that happened before the measurementof the mediator, or have to exclude all events that occurred inthis period, leaving us with a ‘black-time-box’ filled withevents not taken into account.

Related to the direction of the association between tem-perament and perceived affection is the direction of the rela-tionship between temperament and stressful events. Asproposed by the social selection principle, individuals mayevoke stressful events based on their temperament. However,the opposite is also true, and exposure to stress has beenfound to be related to (non-normative) changes in tempera-ment (Laceulle, Nederhof, Karreman, Ormel, & van Aken,2012). The corresponsive principle explicitly accounts forthese bidirectional associations between temperament andstressful events, postulating that change in temperamentresults from mutually reinforcing person–environment trans-actions, including both social selection and social influence(i.e. temperament can affect stress exposure and stress canaffect temperament; Caspi, Roberts, & Shiner, 2005;Jeronimus et al., 2014; Riese et al., 2014). A recent studyon the TRAILS data supported the corresponsive principle,although the effect varied between the different temperamenttraits (Laceulle et al., 2014). Whereas stressful events werefound to predict subsequent fear, stressful events were pre-dicted by but not predictive of shyness and affiliation. For ef-fortful control and frustration, a fully reciprocal model wasfound. Consequently, future research may explore whatmakes frustration and effortful control different, and furtherresearch including multiple waves of temperament, per-ceived affection and stressful social event data may allowfor more detailed test of mediation and shed more light onthe causal order of the various associations.

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


CONCLUSION

In conclusion, in this study a model was tested in which weproposed that adolescent perceived relationship affection me-diates the association between adolescent temperament andevocation of stressful social events in three social domains(parents, peers and romantic partners). Findings indicate thatindividuals may evoke subsequent stressful social eventsbased on their temperament and that this associations is par-tially mediated by adolescents’ perceived relationship affec-tion. The observed effect sizes were small but rangedaround the average in psychology. This suggests that asearch for third variables may be a promising way to learnto understand the mechanism that underlie the stress selec-tion principle and that perceived relationship affect may beone of the candidates.

ACKNOWLEDGEMENTS

This research is part of the TRAILS. Participating centres ofTRAILS includes various departments of the UniversityMedical Center and University of Groningen, the ErasmusUniversity Medical Center Rotterdam, the University ofUtrecht, the Radboud Medical Center Nijmegen, and theParnassia Bavo group, all in the Netherlands. TRAILS hasbeen financially supported by various grants from theNetherlands Organization for Scientific Research (NWO;Medical Research Council Programme grant GB-MW 940-38-011; ZonMW Brainpower grant 100-001-004; ZonMwRisk Behaviour and Dependence grants 60-60600-98-018and 60-60600-97-118; ZonMw Culture and Health grant261-98-710; Social Sciences Council medium-sized invest-ment grants GB-MaGW 480-01-006 and GB-MaGW 480-07-001; Social Sciences Council project grants GB-MaGW457-03-018, GB-MaGW 452-04-314 and GB-MaGW 452-06-004; NWO large-sized investment grant 175.010.2003.005); the Sophia Foundation for Medical Research (projects301 and 393), the Dutch Ministry of Justice (WODC), theEuropean Science Foundation (EuroSTRESS project FP-006), and the participating universities. We are grateful toall adolescents, their parents and teachers who participatedin this research and to everyone who worked on this projectand made it possible.

SUPPORTING INFORMATION

Additional supporting information may be found in theonline version of this article at the publisher’s web-site.

