You are here

Observational attachment theory-based parenting measures predict children's attachment narratives independently from social learning theory-based measures

Carla Matias, Thomas G. O’Connor, Annabel Futh & Stephen Scott (2014). Observational attachment theory-based parenting measures predict children’s attachment narratives independently from social learning theory-based measures. Attachment & Human Development, 16:(1), 77-92.


The goal was to examine the extent to which widely-used, good quality observational measures derived from attachment and social learning theory overlapped, and whether they differentially predicted two independently-rated key outcomes in young school-aged children, one particular to attachment theory, attachment narratives, and one that is more general, peer-nominated social competence.

Design: Observational methodology; cross-sectional

Variables Measured, Instruments Used

  1. The Coding of Attachment-Related Parenting (CARP): Measures parent-child interaction quality
  2. The Parent Behavior Coding Scheme (PBCS): Assesses contingent maternal utterances
  3. The Manchester Child Attachment Story Task (MCAST):A narrative story stem task to elicit attachment representations in school-age children 
  4. British Picture Vocabulary Test: A standard measure of child verbal intelligence
  5. Peer nominations: Measured social adjustment  
  • Peer-rated popularity
  • Peer rejection
  • Fighting behavior          
    1. Socio-demographic Factors
  • Parental education: low = left school by 16 yrs: mid = secondary/technical qualification: high = degree
  • Income: low = <£175 p/week; mid = £176 to £325 p/week; and high = >£326 p/week
  • Single-parent status
  • Housing type
  • Self-designated ethnicity of parents


  • N=113
  • Participant ages: five-six year-old children
  • Location: Inner-city borough of London, UK, Home and school settings               
  • Eligibility: The study is based on the first (pre-treatment) wave of data from the Primary Age Learning Skills (PALS) study
    • Only children from the second and third cohorts were involved because the attachment narrative measure was not included in the first cohort
  • SES: 79% had state assisted housing; 38% had household income of £175 per week or less
  • Additional:                                                       
    • 47% of mothers were Black African (first generation immigrants),  20% African Caribbean origin, 20% White British/European, 11% “Other”
    • Of the children, 49% were male                                


  1. Given the inclusion of high-risk, ethnically diverse families,  findings may not generalize to other samples
  2. It is possible that recording parenting behavior under more stressful situations such as when the child is worried, ill or unexpectedly separated would have led to stronger associations with child narratives.
  3. Whereas attachment theory measures were coded on global scales, social learning theory measures were based on event sampling. Differences between measures and their predictions might reflect this different coding strategy.
  4. Given the cross-sectional nature of the design, one cannot draw causal conclusions about direction of effects


1. Overlap between attachment and social learning theory measures                                                       

  • Correlations between parenting measures within each theory indicated substantial overlap between Mutuality and Sensitive Responding                                                        
  • Within the social learning measures, there was modest overlap between Positive Attending and Directive Parenting; Criticism was unrelated to Positive Attending but modestly associated with Directive Parenting                                                            
  • Correlations between measures from each theory indicated the highest degree of overlap for positive parenting measures but no relationship between attachment measures and Directive Parenting, and a modest negative relationship between the two attachment measures and Criticism

2. Prediction of child social adjustment from parenting measures                                                 

  • There were modest to moderate correlations between the two attachment-based measures of parenting and attachment security and peer nomination                                                           
  • Positive Attending and Criticism were associated with peer nominations of fighting                      
  • Mutuality was significantly associated with secure attachment narrative independent of all covariates                                                                    
  • Of the other variables in the model, only child gender, and verbal IQ were also significantly associated with attachment security                                                                  
  • Attachment-based ratings did not predict significant independent variance in peer nominations whereas the social learning theory construct of Criticism did; male gender was also a significant predictor.
  • Regression models for peer nominations of Liked and Disliked indicated that neither attachment nor social learning theory measures predicted peer ratings independently; This implies that the attachment theory-based and social learning theory-based measures account for overlapping variance.