Test of item-response bias in the CES-D scale. experience from the New Haven EPESE study

We present results of item-response bias analyses of the exogenous variables age, gender, and race for all items from the Center for Epidemiologic Studies Depression (CES-D) scale using data (N = 2340) from the New Haven component of the Established Populations for Epidemiologic Studies of the Elderly (EPESE). The proportional odds of blacks responding higher on the CES-D items "people are unfriendly" and "people dislike me" were 2.29 (95% confidence interval: 1.74, 3.02) and 2.96 (95% confidence interval: 2.15, 4.07) times that of whites matched on overall depressive symptoms, respectively. In addition, the proportional odds of women responding higher on the CES-D item "crying spells" were 2.14 (95% confidence interval: 1.60, 2.82) times that of men matched on overall depressive symptoms. Our data indicate the CES-D would have greater validity among this diverse group of older men and women after removal of the crying item and two interpersonal items.