Psychometric Properties of the Edinburgh Postpartum Depression Scale in Mexican Mothers

The Edinburgh Postnatal Depression Scale (EPDS) is widely used in epidemiological, research, and clinical contexts. Although it has clearly established cut-off points, its psychometric properties have not been thoroughly analyzed—particularly considering the ordinal nature of its items. To address this gap, the present psychometric study aimed to analyze the items, test a unidimensional (single-factor) model, estimate overall internal consistency, describe score distribution, and evaluate concurrent construct validity in relation to perceived social support (negative association), depression during pregnancy (risk factor), and marital status (being a single mother as a risk factor). The EPDS, along with a Perceived Social Support Scale (PSSS), was administered to an incidental sample of 142 Mexican women who were at least two months postpartum. The ten items demonstrated discriminating power and internal consistency. The one-factor model was validated with an acceptable to good fit using the weighted least squares mean and variance adjusted estimation method. Internal consistency was very high (McDonald’s ω = .93; Green-Yang’s ω = .91). The total score distribution deviated from normality due to positive skewness. Expectations were met for all three aspects of validation. The correlation between the EPDS total score and the PSSS total score was negative (r = −.66, 95% BCa CI [−.75, −.54]). The mean score of mothers with depression during pregnancy was significantly higher than that of mothers without depression during pregnancy (md = 6.15, 95% BCa CI [3.96, 8.83]). Single mothers without partners had a significantly higher mean rank than married and cohabiting mothers. It is concluded that the EPDS is a reliable and valid instrument and should be interpreted using percentile ranks. Future studies are encouraged to replicate these findings and to treat marital status as a polytomous qualitative variable, nuanced by the presence or absence of a partner, when included in statistical analyses.