SEARCH

SEARCH BY CITATION

Keywords:

  • Temperament;
  • attachment;
  • behavior problems;
  • social anxiety;
  • longitudinal

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

Results from the Uppsala Longitudinal Study (ULS), which started in 1985, are reported in two sections. The first section gives a summary of longitudinal data from infancy to middle childhood (age 9 years; = 96) concerning predictions of social functioning aspects from the theoretical perspectives of temperament, attachment, and health psychology (social factors). The second section presents the first results emanating from a follow-up when participants were 21 years old (= 85). The developmental roots of social anxiety symptoms were studied from the same perspectives as above, although with a special focus on the predictive power of the temperament trait of shyness/inhibition. Results for middle childhood outcomes showed that temperament characteristics were relevant for most outcomes, whereas the contribution of attachment was most convincingly shown in relation to social competence and personality. Social factors were found to have moderating functions, but direct effects were also shown, the most interesting perhaps being positive effects of non-parental day care. Results from the 21-year data confirmed the expected predictive relation from shyness/inhibition to symptoms of social anxiety and further showed this relation to be specific; the relation to symptoms of depression did not survive control for social anxiety, although the opposite was true. The broad analysis of predictor associations with social anxiety, showing the relevance of other temperament factors as well as interactive effects, again attested to the need for multi-faceted models to analyze developmental trajectories.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

The contribution of early-appearing individual differences to later personality and adjustment was the general question motivating the initiation of the Uppsala Longitudinal Study (ULS) in 1985. Our thinking about influences on social and emotional development was then, and continues to be, based on three sets of factors, the individual, the immediate family, and social factors impinging on individual and family functioning (cf. Bronfenbrenner, 1986). Theoretical and methodological guidelines were found in theory and research in the temperament (individual characteristics), attachment (child-parent relationship) and health psychology (social factors) fields. The study began when the participants were 6 weeks old, and we recently completed a data collection when they were 21 years of age.

A short theoretical and methodological introduction to the project will be given, followed by two empirical sections. First, we will present an overview of results pertaining to the predictive power of temperament, attachment and selected social factors for socio-emotional adjustment during childhood, that is up to 8–9 years. Most of these results have been published, but some new data will be provided to complete the picture. The second empirical part presents new results from the data collection at 21 years with an inquiry focusing on the importance of early temperament traits in combination with attachment and social factors for development of social anxiety symptoms in young adulthood.

Temperament

The majority of temperament researchers agree that temperament refers to biologically rooted individual characteristics, observed in overt behavior in early life, and showing some stability over situations and time (e.g., Bates, 1989). However, the behavioral expressions of temperament are formable during development, for instance in processes that emanate from environmental responses to the individual’s behavior or steer a person to seek environmental stimulation, dependent on his/her particular temperament characteristics (Caspi, Roberts & Shiner, 2005). Temperament has generally been conceptualized as a set of traits, rationally or theoretically derived. Most often temperament structures have been formed via factor analyses of behavioral expressions.

Attachment

Attachment is defined as a bond that develops between the infant and his/her caregiver, over the first year of life (Bowlby, 1969/1978); the quality of that bond may be assessed in the Strange Situation procedure, developed by Ainsworth (Ainsworth, Blehar, Waters & Wall, 1978). Three basic patterns of attachment have been derived from and replicated in observations of normal samples of infants in numerous studies (van IJzendoorn & Sagi-Schwartz, 2008).These patterns are theoretically related to antecedent caregiver-infant interactions in particular ways. A secure pattern is hypothesized to result from interactions with a caregiver who is “sensitively responsive” to her infant’s signals, reading them accurately and responding appropriately. A less responsive, slightly rejecting caregiver may have a child who avoids contact when stressed, with minimal emotional expression: an insecure-avoidant pattern. The insecure-ambivalent infant is inconsistent in his/her attempts to get comfort from the caregiver, who is assumed to have been over-involved, intrusive, or inconsistent in his/her responses to the child. Based on such experiences in infant-caregiver interactions, internal working models develop that regulate views on self and others and influence socio-emotional development throughout life (e.g., Bretherton & Munholland, 2008).

Social factors

In the present context, three social factors in and surrounding the family are relevant. An important family support factor in Sweden is the availability of non-parental day care. In international comparisons Swedish day care appears to be of generally high quality due to factors such as teacher education, educational programs, localities, and equipment (Hwang & Broberg, 1992). Previous studies of non-parental care in Sweden have shown either no differences between home-reared children and children in non-parental care or positive effects of such care regarding socio-emotional development (Andersson, 1992; Wessels, Lamb, Hwang & Broberg, 1997).

Stress experiences are presumably caused by negative events (loss of close relationships, loss of employment, accidents, etc.) and daily hassles (minor irritating events) with adverse effects on subsequent functioning. Infants and young children are likely to experience stress in indirect ways, mediated by parenting and functioning of the family system; older children may also experience stress directly. There is correlational evidence of associations between perceived parenting stress and child adjustment problems; the direction of effects is, however, often unclear (Crnic & Low, 2002).

Socioeconomic status (SES) can be defined based on a variety of social and economic indices, such as education, occupation, income, and values (Reber, 1985). In the ULS we use parental education as a proxy for socioeconomic status, because most other indices are connected to educational level.

Socio-emotional adjustment

By socio-emotional adjustment we mean the relation an individual establishes with respect to the environment, and as such it has positive and negative, social and emotional aspects (cf. Reber, 1985). In the ULS, important positive aspects pertain to the construct of social competence, that is, the effective functioning in social contexts, studied as prosocial orientation, social initiative, popularity, and other positive social behaviors (Cavell, 1992; Rydell, Hagekull & Bohlin, 1997). On the negative side, a common description of problems is made in terms of the broadband dimensions externalizing and internalizing behaviors, with externalizing denoting aggressive, hyperactive, and concentration problems, and internalizing referring to anxious, depressed, fearful behaviors, and sometimes also psychosomatic complaints, including feeding/eating problems (Achenbach & Edelbrock, 1978).