REFERENCES

Ackerman, R., & Corretti, C. (2015). Pathological personality traitsand intimacy processes within roommate relationships. EuropeanJournal of Personality, 29, 152–172

Back, M. D., Baumert, A., Denissen, J. J. A., Hartung, F., Penke, L.,Schmukle, S. C., et al. 2011). PERSOC: A unified framework forunderstanding the dynamic interplay of personality and social


relationships. European Journal of Personality, 25(2), 90–107.DOI: 10.1002/per.811

Beam, C. R., & Turkheimer, E. (2013). Phenotype-environmentcorrelations in longitudinal twin models. Development and Psy-chopathology, 25(1), 7–16. DOI: 10.1017/S0954579412000867

Belli, R. F., Shay, W. L., & Stafford, F. P. (2001). Event history cal-endars and question list surveys: A direct comparison ofinterviewing methods. Public Opinion Quarterly, 65(1), 45–74.DOI: 10.1086/320037

Berry, D., & Hansen, J. (1996). Positive affect, negative affect, andsocial interaction. Journal of Personality and Social Psychology,71(4), 796–809. DOI: 10.1037//0022-3514.71.4.796

Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R.(2009). Introduction to meta-analysis. United Kingdom: JohnWiley & Sons Ltd.

Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage:Review and critique. Psychological Bulletin, 107(1), 3–33. DOI:10.1037/0033-2909.107.1.3

Branje, S. J. T., Van Lieshout, C. F. M., & Gerris, J. R. M. (2007).Big Five personality development in adolescence and adulthood.European Journal of Personality, 21(1), 45–62.

Caspi, A., Moffitt, T. E., Thornton, A., & Freedman, D. (1996). Thelife history calendar: A research and clinical assessment methodfor collecting retrospective event-history data. InternationalJournal of Methods in Psychiatric Research, 6(2), 101–114.DOI: 10.1002/(SICI)1234-988X(199607)6:2<101::AID-MPR156>3.3.CO;2-E

Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality de-velopment: Stability and change . Annual Review of Psychology,56(1), 453–484. DOI: 10.1146/annurev.psych.55.090902.141913

Caspi, A., & Shiner, R. (2011). Temperament and personality. In M.Rutter, D. Bishop, D. Pine, S. Scott, J. Stevenson, E. Taylor & A.Thapar (Eds.), Rutter’s child and adolescent psychiatry (5th ed.,pp. 182). Malden, Massachusetts, USA: Blackwell PublishingLimited.

Caspi, A., Bem, D. J., & Elder, G. H. (1989). Continuities and con-sequences of interactional styles across the life course. Journal ofPersonality, 57(2), 375–406. DOI: 10.1111/j.1467-6494.1989.tb00487.x

Caspi, A., & Moffitt, T. E. (1993). When do individual differencesmatter? A paradoxical theory of personality coherence. Psycho-logical Inquiry, 4(4), pp. 247–271.

Chan, S. W. Y., Goodwin, G. M., & Harmer, C. J. (2007). Highlyneurotic never-depressed students have negative biases in infor-mation processing. Psychological Medicine, 37(9), 1281–1291.DOI: 10.1017/S0033291707000669

Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1),155.

Collins, W. A., Welsh, D. P., & Furman, W. (2009). Adolescent ro-mantic relationships. Annual Review of Psychology, 60(1),631–652.

Cooley, C. H. (1902). The looking-glass self. Human nature and thesocial order (pp. 179–185). New York: Scribner’s.

Cyranowski, J., Frank, E., Young, E., & Shear, M. (2000). Adoles-cent onset of the gender difference in lifetime rates of major de-pression: A theoretical model. Archives of General Psychiatry,57(1), 21–27. DOI: 10.1001/archpsyc.57.1.21

De Winter, A., Oldehinkel, A., Veenstra, R., Brunnekreef, J.,Verhulst, F., & Ormel, J. (2005). Evaluation of non-response bias in mental health determinants and outcomesin a large sample of pre-adolescents. European Journal ofEpidemiology, 20(2), 173–181. DOI: 10.1007/s10654-004-4948-6

Downey, G., Freitas, A., Michaelis, B., & Khouri, H. (1998). Theself-fulfilling prophecy in close relationships: Rejection sensitiv-ity and rejection by romantic partners. Journal of Personality andSocial Psychology, 75(2), 545–560. DOI: 10.1037/0022-3514.75.2.545

