Effect of the Data Collection Method on Mobile Phone Survey Participation in Bangladesh and Tanzania: Secondary Analyses of a Randomized Crossover Trial
Bibliographic record
Abstract
BACKGROUND: Mobile phone surveys provide a novel opportunity to collect population-based estimates of public health risk factors; however, nonresponse and low participation challenge the goal of collecting unbiased survey estimates. OBJECTIVE: This study compares the performance of computer-assisted telephone interview (CATI) and interactive voice response (IVR) survey modalities for noncommunicable disease risk factors in Bangladesh and Tanzania. METHODS: This study used secondary data from a randomized crossover trial. Between June 2017 and August 2017, study participants were identified using the random digit dialing method. Mobile phone numbers were randomly allocated to either a CATI or IVR survey. The analysis examined survey completion, contact, response, refusal, and cooperation rates of those who received the CATI and IVR surveys. Differences in survey outcomes between modes were assessed using multilevel, multivariable logistic regression models to adjust for confounding covariates. These analyses were adjusted for clustering effects by mobile network providers. RESULTS: For the CATI surveys, 7044 and 4399 phone numbers were contacted in Bangladesh and Tanzania, respectively, and 60,863 and 51,685 phone numbers, respectively, were contacted for the IVR survey. The total numbers of completed interviews in Bangladesh were 949 for CATI and 1026 for IVR and in Tanzania were 447 for CATI and 801 for IVR. Response rates for CATI were 5.4% (377/7044) in Bangladesh and 8.6% (376/4391) in Tanzania; response rates for IVR were 0.8% (498/60,377) in Bangladesh and 1.1% (586/51,483) in Tanzania. The distribution of the survey population was significantly different from the census distribution. In both countries, IVR respondents were younger, were predominantly male, and had higher education levels than CATI respondents. IVR respondents had a lower response rate than CATI respondents in Bangladesh (adjusted odds ratio [AOR]=0.73, 95% CI 0.54-0.99) and Tanzania (AOR=0.32, 95% CI 0.16-0.60). The cooperation rate was also lower with IVR than with CATI in Bangladesh (AOR=0.12, 95% CI 0.07-0.20) and Tanzania (AOR=0.28, 95% CI 0.14-0.56). Both in Bangladesh (AOR=0.33, 95% CI 0.25-0.43) and Tanzania (AOR=0.09, 95% CI 0.06-0.14), there were fewer completed interviews with IVR than with CATI; however, there were more partial interviews with IVR than with CATI in both countries. CONCLUSIONS: There were lower completion, response, and cooperation rates with IVR than with CATI in both countries. This finding suggests that, to increase representativeness in certain settings, a selective approach may be needed to design and deploy mobile phone surveys to increase population representativeness. Overall, CATI surveys may offer a promising approach for surveying potentially under-represented groups like women, rural residents, and participants with lower levels of education in some countries.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.030 | 0.048 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.004 | 0.005 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.008 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".