Another perspective on socio-emotional functioning was captured in the Five Factor Model of personality (FFM or the Big Five; Halverson, Kohnstamm & Martin, 1994). The FFM is similar to temperament models, conceptually and methodologically and includes the “superfactors” Extraversion (vs. Introversion), Agreeableness (vs. Hostility), Conscientiousness (or Will to Achieve), Neuroticism (vs. Emotional Stability), and Openness/Intellect. It has been said that temperament is “the hard iceball around which the softer snowball of personality accumulates developmentally” (Graziano, Jensen-Campbell & Sullivan-Logan, 1998, p. 1267). Thus, personality is seen as resulting from additive and interacting temperament and environmental influences. Overlaps between temperament traits and the Big Five have been explored both conceptually and empirically (e.g., Ahadi & Rothbart, 1994; Hagekull, 1994), but few long-term prospective studies have investigated links between early childhood temperament and personality (see reviews by Caspi & Silva, 1995; Rothbart & Bates, 2006).

Method

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

A summary of various methodological aspects is found below. The reader is referred to the original publications for details about participants, procedures, and measures (scale content, response alternatives, psychometric information, etc.).

Participants and procedures

Mothers of all infants born during 11 specified weeks in 1985 at the Academic Hospital in Uppsala were approached. Of these, 62% (= 123) of the mothers agreed to participate in the study; from infant age 10 months, 104 fathers were included (due to increased funding). At age 9 years, 96 families were still in the study. The study site is a university town and its surroundings, and parents were generally highly educated, although parents with no or only a 2-year education beyond the obligatory 9 school years also participated.

The first home visit was conducted at infant age 6 weeks. The contacts were intensive during the first four years (13 data waves). Thereafter, three data collections were undertaken between ages 8 and 9 years. Information was gathered via video recorded lab visits, paper and pencil recorded observations in homes, preschool and school settings, questionnaires, and interviews with parents, preschool and school teachers, and the school children themselves.

At the age of 21 years, 112 participants were invited, 99 persons responded, and 85 agreed to take part. Main attrition causes were lack of time and living abroad. Mean age of the participants was 21 years, 3 months (SD = 5 months). Their occupations were as students (university 36%, high school level 9%, professional program 6%), and in the work force (full-time 27%, part-time 27%; students often had part-time jobs); 6% were in search of work.

A questionnaire containing among other instruments social anxiety measures was sent to the participants 2 weeks before their scheduled visit to the department for interviews and other data collection procedures. Depression questions were responded to during the visit.

Measures

Temperament.  The factor analytically derived temperament traits, intensity/activity, regularity, approach-withdrawal, sensory sensitivity, attentiveness, and manageability (reflecting irritability/negative emotionality), were measured by the Toddler Behavior (TBQ) parent questionnaire for children aged 10 to 15 months (Hagekull & Bohlin, 1981). In late infancy, toddler, preschool, and school ages, negative emotionality, activity, sociability, and shyness/inhibition were studied with parent ratings in the EAS scales, which are based on the criteria of heredity, stability, and presence in early life (Buss & Plomin, 1984).

Attachment.  At 15 months, the standard Strange Situation procedure (Ainsworth et al., 1978) was performed in the laboratory, and infant attachment to mother was classified as secure, insecure avoidant, or insecure ambivalent. We also studied Ainsworth’s construct maternal sensitivity. At 48 months, maternal sensitivity in directing her child was rated from video recordings of a meal in the home (Bohlin & Hagekull, 2000).

Social factors. Non-parental day care was measured as mean number of hours per week during the first 48 months of life. Day care quality was measured in interviews and observations at child age 28 months (Hagekull & Bohlin, 1995). In predictions of middle childhood outcomes child stress experiences in connection to negative life events during age 6–7 years were reported by mothers (Hagekull & Bohlin, 1998). In the 21-year analyses maternal stress experiences were captured in an aggregate of negative life events, daily hassles, and stress due to work load during the first four years. The SES proxy, parental education, was coded in five levels with academic degree as the highest and 9 years of schooling as the lowest level (Hagekull & Bohlin, 1995).

Socio-emotional adjustment.  Parent and teacher rating scales were used to measure social competence in middle childhood, that is, at age 8–9 years (prosocial orientation, social initiative, popularity; Rydell et al., 1997) and problem behaviors at 4 and 8–9 years (externalizing, internalizing, psychosomatic; Rutter, Tizard & Whitmore, 1970). The FFM factors were studied by parent and teacher reports at 8–9 years (Lanthier, 1993).

In adulthood we used the Social Interaction Anxiety Scale (SIAS) and the Social Phobia Scale (SPS), which were developed as twin self-report instruments to distinguish between difficulties mixing or interacting with others (fears of being inarticulate, boring, or ignored), and the more general phobic aspects speech anxiety and scrutiny fears (when eating, drinking, working etc.) in the presence of others (Mattick & Clarke, 1998). Each instrument consists of 20 items with five-point response scales with verbal descriptors (α  =  0.85 for both SIAS and SPS). The Depression scale with 13 items (five-point response scales; α  =  0.87) was taken from a Swedish version of the self-report Symptom Check List (SCL-90; Fridell, Cesarec, Johansson & Thorsen, 2002).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

Prediction of social adjustment during childhood

Based on ULS data, a number of topics have been investigated. Phenomena such as type A behavior, hyperactivity, infant social referencing, maternal adjustment, parents’ perceived control, as well as methodological issues have been dealt with, as seen in the ULS publication list (http://www.psyk.uu.se/hemsidor/devpsy/Webbplats/Publications.html). The present review, however, is focused on the central predictors (temperament, attachment, selected social factors) and the most important child outcomes. The majority of the childhood results presented below have been published, but some unpublished correlations appear in Table 1. Results regarding the FFM predictions will be presented in the text only.