Finn, C., Mitte, K., & Neyer, F. J. (2013). The relationship-specificinterpretation bias mediates the link between neuroticism and

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


satisfaction in couples. European Journal of Personality, 27(2),200–212. DOI: 10.1002/per.1862

Fraley, R. (2002). Attachment stability from infancy to adulthood:Meta-analysis and dynamic modeling of developmental mecha-nisms. Personality and Social Psychology Review, 6(2),123–151. DOI: 10.1207/S15327957PSPR0602_03

Freedman, D., Arland Thornton, Camburn, D., Alwin, D., &Young-DeMarco, L. (1988). The life history calendar: A tech-nique for collecting retrospective data. Sociological Methodol-ogy, 18, 37–68.

Furman, W., & Shaffer, L. (2003). The role of romantic relation-ships in adolescent development. In P. Florsheim (Ed.), Adoles-cent romantic relations and sexual behavior: Theory, research,and practical implications (pp. 3). Mahwah: Psychology Press.

Furman, W. (2002). The emerging field of adolescent romantic rela-tionships. Current Directions in Psychological Science, 11(5),177–180.

Furman, W., & Buhrmester, D. (1992). Age and sex differences inperceptions of networks of personal relationships. Child Develop-ment, 63(1), 103–115. DOI: 10.1111/j.1467-8624.1992.tb03599.x

Furman, W., & Shomaker, L. B. (2008). Patterns of interaction inadolescent romantic relationships: Distinct features and links toother close relationships. Journal of Adolescence, 31(6),771–788. DOI: 10.1016/j.adolescence.2007.10.007

Hagemeyer, B., Neyer, F. J., Neberich,W., &Asendorpf, J. B. (2013).The ABC of social desires: Affiliation, being alone, and closenessto partner. European Journal of Personality, 27, 442–457.

Harris, J. R. (1995). Where is the childs environment? A group so-cialization theory of development. Psychological Review, 102(3),458–489. DOI: 10.1037/0033-295X.102.3.458

Harris, J. R. (2009). The nurture assumption: Why children turn outthe way they do (2nd Edn.). New York, USA: Free Press, Simon& Schuster, Inc.

Hartman, C. A. (2000). Dutch translation of the early adolescenttemperament questionnaire. Internal Report, Department of Psy-chiatry, University of Groningen: the Netherlands,

Hartup, W. W. (1989). Social relationships and their developmentalsignificance. American Psychologist, 44(2), 120–126. DOI:10.1037/0003-066X.44.2.120

Hartup, W., & Laursen, B. (1999). Relationships as developmentalcontexts: Retrospective themes and contemporary issues. Rela-tionships as Developmental Contexts, 30, 13–35.

Hay, D. F., Payne, A., & Chadwick, A. (2004). Peer relations inchildhood. Journal of Child Psychology and Psychiatry, 45(1),84–108. DOI: 10.1046/j.0021-9630.2003.00308.x

Holmes, T. H., & Rahe, R. H. (1967). Social readjustment ratingscale. Journal of Psychosomatic Research, 11(2), 213–218.

Ilmarinen, V. J., Vainikainen, M. P., Verkasalo, M., & Lönnqvist, J.E. (2015). Why are extraverts more popular? Oral fluency medi-ates the effect of extraversion on popularity in middle childhood.European Journal of Personality, 29, 138–151.

Jeronimus, B. F., Ormel, J., Aleman, A., Penninx, B. W. J. H., &Riese, H. (2013). Negative and positive life events predict smallbut long-term change in neuroticism. Psychological Medicine,43(11), 2403–2415. DOI: 10.1017/S0033291713000159

Jeronimus, B. F., Riese, H., Oldehinkel, A. J., & Ormel, J. (2015,submitted). Why does frustration predict psychopathology? Mul-tiple prospective pathways over adolescence: A TRAILS study.

