Maternal Sensitivity & Child Attachment Security: A Meta-Analysis

Sensitive responsiveness of caregivers refers to a caregiver’s ability to accurately perceive, interpret, and promptly respond to a child’s signals and needs appropriately. It involves being attuned to the child’s emotional state, providing comfort when needed, and supporting their exploration and autonomy. This caregiving behavior is crucial for fostering secure attachment relationships.

Madigan, S., Deneault, A.-A., Duschinsky, R., Bakermans-Kranenburg, M. J., Schuengel, C., van IJzendoorn, M. H., Ly, A., Fearon, R. M. P., Eirich, R., & Verhage, M. L. (2024). Maternal and paternal sensitivity: Key determinants of child attachment security examined through meta-analysis. Psychological Bulletin, 150(7), 839–872. https://doi.org/10.1037/bul0000433

Key Points

• The primary methods of this meta-analysis included a comprehensive literature search, coding of study characteristics and effect sizes, and multilevel random-effects meta-analytic models to examine associations between caregiver sensitivity and child attachment security.
• Factors like type of attachment measure, reliability of attachment coding, type of sensitivity measure, duration of sensitivity observation, and time between sensitivity and attachment assessments significantly affected the strength of associations between sensitivity and attachment security.
• For fathers specifically, child age and paternal age moderated associations, with stronger effects for older children and fathers.
• The research, while extensive, has certain limitations such as a lack of diverse samples beyond North America and Europe and few studies on fathers compared to mothers.
• This meta-analysis provides critical insights into the universal importance of caregiver sensitivity for promoting secure child-caregiver attachments across diverse populations and family structures.

Rationale

This meta-analysis aimed to comprehensively synthesize over 50 years of research on associations between caregiver sensitivity and child attachment security.

Previous meta-analyses on this topic have been limited in scope or sample size (De Wolff & Van IJzendoorn, 1997; Lucassen et al., 2011; Zeegers et al., 2017).

With the growth of attachment research across diverse populations and family structures in recent decades, an updated synthesis was needed to elucidate when and for whom sensitivity most strongly predicts attachment.

Specifically, this study sought to address three key gaps in the literature:

1. Directly compare effect sizes for maternal versus paternal sensitivity in predicting child attachment security. Previous meta-analyses suggested stronger effects for mothers, but were conducted decades apart with different inclusion criteria.
2. Identify moderators that may explain heterogeneity in effect sizes across studies, as few consistent moderators have been identified in past meta-analyses.
3. Examine associations between sensitivity and specific insecure attachment classifications (avoidant, resistant, disorganized), which are often combined in research despite potentially distinct antecedents.

By addressing these gaps, this meta-analysis aimed to provide an authoritative update on the role of caregiver sensitivity in attachment formation to inform theory, research, and clinical practice.

Method

The authors conducted a comprehensive literature search as part of a larger project called the Child Attachment Studies Catalogue and Data Exchange (CASCADE).

Electronic databases searched included PsycInfo, MEDLINE, Embase, and Dissertation Abstracts International for published and unpublished studies from 1967 to August 2020.

Search terms included variations of “strange situation” and “attachment.” The authors also reviewed reference lists of relevant meta-analyses and included studies.

The search identified 29,980 unique records. After screening, 2,378 studies were included in CASCADE.

Studies were included if they had:

1. An observational measure of caregiver sensitivity
2. An observational measure of child attachment
3. Sufficient statistics to extract an effect size

Studies were excluded if:

1. Sensitivity was measured after attachment
2. Associations were assessed following an intervention (without baseline data)
3. The sample overlapped with another included study

The final sample included 181 studies with 238 effect sizes.

The authors adhered to PRISMA guidelines and provided a flow diagram of the study selection process.

Statistical Measures

The authors used multilevel random-effects meta-analytic models to account for dependencies between multiple effect sizes from the same study.

They used the robumeta package in R and set the assumed correlation between effect sizes within studies to ρ = .80.

Effect sizes were converted to Fisher’s z and then back to correlations for reporting. Heterogeneity was assessed using τ2 and I2 statistics. Publication bias was evaluated using funnel plots, Egger’s test, and p-curve analyses.

Moderator analyses were conducted using meta-regressions, first at the univariate level and then in multivariate models for significant moderators. Categorical moderators were only tested with sufficient degrees of freedom (df > 4).

Results

Main Effects:

Across all caregivers (k = 174 studies, 230 effect sizes, N = 22,914), the pooled correlation between sensitivity and attachment security was r = .25 (95% CI [.22, .28]).

• For mothers only (k = 159, 202 effect sizes, N = 21,483), r = .26 (95% CI [.22, .29]).
• For fathers only (k = 22, 23 effect sizes, N = 1,626), r = .21 (95% CI [.14, .27]).

Moderator Analyses:

The difference between mother and father effect sizes was not statistically significant.