Table 1. Product moment correlations between temperament, attachment, social factors and outcomes (ratings only) at 4, 8–9, and 21 years
 Externalizing behaviorInternalizing behaviorSocial competenceSocial anxiety 21 yrsDepression
4 yrs8–9 yrs4 yrs8–9 yrs8–9 yrsSIASSPS21 yrs
  1. Note: SIAS = Social Interaction Anxiety Scale, SPS = Social Phobia Scale.

  2. a 10–15 months, no other infant temperament scale correlated significantly with these outcomes.

  3. b 20–48 months.

  4. c 15 months.

  5. d 48 months.

  6. e 2 years for 4-year outcomes, 6–7 years for later outcomes.

  7. f 0–48 months.

  8. g Controlling for parents’ SES level, the coefficients were –0.32, –0.25, and –0.49.

  9. ns p  0.10; + < 0.10; *< 0.05; **< 0.01; ***< 0.001; = 105 (4-year data); = 96 (8–9-year data); = 85 (21-year data).

Temperament
Manageabilitya–0.31**–0.21*nsnsnsnsnsns
Negative emotionalityb0.31***0.19+0.32**0.27**–0.29**nsnsns
Activityb0.45***0.26**–0.26**–0.17+0.27**–0.37***–0.24**–0.20+
Sociabilitybnsns–0.21*–0.23*0.31**–0.29**nsns
Shynessb–0.20*ns0.32***0.30**–0.39***0.36***0.42***0.22*
Attachment
Infant securitycnsnsnsns0.26**nsnsns
Maternal sensitivityd–0.33***–0.28**nsnsnsnsnsns
Social factors
Child stressensns0.22*0.39***nsnsns0.20+
Nonparental carefnsnsns0.29***0.25*–0.33**g–0.26*g–0.47***g
SES–0.32***–0.28**nsnsnsnsnsns

Main effects of temperament

The prospective analyses of temperament in relation to behavior problems demonstrated a role of infant/preschool temperament in the development of externalizing behavior problems in both a short-term and long-term perspective. Low manageability/high negative emotionality in infancy/preschool age predicted high degrees of externalizing behaviors at 4 and 8–9 years (Bohlin & Hagekull, 2000) and so did high activity in preschool children (Table 1, left).

Links from early temperament to later internalizing problems included infant and preschool negative emotionality showing positive correlations with headache and stomachache complaints at school age, and also positive links between negative emotionality and picky eating (Hagekull & Bohlin, 2004). No other significant infant links to internalizing problems or social competence were found (Table 1). Preschool negative emotionality carried positive relations, and preschool activity and sociability showed negative associations, with internalizing problems at 4 and 8–9 years. Predicted links between preschool shyness/inhibition and internalizing problems were empirically established at both ages. Further, preschool shyness/inhibition and negative emotionality were negatively related to school age social competence with peers (Bohlin, Hagekull & Andersson, 2005), and sociability and activity levels showed significant positive correlations with social competence at school age.

Predictions from infant (20 months) and preschool (28–48 months) temperament to FFM personality traits at age 8–9 years were tested (Hagekull & Bohlin, 1998, 2003). As hypothesized, high extraversion was significantly predicted by the three infancy EAS traits high activity, high sociability, and low negative emotionality. No other personality trait had roots in infancy. The preschool predictions were stronger with the highest amount of variance explained in extraversion by activity, sociability, and shyness/inhibition (reversed). Agreeableness was negatively predicted by activity and positively by shyness/inhibition, high conscientiousness had significant roots in high shyness/inhibition. Negative emotionality correlated positively with neuroticism. Finally, openness had the same temperament bases as extraversion.

Main effects of attachment

Attachment theory has clear implications for social functioning. A secure attachment should foster experiences of trust in oneself and others, resulting in good social skills and low problem load. In tests of such hypotheses we found significant predictions to 4-year externalizing problems only for insecure avoidant children, who showed significantly more problems than secure children (Bohlin & Hagekull, 2000). Predictions to middle childhood ratings of externalizing and internalizing problems were non-significant.

Infant attachment security contributed positively to school age social competence, and this effect was found to be independent of temperament in two studies (Bohlin, Hagekull, Rydell, Berlin & Andersson, 2004; Bohlin et al., 2005). The results reported by Bohlin, Hagekull, and Rydell (2000) broaden the picture; secure children were found to be more socially active, positive, popular, and tended to report less social anxiety. Infant security was also found to be of great importance for FFM personality in middle childhood: extraversion and openness were positively, and neuroticism negatively related to attachment security at 15 months (Hagekull & Bohlin, 2003). Furthermore, maternal sensitivity at 48 months was positively associated with the two remaining FFM traits, agreeableness and conscientiousness (Hagekull & Bohlin, 1998).

Main effects of social factors

Ongoing debates concerning effects of non-parental day care have made this issue urgent to investigate. Only positive effects appeared in the ULS; few problems (internalizing, externalizing), low neuroticism, high social competence and extraversion characterized children who had spent many hours per week in non-parental care. High quality care was also found to be associated with few problems (Bohlin & Hagekull, 2000; Bohlin et al., 2004; Hagekull & Bohlin, 1995, 1998).

Low SES was connected to a high degree of externalizing problems and neuroticism, and also to low agreeableness and conscientiousness (Bohlin & Hagekull, 2000; Hagekull & Bohlin, 1995, 1998). Child stress was positively related to internalizing problems at both 4 and 8–9 years (Table 1). Both maternal and child stress experiences were positively related to later child neuroticism (Hagekull & Bohlin, 1998).

