Response stability in social research: Lessons from a longitudinal study of the family-building process

Abstract

Responses to questions, phrased to elicit identical answers from the same individuals over time, are compared. The questions were used in a South African longitudinal study on the family-building process. Although consistency for certain questions was high, consistency was particularly low for questions on family-planning methods used and desired number of children at the time of marriage. Type of question (e.g. factual or attitudinal), its importance to the respondent, and changing norms influenced response stability. Although disturbing, the degree of response stability was not uncommonly weak when compared with results from other similar studies. The need to evaluate the quality of data obtained in survey research, especially where heavy reliance is placed on memory recall, is stressed.

S. Afr. J. Sociol. 1986, 17(1): 1–6

Responsies op vrae wat geformuleer is om ooreenstemmende antwoorde van dieselfde persone na ’n tydsverloop te verkry, word vergelyk. Die vrae is in ’n Suid-Afrikaanse longitudinale studie op die gesinsbouproses gebruik. Hoewel ooreenstemming vir sommige vrae hoog was, was dit besonder laag vir vrae insake kontraseptiewe metodes gebruik en begeerde kindertal ten tyde van huweliksluiting. Die tipe vraag (bv. feitelik of houding), die belangrikheid daarvan vir die respondent, en veranderende norme het stabiliteit in responsie beïnvloed. Hoewel kommerwekkend, is die graad van responsiestabiliteit nie buitengewoon laag indien dit vergelyk word met resultate van ander dergelike studies nie. Die nodigheid om die kwaliteit van gegewens te evalueer wat deur middel van opnames verkry is, veral waar op respondente se geheue staatgemaak word, word benadruk.

S.-Afr. Tydskr. Sosiol. 1986, 17(1): 1–6