CAPI Surveys: What They Are, How They Work, and How to Get Your Fieldwork Right

Before digital tools changed how field surveys worked, an enumerator would arrive at a household with a paper questionnaire, a pen, and a clipboard. They would fill in the answers by hand, carry the forms back to the office, and wait for a data entry team to type everything into a system. At every step: transcription errors, lost forms, misread handwriting, data entry mistakes.

CAPI, Computer-Assisted Personal Interviewing, replaced that process. The enumerator still goes to the household. The interview still happens face-to-face. But the questionnaire lives on a tablet or smartphone, and every answer goes directly into a digital form with built-in logic, constraints, and quality checks.

It is now the standard for professional face-to-face survey research globally. Here is how it works and what you need to run it well.

What Makes CAPI Different from Paper-Based Fieldwork

• Skip logic enforced automatically: the form shows only the questions that are relevant based on previous answers. An enumerator cannot accidentally ask a question that should have been skipped, or skip one that should have been shown.
• Validation in real time: the form can enforce constraints at the point of entry. A respondent's age cannot be entered as 200. A phone number must be 11 digits. An income field cannot accept text. Errors are caught before the interview ends, not after the data is analyzed.
• GPS metadata: the device records where each interview was conducted. This is one of the most important quality control mechanisms in field research, because it allows supervisors to verify that interviews were conducted in the right locations and not fabricated at the enumerator's desk.
• Offline capability: platforms like KoboToolbox, SurveyCTO, and ODK allow enumerators to collect data without an internet connection. Completed forms sync to the server when connectivity is available.
• Real-time submission and monitoring: supervisors can review incoming data as it arrives, track completion rates by enumerator, and flag anomalies without waiting for end-of-day reports.

The Main CAPI Platforms

The three most widely used platforms in international development and applied research are:

• KoboToolbox: free, open-source, widely used by NGOs, UN agencies, and academic researchers. Very strong community support and extensive documentation.
• SurveyCTO: more feature-rich and better suited for large-scale and longitudinal studies. Preferred by development economists, J-PAL, IPA, and government statistical offices. Paid subscription required.
• ODK (Open Data Kit): the original open-source mobile data collection platform on which both KoboToolbox and SurveyCTO are based. Still used directly by some technical teams. Less user-friendly for non-developers but highly flexible.

How to Run CAPI Fieldwork Well

A. Design Your Form With the Enumerator in Mind

A form that works correctly in testing can still confuse enumerators in the field. Use clear, short question labels. Add enumerator instructions (visible only to the interviewer, not the respondent) for questions that require explanation or that have unusual skip logic. Build audio prompts for enumerators who are working in language-diverse areas where reading fatigue is a concern.

B. Pilot Every Form Before Full Deployment

A pilot with 15 to 20 real respondents in a representative setting will reveal almost every logic error, translation problem, and flow issue that desk testing misses. Any survey form that has not been piloted before full deployment is a data quality risk.

C. Train Enumerators on the Form, Not Just the Device

Device familiarity is not the same as form familiarity. Enumerators need to understand why each question is asked, how to probe appropriately without leading, and how to handle the specific respondent situations that create difficulty for that form. Training that only covers how to use the app is insufficient.

D. Monitor Incoming Data Daily

Set up a monitoring system before fieldwork begins, not after. Key things to check daily: GPS locations for interviews submitted outside the target area, interview duration per enumerator (flag those consistently faster than the expected minimum), completion rates against targets, and any constraint violations that were overridden.

E. Conduct Back-Checks

A back-check involves contacting a random subset of respondents (typically 10 percent) after the interview to verify that it took place and that key responses were recorded correctly. Back-checks are the most reliable way to detect enumerator fabrication and are standard practice in professional CAPI research.