Moderator models: Interactions between temperament, attachment, and social factors

Can certain temperament traits in combination with an insecure relationship increase the risk of negative outcomes? Will attachment security or non-parental day care experience buffer effects of child negative temperament characteristics in the long run? Can stress reactions combined with particular temperament dispositions be detrimental for later development? The theoretically most relevant moderator models have concerned such interactions. Regarding middle childhood outcomes, secure inhibited children developed significantly better social competences than did insecure inhibited children, whereas security was not important for non-inhibited children (Bohlin et al., 2005). High emotionality preschoolers who were insecure as infants tended to develop more picky eating at 8–9 years; emotionality did not matter for securely attached children (Hagekull & Bohlin, 2004).

Concerning social factors, an association between high maternal sensitivity and low aggressiveness in 4-year-olds held only for children who had experienced high quality non-parental care (Bohlin & Hagekull, 2000). Other interactions showed that high quality care was most important for reducing aggressiveness in low SES children and for reduction of boys’ internalizing problems (Hagekull & Bohlin, 1995). Thus, good non-parental care had compensating/buffering effects. We also found that negative events constituted a risk for more neuroticism only in combination with low sociability. Children with high sociability were not affected by such experiences (Hagekull & Bohlin, 1998).

Prediction of social anxiety symptoms at age 21 years

Data analyses for the 21-year data have just begun. Here we will present a first set of results, which concern precursors of social anxiety symptoms, with a special focus on links from childhood shyness/inhibition. Shyness/inhibition and social anxiety are similarly conceptualized, in that both phenomena pertain to fear and anxiety experienced in performance and other social situations, although social anxiety is a more debilitating condition (cf. American Psychological Association, 2000). The ULS analyses of middle childhood data had shown that high shyness/inhibition predicted poor social competence as well as provided evidence indicating that a secure attachment relation to mother could mitigate this effect, results that motivated interest in predictive relations to 21-year data.

In the temperament field, shyness/inhibition has been pointed out specifically as a precursor of social anxiety (e.g., Schwartz, Snidman & Kagan, 1999). Degnan and Fox (2007) report that this hypothesis has been supported in studies of both adolescents and adults, although no study on a normal sample seems to have had the ULS kind of prospective, longitudinal design, spanning from infancy to adulthood. Degnan and Fox (2007) also point out that in spite of the significant stability of shyness/inhibition and its predictive value for later anxiety disorders, there is also evidence for discontinuity, which suggests the possibility of a resilience process, that is, “achieving positive adaptation despite experiencing significant threat, adversity, or risk” (p. 733). Thus, influences from, for instance, attachment processes and social factors could moderate the temperamental tendencies, thus constituting protective factors acting against development of high social anxiety. In the present analyses, our broad approach with temperament, attachment, and social factors was applied to achieve a deeper understanding of social anxiety development, using linear as well as moderator models. In line with this general approach it was of interest to apply a broad perspective also in elucidating temperamental influences in the trajectory to social anxiety. Therefore, the other EAS traits, negative emotionality, activity, and sociability (Buss & Plomin, 1984) were explored in addition to shyness/inhibition. Previous studies point to high negative emotionality as a nonspecific factor predictive of both social anxiety and depression (Hyde, Mezulis & Abrahamson, 2008). However, the distinction between negative emotionality in reactions to novelty (shyness/inhibition) and general negative reactions has frequently not been upheld. As a test of the specificity of social anxiety predictors, we studied independent contributions from emotionality and shyness, as well as from other temperament traits in predicting depression and social anxiety symptoms. Thus, the research questions concerned prediction of social anxiety symptoms in adulthood from preschool temperament and experiences, and whether the precursors of social anxiety symptoms could be distinguished from those of depression symptoms, which often include features of social anxiety (American Psychiatric Association, 2000).

Shyness/inhibition and attachment security were expected to contribute to long-term social adjustment via direct links. We also expected moderating influences on effects of high shyness/inhibition; a sensitive mother should know how to handle her child’s social withdrawal tendencies (cf. Degnan & Fox, 2007). Further, child care arrangements are potentially important for social development (cf. Degnan & Fox, 2007). Non-parental day care in Sweden has previously shown its capacity to contribute to positive social adjustment in both direct and moderator models, and may be expected to lessen social reticence also in a long-term perspective. Finally, child stress experiences in connection with negative life events could be a risk factor for negative development, both directly and in interaction with temperament (Hagekull & Bohlin, 1998).

Childhood predictors of social anxiety and depression symptoms

Note that in using the terms social anxiety and depression below, we are referring to symptoms, not diagnosed disorders. As seen in Table 1 (right), early negative emotionality predicted neither social anxiety nor depression, whereas shyness/inhibition showed substantial positive correlations with social anxiety measures and a low but significant positive relation to depression. Activity was at least to some extent negatively associated with all three outcomes, and sociability showed a negative significant correlation with social interaction anxiety only.

Attachment security and maternal sensitivity had no direct links to the outcomes. A marginal positive relation was found between child stress experiences and depression in adulthood, whereas non-parental care yielded significant negative correlations with all outcomes, most notably with depression.

Independent and interaction effects on social anxiety and depression

Regression analyses were executed to study independent contributions from the four temperament traits. Only activity carried a unique (negative) relation with social interaction anxiety, = –2.51, < 0.02, and shyness/inhibition was the sole independent contributor to general social anxiety, = 3.10, < 0.01 (positive relation). No trait was uniquely predictive of depression. When excluding sociability (because of its correlation of –0.38 with shyness) both activity, < 0.01, and shyness/inhibition, < 0.07, predicted SIAS scores; no changes were found for SPS and depression scores.