Jeronimus, B. F., Riese, H., Sanderman, R., & Ormel, J. (2014).Mutual reinforcement between neuroticism and life stressors: Afive-wave, sixteen-year study to test reciprocal causation. Journalof Personality and Social Psychology, 107(4), 751.

Kandler, C., Bleidorn, W., Riemann, R., Angleitner, A., &Spinath, F. M. (2012). Life events as environmental states andgenetic traits and the role of personality: A longitudinal twinstudy. Behavior Genetics, 42(1), 57–72. DOI: 10.1007/s10519-011-9491-0

Kendler, K. S., & Baker, J. H. (2007). Genetic influences on mea-sures of the environment: A systematic review. PsychologicalMedicine, 37(5), 615–626. DOI: 10.1017/S0033291706009524


Kotov, R., Gamez, W., Schmidt, F., & Watson, D. (2010). Linking“big” personality traits to anxiety, depressive, and substance usedisorders: A meta-analysis. Psychological Bulletin, 136(5),768–821. DOI: 10.1037/a0020327

Küfner, A. C. P., Nestler, S., & Back, M. D. (2013). The two path-ways to being an (un-) popular narcissist. Journal of Personality,81(2), 184–195. DOI: 10.1111/j.1467-6494.2012.00795.x

Laceulle, O.M., van Aken, M.A.G., Ormel, J., & Nederhof, E.(2014). Stress-sensitivity and reciprocal associations betweenstressful events and adolescent temperament. Personality and In-dividual Differences, DOI: 10.1016/j.paid.2014.12.009.

Laceulle, O. M., Nederhof, E., Karreman, A., Ormel, J., & vanAken, M. A. G. (2012). Stressful events and temperament changeduring early and middle adolescence: The TRAILS study. Euro-pean Journal of Personality, 26(3), 276–284. DOI: 10.1002/per.832

Larsen, R. J., & Buss, D. M. (2013). Personality psychology: Do-mains of knowledge about human nature. New York: McGraw-Hill.

Laursen, B., & Williams, V. A. (1997). Perceptions of interdepen-dence and closeness in family and peer relationships among ado-lescents with and without romantic partners. In W. A. Collins(Ed.), Romantic relationships in adolescence: Developmentalperspectives. (pp. 3–20). San Francisco, CA US: Jossey-Bass.

Lempers, J. D., & Clark-Lempers, D. S. (1992). Young, middle, andlate adolescents’ comparisons of the functional importance of fivesignificant relationships. Journal of Youth and Adolescence, 21(1),53–96. DOI: 10.1007/BF01536983

Li, X., Wong, W., Lamoureux, E. L., & Wong, T. Y. (2012). Arelinear regression techniques appropriate for analysis when the de-pendent (outcome) variable is not normally distributed? Investi-gative Ophthalmology & Visual Science, 53(6), 3082–3083.DOI: 10.1167/iovs.12-9967

Linder, J. R., Crick, N. R., & Collins, W. A. (2002). Relational ag-gression and victimization in young adults’ romantic relation-ships: Associations with perceptions of parent, peer, andromantic relationship quality. Social Development, 11(1),69–86. DOI: 10.1111/1467-9507.00187

London, B., Downey, G., Bonica, C., & Paltin, I. (2007). Socialcauses and consequences of rejection sensitivity. Journal of Re-search on Adolescence, 17(3), 481–506. DOI: 10.1111/j.1532-7795.2007.00531.x

Lüdtke, O., Roberts, B. W., Trautwein, U., & Nagy, G. (2011). Arandom walk down university avenue: Life paths, life events,and personality trait change at the transition to university life.Journal of Personality and Social Psychology, 101(3),620–637. DOI: 10.1037/a0023743

Luhmann, M., Orth, U., Specht, J., Kandler, C., & Lucas, R. E.(2014). Studying changes in life circumstances and personality:It’s about time. European Journal of Personality, 28(3),256–266. DOI: 10.1002/per.1951

