The Attachment in Middle Childhood Screening Tool

Student thesis: Doctoral ThesisDoctor of Clinical Psychology


Background: Attachment relationships in middle childhood have immediate and long-term implications for an individual’s social and emotional development, well-being, and family functioning. This highlights the need for accurate assessment and intervention to reduce the impact. There currently is a lack of validated and reliable middle childhood attachment measures, and hence the development of an attachment screening tool has been recommended (NICE, 2015).

Aims: To use state-of-the-art, contemporary methods in psychometrics to investigate, for the first time, the Attachment in Middle Childhood Screening Tool (AMCST), a new, uniquely designed, parent completed questionnaire. The AMCST consists of nine attachment related scenarios, accompanied by 14 items, with five responses that capture the different attachment styles and an unsure response. To explore the functionality of the scenarios in measuring attachment style as latent trait (underlying continuum). To evaluate how informative, discriminative and difficult each item is individually for each attachment style. To establish the reliability and validity of the tool.

Method: At time point one, 239 parents of children aged 6-12 years, recruited through schools and social media, completed the AMCST and five other measures. At time point two, 138 parents completed the AMCST. Item factor analysis-confirmatory factor analysis (IFA- CFA) examined dimensionality. Item response theory (IRT) models focused on each item and adjoining scenario separately to evaluate their discrimination, difficulty and information. Classical test theory results present the AMCST reliability and construct validity.

Results: IFA-CFA confirmed unidimensionality of the AMCST. IRT results indicated high endorsement for secure and disorganised attachment style item responses but indicated low endorsement for insecure responses. Therefore, suggesting that the AMCST measures secure and disorganised attachment but not insecure attachment. Four items were deemed more informative, discriminative, and difficult than others. Good test re-test reliability and construct validity was established.
Conclusion: The preliminary study of the psychometric properties of the uniquely designed AMCST using contemporary psychometrics has resulted in a wealth of information regarding attachment in middle childhood. Future research could be used to strengthen the performance of the measure’s insecure attachment responses and investigate the criterion validity of the AMCST.
Date of Award1 Nov 2022
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorSilia Vitoratou (Supervisor), Crispin Day (Supervisor) & Patrick Smith (Supervisor)

Cite this