Response probabilities and response-mode preferences in a self-administered survey

Area of Science:

• Health Services Research
• Survey Methodology
• Patient Experience Research

Background:

• Declining survey response rates pose a challenge to data representativeness.
• The shift towards online surveys for cost and time efficiency may alter respondent composition.
• Understanding factors influencing survey participation and format preference is crucial for accurate data collection.

Purpose of the Study:

• To classify women in a national maternity care survey by their probability of responding, using registry data.
• To classify respondents by their likelihood of choosing a paper versus online questionnaire, using registry and self-reported data.

Main Methods:

• Utilized registry data to predict response probability based on demographic and obstetric variables.
• Employed registry and self-reported data to determine questionnaire format preference.
• Analyzed a national survey dataset on patient experiences with maternity care.

Main Results:

• Response probability varied significantly based on age, number of previous births, and geographic origin (0.25-0.73).
• Education level was a significant predictor of questionnaire format choice.
• Women with lower education levels were more likely to choose paper questionnaires (0.50 probability) than online (0.38 probability).

Conclusions:

• Demographic and obstetric factors significantly influence survey response rates.
• Educational attainment is associated with a preference for paper over online survey formats.
• These findings highlight potential biases in survey data due to varying response probabilities and format preferences.