For mothers, significant moderators included:

• Attachment measure (larger effects for AQS vs. Strange Situation)
• Sensitivity measure (larger effects for MBQS vs. other measures)
• Reliability of attachment coding (larger effects for excellent/adequate vs. poor reliability)
• Duration of sensitivity observation (larger effects for longer observations)
• Time between assessments (smaller effects with longer intervals)

For fathers, significant moderators were:

• Child age (larger effects for older children)
• Father age (larger effects for older fathers)

Insecure Attachment Subtypes:

Maternal sensitivity was negatively associated with:

• Avoidant attachment: r = -.24 (k = 35, N = 4,798)
• Resistant attachment: r = -.14 (k = 36, N = 4,838)
• Disorganized attachment: r = -.22 (k = 23, N = 3,799)

Effect sizes did not significantly differ between insecure subtypes.

Insight

This meta-analysis provides the most comprehensive synthesis to date on associations between caregiver sensitivity and child attachment security.

The overall effect size (r = .25) was remarkably consistent with previous meta-analyses, despite a much larger and more diverse sample.

This speaks to the robustness of sensitivity as a key determinant of attachment across diverse populations and study methodologies.

A key finding was the lack of significant difference between maternal and paternal sensitivity effect sizes. This contrasts with earlier meta-analyses suggesting stronger effects for mothers, likely reflecting societal shifts in father involvement and more recent high-quality studies on fathers.

It highlights the importance of sensitive caregiving from both parents for promoting secure attachments.

The identification of several methodological moderators (e.g., type/reliability of measures, observation length) provides valuable guidance for future attachment research.

It suggests the need for standardized, high-quality measures and procedures to accurately capture sensitivity-attachment links.

For fathers specifically, the moderating effects of child and father age suggest sensitivity may become increasingly influential as children develop and fathers gain caregiving experience. This aligns with research on increasing father involvement across early childhood.

The similar effect sizes for avoidant, resistant, and disorganized attachment challenge theoretical notions of distinct caregiving antecedents for these subtypes.

However, the authors note this may reflect limitations in statistical power or the measures used, rather than true equivalence.

Future research should:

1. Include more diverse samples beyond North America/Europe
2. Examine multiple dimensions of parenting beyond sensitivity
3. Consider reciprocal influences between child attachment and parental sensitivity
4. Develop clinical measures of sensitivity for practice settings

Strengths

The study had many methodological strengths including:

1. Comprehensive literature search spanning over 50 years
2. Large sample size (181 studies, 22,914 participants)
3. Inclusion of unpublished studies to mitigate publication bias
4. Multilevel meta-analytic models to account for dependencies
5. Examination of multiple potential moderators
6. Separate analyses for mothers and fathers
7. Analysis of associations with specific insecure attachment subtypes
8. Rigorous assessment of publication bias and p-hacking

Limitations

1. Geographical restrictions: 87% of studies from North America/Europe
2. Demographic limitations: Samples predominantly White, middle-class
3. Few studies on fathers (7% of total sample) compared to mothers
4. Correlational nature of data precludes causal conclusions
5. Potential language bias from English-language search strategy
6. Reliance on study-level moderators rather than individual participant data
7. Inability to test some moderators for fathers due to limited studies

Clinical Implications

The results have significant real-world implications for understanding and promoting healthy child development. They confirm sensitive caregiving as a critical target for parenting interventions aiming to foster secure attachments.

The similar effect sizes for mothers and fathers suggest both parents play important roles in attachment formation. This supports policies promoting father involvement in early childcare, such as parental leave.

The moderating effects of methodological factors highlight the need for high-quality, standardized measures in both research and clinical practice. Developing brief, valid sensitivity assessments for community settings could improve early identification of at-risk dyads.

For clinicians, the results reinforce the importance of supporting sensitive parenting behaviors across diverse family structures. They also suggest potential benefits of targeting both mothers and fathers in attachment-based interventions.

The lack of clear distinctions between insecure subtypes indicates a need for further research on their unique antecedents and sequelae. This could inform more tailored intervention approaches for different attachment patterns.

Overall, this meta-analysis provides strong evidence for the universal importance of caregiver sensitivity in promoting secure attachments and healthy child development.

It highlights promising directions for future research, practice, and policy to support positive caregiver-child relationships across diverse populations.

References

Primary reference

Madigan, S., Deneault, A.-A., Duschinsky, R., Bakermans-Kranenburg, M. J., Schuengel, C., van IJzendoorn, M. H., Ly, A., Fearon, R. M. P., Eirich, R., & Verhage, M. L. (2024). Maternal and paternal sensitivity: Key determinants of child attachment security examined through meta-analysis. Psychological Bulletin, 150(7), 839–872. https://doi.org/10.1037/bul0000433

Other references

De Wolff, M. S., & Van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development, 68(4), 571-591.

Lucassen, N., Tharner, A., Van IJzendoorn, M. H., Bakermans-Kranenburg, M. J., Volling, B. L., Verhulst, F. C., Lambregtse-Van den Berg, M. P., & Tiemeier, H. (2011). The association between paternal sensitivity and infant–father attachment security: A meta-analysis of three decades of research. Journal of Family Psychology, 25(6), 986-992.

Zeegers, M. A. J., Colonnesi, C., Stams, G. J. M., & Meins, E. (2017). Mind matters: A meta-analysis on parental mentalization and sensitivity as predictors of infant–parent attachment. Psychological Bulletin, 143(12), 1245-1272.