Combining initial mail questionnaires with phone sampling of non-respondents minimizes data collection costs while eliminating non-response selection bias.

Download image · 1200 × 630 PNG

Research steps

Follow postal replies with telephone sampling

  1. Mail survey; remind after two weeks
  2. Two weeks later: sample 15% of non-respondents
  3. Survey by phone; weight both response routes

Illustrative summary of the source-described two-stage survey. Postal respondents have weight 1; telephone respondents have the inverse sampling-rate weight. Phone participation requires agreement.

See source →

How

When to use

Evaluating health prevention programs among elderly populations often suffers from low questionnaire response rates and non-response selection bias. This method combines a broad mail survey with targeted phone follow-ups of non-respondents to measure risk perceptions and private investments. It is designed for researchers and public agencies seeking cost-effective data collection without compromising representative sample accuracy.

Conduct a two-stage survey combining low-cost postal mailings with targeted telephone sampling of non-respondents.

  1. Mail a written questionnaire simultaneously with program outreach materials to all targeted participants.
  2. Send a second copy of the mail questionnaire two weeks later to individuals who have not yet responded.
  3. Randomly select a 15% sample from the remaining non-respondent population two weeks after the second mailing.
  4. Call the sampled non-respondents by phone to administer the survey directly upon agreement.
  5. Apply sampling weights equal to 1 for postal respondents and the inverse of the sampling rate for telephone respondents to calculate unbiased population estimates.

Why it works

Weighting telephone responses enables researchers to achieve a high overall effective response rate across large study populations.

Example

In Île-de-France, CNAV mailed 40,670 questionnaires to program recipients and received 13,499 postal replies (33.2%), then sampled 4,075 non-respondents by phone to obtain 1,031 additional completes (25.3%), achieving a 50.1% overall response rate.

About the study

Participants

40,670 retired individuals with valid addresses living in Île-de-France (Paris region), France.

Study Design

Two-stage survey design featuring two postal waves followed by 15% phone sampling of non-respondents to collect data on perceived fall risks and financial barriers.

Results

The two-stage protocol increased the total response rate from 33.2% (postal only) to 50.1% overall, showing that 44.9% of retirees perceived no home fall risk and financial fear was a barrier for only 5.3%.

Source

See source →