In a second set of regression analyses, potentially moderating effects of maternal sensitivity, child stress experiences, and non-parental care on the effects of temperament traits were examined. (Attachment security was not included as it is problematic to regard a relationship studied at 15 months as a moderator of effects of later studied temperament traits.) Each moderator was examined together with each trait in a regression equation for each of the three outcomes. Six significant interactions were found, one each related to the two social anxiety aspects and four related to depression (Fig. 1). The interactions showed that in the development of both general and social interaction anxiety, effects of high shyness/inhibition could be reduced by high maternal sensitivity at age 4 years, =  –2.73, < 0.01, and =  –2.15, < 0.05, respectively. Concerning depression, child stress experiences constituted a risk factor in connection with three temperament characteristics, high negative emotionality, = 2.37, < 0.05, low activity, =  –2.64, < 0.01, and low sociability, =  –2.49, < 0.05. Low childhood sociability in combination with low maternal sensitivity was also connected to a high degree of depression symptoms, = 2.34, < 0.05.

image

Figure 1. Top left: Interaction effect of childhood shyness/inhibition and maternal sensitivity on social interaction anxiety symptoms at age 21 years. Top right: Interaction effect of childhood shyness/inhibition and maternal sensitivity on general social anxiety symptoms at age 21 years. Middle left: Interaction effect of childhood negative emotionality and child stress experiences on depression symptoms at age 21 years. Middle right: Interaction effect of childhood activity and child stress experiences on depression symptoms at age 21 years. Bottom left: Interaction effect of childhood sociability and child stress experiences on depression symptoms at age 21 years. Bottom right: Interaction effect of childhood sociability and maternal sensitivity on depression symptoms at age 21 years.

Download figure to PowerPoint

Specificity in predictions of social anxiety

In a final set of regressions, the specificity of shyness/inhibition as a predictor of social anxiety in comparison to predictions of depression was evaluated. With depression controlled for, shyness/inhibition remained a significant predictor of social interaction as well as general social anxiety, t = 2.84, p < 0.01 (SIAS scores), and = 3.52, < 0.001 (SPS scores). When SIAS or SPS scores were controlled for in predictions of depression, the shyness/inhibition variable became non-significant in both equations, = 0.85, ns, and =  –0.07, ns. Thus, shyness/inhibition was uniquely predictive of social anxiety, whereas the connection between shyness/inhibition and depression was dependent on co-occurring social anxiety problems.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

The ULS from infancy to middle childhood

The ULS findings concerning the importance of temperament, attachment, and social factors in the development of socio-emotional functioning from infancy to early school age may be briefly summarized by saying that temperament characteristics were predictive of problem behaviors as well as personality and social competence, whereas security of attachment was confined to predicting social competence and personality. Although social factors were mainly considered because of their possible moderating function, there were also direct effects in relation to problem behaviors and social competence, the most interesting perhaps being positive effects of non-parental day care.

Temperament.  Infant temperament had few links to later outcomes, the association of poor manageability/high irritability with later externalizing and psychosomatic problems being the only significant ones. However, temperament dispositions as displayed during toddlerhood and the early preschool period showed several longitudinal predictive relations; shyness/inhibition, emotionality, sociability, and activity all predicted problem behaviors in both 4- and 8–9-year-olds, as well as social competence and personality factors in the 8–9-year-olds. The associations of shyness/inhibition to internalizing problems, low social competence and low extraversion were in line with expectations in that this temperament trait has been shown to be stable in the ULS, and in other studies shown to be a precursor of social anxiety in childhood (Degnan & Fox, 2007). Some support was also obtained for the less well-studied protective effect of shyness/inhibition in relation to externalizing problems (e.g., Rothbart & Bates, 2006, p. 138), although this was only found at 4 years. More surprising were the findings that high shyness/inhibition was associated with 8–9 year low openness and low social competence. Together our results support and extend those previously reported concerning negative developmental effects for socially withdrawn and fearful children (Rubin, 1993; see also Degnan & Fox, 2007).

Negative emotionality was related in expected ways to problem behavior, high emotionality foreboded high levels of both internalizing and externalizing behaviors. This is interesting as it suggests that this temperament trait may be the source behind the co-occurrence between the two types of problem behaviors (e.g., Cunningham & Boyle, 2002). The relation to later rated neuroticism/emotional instability is in accordance with expectations and can perhaps be interpreted in terms of continuity. More interesting from a developmental perspective is the finding that social competence was lower for children who had evidenced high levels of negative emotionality several years earlier. Whether or not this reflects co-occurrence between emotionality and problem behaviors or constitutes a genuine effect is worth further study.

The longitudinal relations for activity were interesting in that high activity was found to be a precursor of negative as well as positive outcomes; high activity was associated with high externalizing, but low internalizing problem behaviors, with high social competence, extraversion and openness, but low agreeableness. In the temperament literature, high activity was originally related to negative outcomes in early childhood (Thomas & Chess, 1977). Our positive outcomes may be explained in several ways: first, these findings could be an extension of infancy data showing that activity is linked to expressions of both negative and positive affects (Rothbart & Bates, 2006, p. 128). Second, our measure may not capture the very high end, hyperactivity; a hyperactivity measure should be predictive of mainly negative outcomes. It could be that moderately high activity protects against development of internalizing problems and also promotes social competence behaviors.

Attachment.  Our findings indicate that attachment security in infancy is relevant for the development of social competence, a conclusion initially drawn from results in the longitudinal Minnesota project (Elicker, Englund & Sroufe, 1992) and established as a robust phenomenon in a review by Schneider, Atkinson, and Tardif (2001). The evidence concerning a predictive effect in relation to problem behaviors is more scattered in our data. This is in concordance with conclusions from much of the attachment literature (DeKlyen & Greenberg, 2008), particularly as the ULS employed a low-risk sample, and the insecure category of disorganized attachment, which has been shown to be related to externalizing problems (van IJzendoorn, Schuengel & Bakermans-Kranenburg, 1999), was not included in the data set. However, it is worth noting that attachment security at 15 months predicted several aspects of personality in middle childhood; children who had been secure as infants were 7 years later rated as more extraverted, more open, and less neurotic. Such individual characteristics could well reflect an ease of communication that lies at the heart of attachment security (Stevenson-Hinde, 2007), and they may be connected with phenomena tapped by the social competence measures. New and intriguing were the findings indicating that attachment security in several instances acted as a moderator of various risk factor effects. For instance, security was found to alleviate effects of infant inhibition in social competence development.

