Maternal Attachment Inventory: psychometric evaluation of the Korean version

This paper is a report of an assessment of the validity and reliability of the Korean language version of the Maternal Attachment Inventory.

Although the Maternal Attachment Inventory has been widely used to measure mother-infant attachment, no psychometric evaluation of the measure with a Korean population has been reported.

Data from 196 mothers within 6 weeks postpartum were collected during July to November 2003 and used to evaluate the validity and reliability of the Korean version of the Maternal Attachment Inventory. Validity was examined using exploratory factor analysis, item-to-subscale correlations, and correlations with other theoretically relevant variables (maternal foetal attachment and maternal sensitivity). For reliability testing, Cronbach's alpha coefficients, item analyses, and subscale-to-total correlations were examined.

Factor analysis showed that the Korean version had three factors explaining 62.9% of the variance. Item-to-subscale correlation coefficients ranged from 0.49 to 0.91. Correlation coefficients between the Maternal Attachment Inventory and both maternal foetal attachment and maternal sensitivity ranged from 0.10 to 0.64. However, correlations between the third factor scores and both maternal foetal attachment scores and maternal sensitivity scores were not statistically significant. Cronbach's alpha coefficients ranged from 0.65 to 0.94. Correlation coefficients of the inter-item analysis ranged from -0.01 to 0.87. Corrected item-total correlation coefficients were 0.31-0.82. Subscale-to-total correlation coefficients of the Maternal Attachment Inventory ranged from 0.52 to 0.92.

Most items of the Korean version of the Maternal Attachment Inventory were adequate for measuring mothers' affectionate attachment. The instrument, however, exhibited weaknesses in the psychometric properties related to validity and reliability, particularly regarding the three items included in the third factor. More studies using factor analysis of the Maternal Attachment Inventory are needed.