one couple, Nana and Nortey is especially revealing (and poignant) in that it reveals the couple-based dynamics involved in Nana’s decision to have an abortion, presumably to meet an ‘unmet need’ for spacing. To date statistical analysis of the relationships between levels of ‘unmet need’, levels of abortion and contraceptive prevalence provide ambivalent findings. In 1996, the year before my study, the UNFPA estimated that about 45 million induced abortions took place worldwide (UNFPA 1996). A 1994 USAID report suggests that abortion remains a common way for women to control their fertility (USAID 1994). As Nana and Nortey’s story will show, it is really Nortey, and not Nana, who had the ‘unmet need’, and yet the survey does not capture this.
Between June 1997 and January 1998 I carried out a census among the junior staff workers of the University of Ghana and their spouses; my final sample size included 125 men (husbands) and 140 women (wives). The resulting couple sample eventually contained 110 dyads. The survey instrument I used was modelled along the lines of the Ghana DHS (GDHS) as far as background and family planning questions were concerned. However, additionally, the survey included a series of questions on family decision-making, financial support, and access to resources not contained in the GDHS.[95] The survey was interviewer-administered, conducted in the respondents’ homes, separately for men and women, and interviews on average lasted between 30 and 40 minutes.
From the survey respondents who had agreed to be re-interviewed I short-listed 30 couples for re-interviewing, who reflected the range of attitudes and behaviours:!) couples in which husbands are advantaged in reproductive outcomes; 2) couples who express conjugal ‘jointness’ or agreement about reproductive outcomes; and 3) couples in which wives are advantaged in reproductive outcomes. I eventually interviewed 12 husbands and 11 wives (hence 11 couples) that reflected such a range.[96]
Although I did not prevent interviewees from diverging from my script, and I myself also probed respondents when they went off on an interesting path, I did adhere to my fixed set of questions thereby systematizing the collection of this qualitative material. Each respondent was interviewed separately from her or his spouse. Generally interviews lasted between one and one and a half hours; a few took as long as up to two hours or longer; however, I never observed respondents getting bored or tired, and no one ever suggested terminating the interview (see Appendix for a summary of respondents’ characteristics).
Arnfred Page 124 Wednesday, March 3, 2004 2:38 PM