Social factors.  Social factors were found to act as moderators in line with expectations; stress experiences and quality of non-parental care were found to promote risk and resilience, respectively, in combination with temperament, relationship, and background factors. However, the direct and positive effects of non-parental day care in relation to social adjustment were striking and unexpected. The debate in scientific as well as popular literature has concerned the opposite, that is, the potential negative effects of non-parental care during early years. Even though neither Bowlby, nor Ainsworth emphasized amount of time spent with a primary caregiver, but pointed to the quality of time spent with the caregiver, negative effects of non-parental day care have been of concern to later attachment researchers, and a hypothesis about disruptive effects on the parent-child relationship has been derived (e.g., Belsky, 2001). However, accumulated evidence now seems to largely refute this hypothesis (Friedman & Boyle, 2008). A recent report (NICHD, 2006), concerning results for behavioral problems and social skills during the period from 15 to 54 months, points to high day care quality as promoting good outcomes, but also that more time spent in non-parental care is connected with more behavioral and social problems. When comparing with the ULS results, which similarly reflected time spent in non-parental care, it should be noted that the number of hours were generally quite low for ULS participants; most children had zero hours per week during the first 12 months (Mdn = 0, range 0–26.3), and during year 4, which showed the highest number of hours, it was about 20 hours/week (Mdn =19.8, range 0–41; only 11 children had no experience of non-parental care at 48 months). Our findings clearly go against concerns about negative effects. Although the time spent in care was relatively limited, we conclude that Swedish children benefit from early non-parental care experiences, thus corroborating Andersson’s results (1992). We interpret the ULS positive effects, which could not be explained by family SES status, as reflecting social training in day care settings of generally high quality characterized by attachment-related aspects such as stable, sensitive, familiar caregivers with whom the child might form good relationships.

The ULS from infancy to 21 years: Predictions of social anxiety symptoms

In this first exploration of longitudinal relations from infancy and early childhood data to the new dataset obtained at age 21 years, the results came out quite clear. General social phobic anxiety symptoms were predicted by shyness/inhibition and activity, with an independent contribution only from the former, whereas social anxiety in interactional contexts had roots in shyness/inhibition, sociability, and activity, with a significant independent contribution from activity and only a marginal contribution from shyness/inhibition. Interestingly, negative emotionality predicted neither anxiety nor depression, although in our analyses of prediction to school age this variable had been a potent predictor of internalizing problems and social competence. Shyness/inhibition and activity also predicted symptoms of depression, although only the shyness/inhibition coefficient reached significance, and neither variable gave an independent contribution. These data point to an interesting specificity in the predictive relations, which was further emphasized in the analyses controlling for anxiety when predicting depression and vice versa; shyness/inhibition was still significantly related to both social anxiety variables when controlling for depression, but it was not related to depression when controlling for social anxiety. This is valuable information in relation to the interest in developmental trajectories of the comorbidity of anxiety and depression (e.g., Merikangas, Zhang, Avenevoli, Acharyya, Neuenschwander & Angst, 2003) by showing that early social anxiety as displayed in shyness/inhibition seems not to predict depression directly, but rather by way of the continuity of social anxiety. This is in accordance with the comment by Merikangas and colleagues that research evidence suggests that anxiety is more likely to predict onset of depression than the reverse.

A new and interesting finding concerned the relations from high levels of activity during preschool age to low social anxiety at 21 years of age, particularly relevant for anxiety in interactional contexts. As noted above concerning our results for the school age data set, activity was found to be relevant for understanding several aspects of adaptation and social functioning. In relation to adult social anxiety the observation that activity in early childhood predicted social competence is particularly interesting. Thus, a chain from early activity to good social competence during school age to low anxiety in social interactions in adulthood may be discerned.

The exploration of relations from earlier relationship experiences with parents used information from infancy and 4 years. Neither infant attachment security nor 4-year maternal sensitivity showed direct relations to the outcomes. However, relational experiences were found to have moderating effects on relations between temperament and all three outcomes; the effect of shyness/inhibition was more pronounced in relation to both of the anxiety variables for children to mothers evidencing low sensitivity at child age 4 years, and low sensitivity in combination with low child sociability predicted high ratings of depression. These results should encourage developmental researchers to cross the boundaries between theoretical territories to expand our knowledge on the mechanisms branching off from developmental trajectories set by temperamental dispositions.

Among the social factors non-parental care showed striking relations to all three outcomes, particularly so for depression, for which this variable explained 22% of the variance. Although positive effects of non-parental care had been observed for social functioning at school age, such strong and direct effects were not expected over this long time period. As previously mentioned, positive effects of early entrance into Swedish day care have also been presented by Andersson (1992). However, Andersson’s data as well as ours date back more than 20 years. It must be acknowledged that the results may not generalize over time due to differences in the organization and quality of non-parental care in different time periods. However, the possibility of positive effects should be noted and discussed. An elucidation of the mechanisms of positive effects would be of great value. As of now we can only speculate that the social training obtained under guidance of well-educated teachers has long-term beneficial effects for most children.

