A nonrandomized experiment carried out in Jharkhand, India, shows how the effects of interventions designed to improve access to family-planning methods can be erroneously regarded as trivial when contraceptive use is utilized as dependent variable, ignoring women's need for contraception. Significant effects of the intervention were observed on met need (i.e., contraceptive use by women who need contraception) but not on contraceptive use (i.e., contraceptive use by women who may or may not need contraception). Met need captures the woman's success in overcoming barriers to access to family planning, whereas contraceptive use confounds this construct with risk of pregnancy and fertility desires. Exceptions to this rule are identified. (Contains 3 tables and 1 figure.)