Response reliability in a longitudinal survey in Thailand

Area of Science:

• Social Sciences
• Demography
• Survey Methodology

Background:

• Assessing response reliability is crucial for the validity of large-scale social and demographic surveys.
• Developing countries present unique challenges for survey data quality due to potential resource and infrastructure limitations.

Purpose of the Study:

• To evaluate individual-level response reliability in a developing country context using the National Longitudinal Study in Thailand.
• To compare the response reliability of the Thai survey with established US fertility surveys.
• To identify factors influencing response reliability, particularly distinguishing between factual and attitudinal questions.

Main Methods:

• Analysis of data from two rounds of the National Longitudinal Study in Thailand.
• Measurement of individual-level response reliability for various question types (factual vs. attitudinal).
• Comparative analysis with reliability data from leading US fertility surveys.

Main Results:

• Individual-level response reliability in the Thai survey varied significantly, being high for factual questions and low for attitudinal questions.
• Despite low individual reliability for attitudinal items, considerable distributional stability was observed.
• The Thai study's overall response reliability was comparable to leading US fertility surveys.

Conclusions:

• Response reliability in large-scale surveys is highly dependent on question type, with attitudinal questions posing a consistent challenge.
• Low individual-level reliability for attitudinal questions is a significant concern for social scientists globally.
• Survey researchers must address the low reliability of attitudinal measures to ensure the integrity of findings in social and demographic studies.