Childhood stress experience was mainly relevant as a moderator of temperament relations to depression symptoms, in that negative emotionality, low activity, and low sociability were all related to depression only in contexts of early stressful experiences. These findings are another indicator of the value of looking at temperamental dispositions as a starting point from which developmental trajectories diverted in various directions depending on encountered life circumstances.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

The ULS started with the idea of joining different theoretical perspectives in understanding socio-emotional development. The results attest to the usefulness of this approach (cf. Vaughn, Bost & van IJzendoorn, 2008), as predictors from several perspectives were often found to be linked to the same outcome in line with the view expressed in the construct of equifinality (Cicchetti & Rogosh, 1996; Sroufe, Egeland, Carlson & Collins, 2005). Interactive effects point to the need for multi-faceted models of risk/buffering factors as well as for a deeper understanding of phenomena of differential susceptibility (Belsky, Bakermans-Kranenburg & van Ijzendoorn, 2007), exemplified in our 21-year data in which child stress experiences constituted a risk factor in relation to symptoms of depression only in combination with certain temperament traits. The ULS follow-up results further attest to the usefulness of long-term prospective studies for gaining an understanding of the many trajectories that individual, relationship, and other factors will act in concert to create.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References

This research was financed by grants from the Swedish Research Council and The Bank of Sweden Tercentenary Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. References
  • Achenbach, T. M. & Edelbrock, C. S. (1978). The classification of child psychopathology: A review and analysis of empirical efforts. Psychological Bulletin, 85, 12751301.
  • Ahadi, S. A. & Rothbart, M. K. (1994). Temperament, development, and the Big Five. In C. F.HalversonJr, G. A.Kohnstamm & R. P.Martin (Eds.), The developing structure of temperament and personality from infancy to adulthood (pp. 189207). Hillsdale, NJ: Erlbaum.
  • Ainsworth, M. D. S., Blehar, M. C., Waters, E. & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Hillsdale, NJ: Erlbaum.
  • American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th edn). Washington, DC: APA.
  • Andersson, B.-E. (1992). Effects of day care on cognitive and socio-emotional competence of thirteen-year-old Swedish schoolchildren. Child Development, 63, 2036.
  • Bates, J. E. (1989). Concepts and measures of temperament. In G. A.Kohnstamm, J. E.Bates & M. K.Rothbart (Eds.), Temperament in childhood (pp. 326). New York: Wiley.
  • Belsky, J. (2001). Developmental risk (still) associated with early child care. Journal of Child Psychology and Psychiatry, 42, 845860.
  • Belsky, J., Bakermans-Kranenburg, M. J. & Van Ijzendoorn, M. H. (2007). For better and for worse: Differential susceptibility to environmental influences. Current Directions in Psychological Science, 16, 300304.
    Direct Link:
  • Bohlin, G. & Hagekull, B. (2000). Behavior problems in Swedish 4-year-olds: The importance of maternal sensitivity and social context. In P.Crittenden (Ed.), The organization of attachment relationships: Maturation, culture, and context. Cambridge: Cambridge University Press.
  • Bohlin, G., Hagekull, B. & Andersson, K. (2005). Behavioral inhibition as a precursor of peer social competence in early school age: The interplay with attachment and non-parental care. Merrill-Palmer Quarterly, 51, 119.
  • Bohlin, G., Hagekull, B. & Rydell, A.-M. (2000). Attachment and social functioning: A longitudinal study from infancy to middle childhood. Social Development, 9, 2439.
  • Bohlin, G., Hagekull, B., Rydell, A.-M., Berlin, L. & Andersson, K. (2004). Behavioral inhibition and attachment as predictors of peer social competence and social anxiety at early school age. In L. A.Leavitt & D. M. B.Hall (Eds.), Social and moral development: Emerging evidence on the toddler years. New Brunswick, NJ: Johnson & Johnson Pediatric Institute.
  • Bowlby, J. (19691978). Attachment. Attachment and loss, vol.1. New York: Penguin Books.
  • Bretherton, I. & Munholland, K. A. (2008). Internal working models in attachment relationships. Elaborating a central construct in attachment theory. In J.Cassidy & P.R.Shaver (Eds.), Handbook of attachment (2nd edn). New York: Guilford.
  • Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research perspectives. Developmental Psychology, 22, 723742.
  • Buss, A. H. & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Erlbaum.
  • Caspi, A., Roberts, B. W. & Shiner, R L. (2005). Personality development: Stability and change. Annual Review of Psychology, 56, 453484.
  • Caspi, A. & Silva, P. A. (1995). Temperamental qualities at age three predict personality traits in young adulthood: Longitudinal evidence from a birth cohort. Child Development, 66, 486498.
  • Cavell, T. A. (1992). Social adjustment, social performance, and social skills: A tri-component model of social competence. Journal of Clinical Child Psychology, 19, 111122.
  • Cicchetti, D. & Rogosh, F. (1996). Editorial: Equifinality and multifinality in developmental psychopathology, Development and Psychopathology, 8, 597600.
  • Crnic, K. & Low, C. (2002). Everyday stresses and parenting. In M. H.Bornstein (Ed.), Handbook of parenting: Vol. 5. Practical issues in parenting (2nd edn, pp. 243267). Mahwah, NJ: Erlbaum.
  • Cunningham, C. E. & Boyle, M. H. (2002). Preschoolers at risk for attention-deficit hyperactivity disorder and oppositional defiant disorder: Family, parenting, and behavioural correlates. Journal of Abnormal Child Psychology, 30, 555569.
  • Degnan, K. A. & Fox, N. A. (2007). Behavioral inhibition and anxiety disorders: Multiple levels of a resilience process. Development and Psychopathology, 19, 729746.
