Cell Phone and Face-to-face Interview Responses in Population-based Surveys

Abstract

Findings on the reliability and cost-effectiveness of the use of cellular phones vis-à-vis face-to-face interviews in investigating health behaviors and conditions are presented for a national epidemiological sample from Lebanon. Using self-reported responses on identical questions, percentage agreement, κ statistics, and McNemar’s test were used to make the comparisons. Concordance was almost perfect (κ statistics, κ > .8) for measures including age, health insurance, cigarette smoking (ever and current), and diabetes; and substantial (0.6 ≤ κ ≤ 0.8) for other measures such as education, water pipe smoking, alcohol consumption, and hypertension. Moreover, cell phone interviewing saved approximately US$14 per person interviewed. Future research on their use for health research purposes in Lebanon and the region should address their use alone or in combination with landlines in obtaining nationally representative samples.