Contemporary study (option): Ding et al. (2014)

Aim and procedure

Aim: to investigate how mother-infant attachment relates to later attachment, cognitive development and behaviour in young children in China.

Design: longitudinal. Ethical approval from the Children's Hospital of Fudan University; informed consent from all mothers.

Sample: 160 mother-infant pairs recruited through child health networks in Shanghai, mostly middle-income; all firstborn, full-term and healthy (82 boys, 78 girls).

Time 1: the Strange Situation at 12 to 18 months, video and audio recorded; two coders, inter-rater agreement 95%.

Follow-up at about 3 years: 118 children returned (42 lost to follow-up). Measures: the Attachment Q-set (90 cards, sorted by mothers; secure or insecure); the Bayley Scales of Infant Development (Mental Development Index, MDI, and Psychomotor Development Index, PDI), given by two testers blind to attachment type; the Child Behavior Checklist (99 items, mothers). Ten mothers or children were randomly re-tested to check reliability.

Results

• First visit (for the 118 followed up): 76 secure (64.4%), 42 insecure (35.6%). Follow-up classification: 78 secure, 40 insecure.
• Stability: 70 children were secure at both times (reported consistency rate 89.7%) and 34 insecure at both (85.0%).
• Cognition: MDI secure 102.9 (SD 12.6) versus insecure 97.5 (SD 14.8), a significant difference (t-test); PDI 98.8 versus 95.6, not significant. Resistant children scored significantly lower than secure; avoidant children were close to secure.
• Behaviour: insecure infants later scored higher on social withdrawal, aggression and total problems. Insecure (especially disorganised) children tended to show several problems at once.

Conclusions and evaluation

Conclusions: attachment security is fairly stable from infancy to age 3; secure infants tend to have better general cognitive development and fewer behaviour problems, especially withdrawal and aggression.

Evaluation:

• Strengths: longitudinal, so changes are tracked in the same children; reliability checks (95% coder agreement, re-tests); blind cognitive testers add objectivity; ethics approval and consent.
• Weaknesses: 42 of 160 (26.3%) dropped out (attrition may bias results); tiny subgroups (only 10 avoidant and 3 disorganised at follow-up); the Q-set and behaviour checklist rely on mothers' reports (subjective); one city, mainly middle-income (limited generalisability); correlational, so attachment may not cause the cognitive differences.

Worked example

Ding et al. report a stability rate of 89.7% for secure children. Show which numbers give this figure.

1. 70 children were secure at both visits.
2. 70 / 78 x 100 = 89.7%, using the 78 classified secure at follow-up.
3. (Using the 76 secure at the first visit would give 70 / 76 x 100 = 92.1%, so check which total is meant.)

Answer: 70 / 78 = 89.7%.

Check yourself

1. What design did Ding et al. (2014) use?
   • Longitudinal
   • Cross-sectional
   • Meta-analysis
   • Single case study
   Show answer

   Longitudinal. The same children were followed from infancy to about age 3.

2. Ding et al. started with 160 infants and 118 completed follow-up. What percentage were lost? Give your answer to 1 decimal place.
   Show answer

   26.3 %. (160 - 118) / 160 x 100 = 26.2%.

3. At follow-up, 34 of the 40 insecure children had been insecure at the first visit too. What percentage is that? Give your answer to 1 decimal place.
   Show answer

   85.0 %. 34 / 40 x 100 = 85.0%.

4. What did Ding et al. find about the Mental Development Index?
   • Insecure children scored significantly higher
   • There was no difference at all
   • It was not measured
   • Secure children scored significantly higher than insecure children
   Show answer

   Secure children scored significantly higher than insecure children. 102.9 versus 97.5, significant on a t-test.

5. Why did Ding et al. use testers who did not know each child's attachment type?
   • To save money
   • To reduce experimenter bias when measuring cognitive development
   • Because mothers asked them to
   • To increase attrition
   Show answer

   To reduce experimenter bias when measuring cognitive development. Blind testing improves objectivity.

6. Which is a weakness of Ding et al. (2014)?
   • No ethical approval was obtained
   • Coders disagreed on most classifications
   • Children were only tested once
   • Very small avoidant and disorganised subgroups
   Show answer

   Very small avoidant and disorganised subgroups. Only 10 avoidant and 3 disorganised at follow-up.

7. Explain why Ding et al. (2014) cannot conclude that secure attachment causes better cognitive development.
   Show answer

   The study is correlational: attachment was not manipulated, so other factors such as parents' education, income or the child's temperament could cause both secure attachment and better cognitive scores.. Correlation plus alternative cause.

Practise with instant feedback Print this lesson

Sources

• Pearson Edexcel International Advanced Subsidiary/Advanced Level in Psychology specification, Issue 6 (November 2024, first assessment January 2026)
• Pearson IAL Psychology summary of studies: WPS03

Last updated 8 November 2026. Spotted a mistake? Tell us.