  • DeKlyen, M. & Greenberg, M. T. (2008). Attachment and psychopathology in childhood. In J.Cassidy & P.R.Shaver (Eds.), Handbook of attachment (2nd edn, pp. 637665). New York: Guilford.
  • Elicker, J., Englund, M. & Sroufe, L. A. (1992). Predicting peer competence and peer relationships in childhood from early parent-child relationships. In R.Parke & G. W.Ladd (Eds.), Family and peer relationships: Modes of linkage (pp. 77106). Hillsdale, NJ: Erlbaum.
  • Fridell, M., Cesarec, Z., Johansson, M. & Thorsen, S. M. (2002). SCL-90. Svensk normering, standardisering och validering av symptomskalan. [SCL-90. Swedish norms, standardization, and validation of the symptom scale] (SiS följer upp och utvecklar 4/02), Statens institutionsstyrelse. Sweden: Stockholm.
  • Friedman, S. L. & Boyle, D. E. (2008). Attachment in US children experiencing nonmaternal care in the early 1990s. Attachment & Human Development, 10, 225261.
  • Graziano, W. G., Jensen-Campbell, L. A. & Sullivan-Logan, G. M. (1998). Temperament, activity, and expectations for later personality and development. Journal of Personality and Social Psychology, 74, 12661277.
  • Hagekull, B. (1994). Infant temperament and early childhood functioning: Possible relations to the Five Factor model. In C.F.Halverson, G.A.Kohnstamm & R.P.Martin (Eds.), The developing structure of temperament and personality from infancy to adulthood. Hillsdale, NJ: Erlbaum.
  • Hagekull, B. & Bohlin, G. (1981). Individual stability in dimensions of infant behavior. Infant Behavior and Development, 4, 97108.
  • Hagekull, B. & Bohlin, G. (1995). Day care quality, family and child characteristics, and socio-emotional development in the preschool period. Early Childhood Research Quarterly, 10, 505526.
  • Hagekull, B. & Bohlin, G. (1998). Preschool temperament and environmental influences on the Five Factor Model personality dimensions in middle childhood. Merrill-Palmer Quarterly, 44, 194215.
  • Hagekull, B. & Bohlin, G. (2003). Early temperament and attachment as predictors of The Five Factor Model of Personality. Attachment & Human Development, 5, 218.
  • Hagekull, B. & Bohlin, G. (2004). Predictors of middle childhood psychosomatic problems: An emotion regulation approach. Infant and Child Development, 13, 389405.
  • Halverson, C. F., Kohnstamm, G. A. & Martin, R. P. (1994). The developing structure of temperament and personality from infancy to adulthood. Hillsdale, NJ: Erlbaum.
  • Hwang, C. P. & Broberg, A. G. (1992). The historical and social context of child care in Sweden. In M.E.Lamb, K.J.Sternberg, C.P.Hwang & A.G.Broberg (Eds.), Child care in context: Cross-cultural perspectives (pp. 2754). Hillsdale, NJ: Erlbaum.
  • Hyde, J. S., Mezulis, A. H. & Abrahamson, L. Y. (2008). The ABCs of depression: Integrating affective, biological, and cognitive models to explain the emergence of the gender difference in depression. Psychological Review, 115, 291313.
  • Lanthier, R. P. (1993). The Big Five personality dimensions in middle childhood and adolescence (Doctoral dissertation, University of Denver, 1993). Dissertation Abstracts International, 54(5-B), 2784.
  • Mattick, R. P. & Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy, 36, 455470.
  • Merikangas, K. R., Zhang, H., Avenevoli, S., Acharyya, S., Neuenschwander, M. & Angst, J. (2003). Longitudinal trajectories of depression and anxiety in a prospective community study: the Zurich Cohort Study. Archives of General Psychiatry, 60, 9931000.
  • NICHD Early Child Care Research Network (2006). Child-care effect sizes for the NICHD study of early child care and youth development. American Psychologist, 61, 99116.
  • Reber, A. S. (1985). Dictionary of psychology. London: Penguin.
  • Rothbart, M. K. & Bates, J. E. (2006). Temperament. In N.Eisenberg, W.Damon & R.M.Lerner (Eds.), Handbook of child psychology: Vol.3: Social, emotional, and personality development (6th edn, pp. 99166). Hoboken, NJ: Wiley.
  • Rubin, K. H. (1993). The Waterloo longitudinal project: Correlates and consequences of social withdrawal from childhood to adolescence. In K. H.Rubin & J. B.Asendorpf (Eds.), Social withdrawal, inhibition, and shyness in childhood (pp. 291—314). Hillsdale, NJ: Erlbaum.
  • Rutter, M., Tizard, J. & Whitmore, K. (1970). Education, health, and behaviour. London: Longman.
  • Rydell, A.-M., Hagekull, B. & Bohlin, G. (1997). Measurement of two social competence aspects in middle childhood. Developmental Psychology, 33, 824833.
  • Schneider, B., Atkinson, L. & Tardif, C. (2001). Parent-child attachment and children’s peer relations: A quantitative review. Developmental Psychology, 37, 86100.
  • Schwartz, C. E., Snidman, N. & Kagan, J. (1999). Adolescent social anxiety as an outcome of inhibited temperament in childhood. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 10081015.
  • Sroufe, L. A., Egeland, B., Carlson, E. A. & Collins, W. A. (2005). The development of the person. The Minnesota study of risk and adaptation from birth to adulthood. New York: Guilford.
  • Stevenson-Hinde, J. (2007). Attachment theory and John Bowlby: Some reflections. Attachment & Human Development, 9, 337342.
  • Thomas, A. & Chess, S. (1977). Temperament and development. New York: Brunner/Mazel.
  • Van IJzendoorn, M. H. & Sagi-Schwartz, A. (2008). Cross-cultural patterns of attachment. In J.Cassidy & P. R.Shaver (Eds.), Handbook of attachment (2nd edn, pp. 880905). New York: Guilford.
  • Van IJzendoorn, M. H., Schuengel, C. & Bakermans-Kranenburg, M. (1999). Disorganized attachment in early childhood: Meta-analysis of precursors, concomitants and sequelae. Development and Psychopathology, 11, 225249.
  • Vaughn, B. E., Bost, K. K. & Van IJzendoorn, M. H. (2008). Attachment and temperament. In J.Cassidy & P. R.Shaver (Eds.), Handbook of attachment (2nd edn, pp. 192216). New York: Guilford.
  • Wessels, H., Lamb, M. E., Hwang, C. P. & Broberg, A. G. (1997). Personality development between 1 and 8 years of age in Swedish children with varying child care experiences. International Journal of Behavioral Development, 21, 771794.