Akademik Veri Yönetim Sistemi
Istanbul Ticaret Üniversitesi Sosyal Bilimler Dergisi, cilt.23, sa.51, ss.2088-2106, 2024 (Hakemli Dergi)
The incorporation of properties of complex sample design in statistical analyses is frequently overlooked although multi-staged, stratified, and cluster sample designs are used in surveys. This is important to make design-based inferences based on probability samples. It is possible to estimate the inevitable impact of such designs on statistics by utilizing standard errors, confidence intervals, and design effects. Several programs including Stata, R, and SPSS enable researchers to incorporate the properties of sample designs. The major goal of this study is to identify how the complex sample design adjusted analysis results should be presented with an example of the preferred age at marriage in Türkiye. The study in particular investigates the impact of marriage cohort on women’s likelihood of reporting marriage age preference below the median age at marriage in Türkiye while controlling for background characteristics. The 2018 Turkey Demographic and Health Survey, a cross-sectional, face-to-face survey, is the data source for this study. The nationally representative data sets provide information on household composition, fertility, family planning, etc. In addition to proportions of marriage age preference by background characteristics of women in Türkiye, design adjusted estimates of binary logistic regression analysis technique was used to reach objectives. The results indicated that there were some variations in the standard errors of proportions and regression coefficients for study variables. In addition to significant effect of marriage cohort, statistical parameters showed the variance inflation in the complex sample design compared to that of a simple random sample design with the same size. Incorporating properties which are generally overlooked in analyses using complex sample designed survey data consists of the main originality. Moreover, the study is original in terms of presenting representative results focusing on marriage cohorts to the literature which is devoted to marriage age preference in Türkiye.