Using a subsample of currently married women from the 1993 Ghana Demographi
c and Health Survey (GDHS), this study examines differentials in contracept
ive use by religious affiliation, namely: Catholic, Protestant, Other, Chri
stian, Muslim, No Religion, and Traditional. Logistic regression is employe
d to explore whether reported religious variations in contraceptive Else ca
n be explained by religion per se (particularized theology hypothesis) or b
y other characteristics that distinguish the religious groups (characterist
ics hypothesis). Generally the findings are congruent with the characterist
ics hypothesis, because the contraceptive use differentials by religious gr
oups is accounted for by the differences in socioeconomic and demographic c
haracteristics of these women. However, for the urban Other Christian women
, even after the necessary controls, religion continued to emerge as signif
icant determinant of contraceptive Else. Policy implications of these resul
ts are discussed.