Middeldorp, C. M., Cath, D. C., Beem, A. L., Willemsen, G., &Boomsma, D. I. (2008). Life events, anxious depression and per-sonality: A prospective and genetic study. Psychological Medi-cine, 38(11), 1557–1565. DOI: 10.1017/S0033291708002985

Moran, L. R., Lengua, L. J., & Zalewski, M. (2013). The interactionbetween negative emotionality and effortful control in earlysocial-emotional development. Social Development, 22(2),340–362. DOI: 10.1111/sode.12025

Muris, P., & Meesters, C. (2009). Reactive and regulative tempera-ment in youths: Psychometric evaluation of the early adolescenttemperament questionnaire-revised. Journal of Psychopathologyand Behavioral Assessment, 31(1), 7–19. DOI: 10.1007/s10862-008-9089-x

Nettle, D. (2007). Personality: What makes you the way you are.New York: Oxford University Press.

Neyer, F. J., & Lehnart, J. (2007). Relationships matter in personal-ity development: Evidence from an 8-year longitudinal studyacross young adulthood. Journal of Personality, 75(3),535–568. DOI: 10.1111/j.1467-6494.2007.00448.x

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per


Neyer, F. J., & Asendorpf, J. B. (2001). Personality–relationshiptransaction in young adulthood. Journal of Personality and So-cial Psychology, 81(6), 1190–1204. DOI: 10.1037/0022-3514.81.6.1190

Neyer, F. J., Mund, M., Zimmermann, J., & Wrzus, C. (2013). Per-sonality-relationship transactions revisited. Journal of Personal-ity, early viewDOI: 10.1111/jopy.12063

Nieboer, A., Lindenberg, S., Boomsma, A., & Van Bruggen, A.(2005). Dimensions of well-being and their measurement: TheSPF-IL scale. Social Indicators Research, 73(3), 313–353. DOI:10.1007/s11205-004-0988-2

Ormel, J., Lindenberg, S., Steverink, N., & Vonkorff, M. (1997).Quality of life and social production functions: A frameworkfor understanding health effects. Social Science & Medicine,45(7), 1051–1063. DOI: 10.1016/S0277-9536(97)00032-4

Peterson, R. A., & Brown, S. P. (2005). On the use of beta coeffi-cients in meta-analysis. Journal of Applied Psychology, 90(1),175–181. DOI: 10.1037/0021-9010.90.1.175

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resamplingstrategies for assessing and comparing indirect effects in multiplemediator models. Behavior Research Methods, 40(3), 879–891.DOI: 10.3758/BRM.40.3.879

Putnam, S. P., Ellis, S. K., & Rothbart, M. K. (2001). The structureof temperament from infancy through adolescence. In A. Eliasz,& A. Engleneiter (Eds.), Advances in research on temperament.(pp. 165–182). Berlin: Pabst Scientist Publicer.

Riese, H., Snieder, H., Jeronimus, B. F., Korhonen, T., Rose, R. J.,Kaprio, J., et al. 2014). Timing effects stressful life events on sta-bility and change of neuroticism scores. European Journal ofPersonality, (2), 193–200. DOI: 10.1002/per.1929

Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Goldberg,L. R. (2007). The power of personality the comparative validityof personality traits, socioeconomic status, and cognitive abilityfor predicting important life outcomes. Perspectives on Psycho-logical Science, 2(4), 313–345. DOI: 10.1111/j.1745-6916.2007.00047.x

Roberts, B. W., Caspi, A., & Moffitt, T. E. (2003). Work experi-ences and personality development in young adulthood. Journalof Personality and Social Psychology, 84(3), 582–593. DOI:10.1037/0022-3514.84.3.582

Robins, R. W., Caspi, A., & Moffitt, T. E. (2002). It’s not just whoyou’re with, it’s who you are: Personality and relationship expe-riences across multiple relationships. Journal of Personality,70(6), 925–964.

Rothbart, M. K. (2011). Becoming who we are. New York, NY:Guilford Press.

