CAPI, CATI and CAWI each solve different fieldwork problems. The best method depends on the target population, sampling frame, connectivity, questionnaire complexity, sensitivity, geography and required level of verification.
There is no universally “best” survey mode
Research teams often ask whether face-to-face CAPI, telephone CATI or online CAWI is best for African studies. The useful answer is not a ranking. It is a fit-for-purpose decision based on who must be reached, what must be measured and what evidence is needed to defend the sample.
Africa contains highly connected urban populations, rural communities with limited internet access, formal business audiences, mobile-first consumers and hard-to-reach groups. A method that performs well for a bank-customer study in Lagos may be unsuitable for a household survey in remote districts. Mode selection should therefore begin with the target population, not with the research platform.
CAPI: strongest when physical reach and verification matter
Computer-Assisted Personal Interviewing uses a tablet or smartphone for face-to-face interviews. CAPI is well suited to household surveys, retail audits, location-based intercepts and studies where interviewers must verify a respondent, outlet, asset or service environment.
Its advantages include the ability to use GPS, timestamps, skip logic, media capture where ethically appropriate, visual stimuli and offline interviewing. Its limitations are higher field logistics, interviewer effects, travel time and the need for close supervision. CAPI quality depends heavily on recruitment, training, route planning, device controls and live data review.
CATI: efficient for reachable populations and rapid follow-up
Computer-Assisted Telephone Interviewing is useful when a valid phone sample exists or can be generated responsibly. It can reduce travel costs, reach dispersed respondents and support faster fieldwork. It is often suitable for customer research, beneficiary follow-ups, business surveys and monitoring studies.
CATI is less suitable for very long or visually complex questionnaires. Unknown numbers may have lower answer rates, phone ownership may be unequal across demographic groups, and network quality can affect call completion. A good CATI design therefore includes call-attempt rules, appointment scheduling, disposition codes and quality checks on interviewer performance.
CAWI: powerful for digitally reachable audiences
Computer-Assisted Web Interviewing allows respondents to complete a questionnaire online. It can be fast, scalable and cost-efficient when the audience is digitally accessible. It works particularly well for employee research, professional audiences, customer databases and verified online panels.
The critical question is coverage. An online sample should not be described as representative of a population that includes substantial numbers of people who have little chance of being recruited online. Researchers also need controls for duplicate participation, bots or fraud, speeding, inattentive responses and poor mobile questionnaire design.
A practical decision framework
Choose CAPI when location verification, visual materials, physical observation or broad household reach is central. Choose CATI when respondents are geographically dispersed but phone-accessible and the questionnaire can be delivered conversationally. Choose CAWI when the population is digitally reachable and self-completion will not systematically exclude important groups.
Then test the decision against five questions: Do we have a defensible sampling frame? Can the mode reach the required subgroups? Will respondents understand the questionnaire without extra explanation? Can quality be verified? Will the method protect privacy and respondent comfort? If any answer is weak, a mixed-mode design may be stronger.
When mixed-mode research makes sense
Mixed-mode designs can improve coverage, but they require discipline. For example, an urban consumer sample might combine online interviewing with face-to-face recruitment for underrepresented groups. A beneficiary study could use CATI as the primary mode and CAPI for locations with weak phone access.
The questionnaire, sample allocation and analysis plan must account for mode effects. Researchers should not simply merge three convenience samples and call the result nationally representative. The value of mixed mode comes from a deliberate coverage strategy, not from using more technologies.
Compare modes on coverage, not only cost
Cost is important, but the cheapest mode can become expensive if it excludes the people the study needs to represent. Coverage should therefore be assessed before fieldwork starts. For CAPI, the main question is whether interviewers can physically reach the sampled locations and respondents safely.
For CATI, researchers need to understand phone ownership, number quality, network reliability and whether some groups are systematically less reachable. For CAWI, the critical issues are internet access, device ownership, digital literacy and whether recruitment sources overrepresent particular types of people. A mode decision is stronger when the team can explain who has a realistic chance of selection and participation, not merely how quickly interviews can be completed.
Questionnaire design should change with the mode
The same questionnaire should not automatically be copied across CAPI, CATI and CAWI. In face-to-face interviews, trained interviewers can explain complex instructions, manage showcards and observe non-verbal confusion. Telephone surveys need shorter sentences, fewer response options and an interview length that respondents can tolerate without visual support.
Online questionnaires need clean mobile design, clear self-completion instructions and careful control of grids, long lists and open-ended questions. Sensitive questions may also perform differently by mode because privacy and social desirability vary. When a study uses more than one mode, researchers should identify items most vulnerable to mode effects and test them during piloting.
Quality assurance differs by technology
Each mode creates its own verification opportunities and risks. CAPI can provide GPS coordinates, timestamps, route information and supervisor observation, but these controls are meaningful only if the research protocol defines acceptable ranges and investigates anomalies. CATI quality assurance can include call recordings where consent and law permit, dial histories, interview duration, disposition codes, silent monitoring and call-backs.
CAWI requires controls for duplicate entries, bots, straight-lining, speeding, impossible answer patterns and suspicious device or network behaviour. The objective is not to collect as much paradata as possible. It is to select evidence that helps confirm respondent eligibility, interview authenticity and data quality without compromising privacy.
When a mixed-mode design is justified
Mixed mode is most useful when it solves a specific coverage or operational problem. A customer survey might begin with CAWI and use CATI follow-up for important segments with low online response. A national beneficiary study might use CATI in areas with strong phone coverage and CAPI in locations where access to functioning numbers is weak.
Researchers should plan how respondents are allocated to modes, whether the same sampling frame is used and how differences in question presentation will be handled analytically. Results should also be checked for systematic mode differences before datasets are combined. Mixed mode is a design choice, not simply a way to accumulate responses from several convenient sources.
How Surveysphere Africa can support
Surveysphere Africa designs CAPI, CATI, CAWI and mixed-mode studies around the population you need to reach, with field controls and sampling logic built into the research plan. Need to reach the right respondents?
