This study illustrates the use of panel data and a fixed-effects estimator
to investigate the impact of family planning program inputs on contraceptiv
e utilization in Morocco during the 1992-1995 period. By controlling the po
tential bias resulting from common unobserved determinants of program resou
rce allocation decisions and program outcomes, the methodology helps overco
me an important constraint to the use of non-experimental study designs in
undertaking meaningful impact assessments. Data from a panel of women inter
viewed in both the 1992 and 1995 Morocco Demographic and Health Surveys wer
e used in the study, along with 'program' data from Service Availability :M
odules undertaken in conjunction with each survey round. The results indica
te that changes in the family planning supply environment, in particular in
creased presence of nurses trained in family planning and the level of infr
astructure at public clinics, played a significant role in the increased us
e of modern contraceptives during the study period.