Rothbart, M. K., & Bates, J. E. (2007). Temperament. In DamonW., Lerner R. & Eisenberg N. (Eds.), Handbook of child psychol-ogy (6th edn., pp. 99–166). New York: John Wiley & Sons, Inc.DOI: 10.1002/9780470147658.chpsy0303

Sadikaj, G., Moskowitz, D., & Zuroff, D. (2015). Felt security indaily interactions as a mediator of the effect of attachment on sat-isfaction with a romantic partner. European Journal of Personality,29, 187–200.


Scarr, S., & McCartney, K. (1983). How people make their own en-vironments: A theory of genotype→ environment effects. ChildDevelopment, 54(2), 424–435.

Selfhout, M., Burk, W., Branje, S., Denissen, J., Van Aken, M., &Meeus, W. (2010). Emerging late adolescent friendship networksand Big Five personality traits: A social network approach. Jour-nal of Personality, 78(2), 509–538. DOI: 10.1111/j.1467-6494.2010.00625.x

Shanahan, M. J., & Flaherty, B. P. (2001). Dynamic patterns of timeuse in adolescence. Child Development, 72(2), 385–401. DOI:10.1111/1467-8624.00285

Simon, V. A., Aikins, J. W., & Prinstein, M. J. (2008). Romanticpartner selection and socialization during early adolescence.Child Development, 79(6), 1676–1692. DOI: 10.1111/j.1467-8624.2008.01218.x

Soto, C. J., John, O. P., Gosling, S. D., & Potter, J. (2008). The de-velopmental psychometrics of Big Five self-reports: Acquies-cence, factor structure, coherence, and differentiation from ages10 to 20. Journal of Personality and Social Psychology, 94(4),718–737. DOI: 10.1037/0022-3514.94.4.718

Specht, J., Egloff, B., & Schmukle, S. C. (2011). Stability andchange of personality across the life course: The impact of ageand major life events on mean-level and rank-order stability ofthe Big Five. Journal of Personality and Social Psychology,101(4), 862–882.

Sroufe, L. A. (1990). An organizational perspective on the self. InD. Cicchetti, & M. Beeghly (Eds.), The self in transition: Infancyto childhood. The John D. and Catherine T. MacArthur Founda-tion series on mental health and development(pp. 281). Chicago:University of Chicago Press.

Vinkhuyzen, A. A. E., Sluis, S. v. d., Geus, E. J. C. d., Boomsma,D. I., & Posthuma, D. (2010). Genetic influences on ‘environ-mental’ factors. Genes, Brain and Behavior, 9(3), 276–287.DOI: 10.1111/j.1601-183X.2009.00554.x

Windle, M. (1992). Temperament and social support in adoles-cence: Interrelations with depressive symptoms and delinquentbehaviors. Journal of Youth and Adolescence, 21(1), 1–21.

Wood, D., & Roberts, B. W. (2006a). Cross-sectional and longitudi-nal tests of the personality and role identity structural model(PRISM). Journal of Personality, 74(3), 779–809. DOI:10.1111/j.1467-6494.2006.00392.x

Wood, D., & Roberts, B. W. (2006b). The effect of age and role in-formation on expectations for Big Five personality traits. Person-ality and Social Psychology Bulletin, 32(11), 1482–1496. DOI:10.1177/0146167206291008

Wrzus, C., Hänel, M., Wagner, J., & Neyer, F. J. (2013). Social net-work changes and life events across the life span: A meta-analysis. Psychological Bulletin, 139(1), 53–80. DOI: 10.1037/a0028601

Zimmer-Gembeck, M. J. (2002). The development of romantic rela-tionships and adaptations in the system of peer relationships.Journal of Adolescent Health, 31(6, Supplement), 216–225.DOI: 10.1016/S1054-139X(02)00504-9

Eur. J. Pers. 29: 125–137 (2015)

DOI: 10.1002/per

why not everyone gets their fair share of stress: adolescent's perceived relationship affection...

Documents