# CAPI and CATI surveys in Jamaica

Tablet interviews with GPS and timing checks, phone interviews from a supervised script, and paper when the setting calls for it.

Balcostics Research fields surveys on tablets, by telephone, and on paper. The mode is a design choice. We program it, train on it, and supervise it the same way in each parish.

## Tablet interviews

CAPI is the default on household and business surveys when the instrument has skips, rosters or a long module. Interviewers carry Android devices we own. The questionnaire is tested on those devices before training.

On the myVoice field collector the interview can be captured with no signal and synced when the device is back online. The case record stores each contact attempt. A household roster and random (Kish) selection pick the respondent inside the household, to the rule in your sample design.

GPS is stored at the start of the interview and at the finish. Supervisors see a flag when the interview is more than 500 metres from the address on file. The app also stores how long each question took. That timing is a supervision tool. It is not, by itself, a reason we delete an interview.

Three roles are built in: enumerator, supervisor and manager. Interviewers use the field collector free of charge. The study owner pays 10,000 myVoice points a month. The product page for the app is [myvoice.page/field](https://myvoice.page/field).

If the contract names a delivery format, we agree that format in writing before fieldwork starts.

## Phone interviews

CATI suits shorter instruments, follow-up waves, and respondents who are hard to find at home in the day. Interviewers work from the programmed script. Supervisors listen, call back, and review timing.

We have used phone as the main mode and as a mix. From 2020 to 2021 we were a vendor to GeoPoll for CATI and CAPI fieldwork in Jamaica. The 2019 Child Protection Compact baseline, where we subcontracted to GeoPoll, was CAPI with some CATI. The 2020 MSME line-of-credit study for the Development Bank of Jamaica, Lumin and UWI was paper plus telephone, 111 completed interviews, and we were the prime contractor, including the analysis.

## Paper when the job needs it

PAPI remains a real mode. We use it when a tablet is a poor fit for the respondent or the site, or when the client's instrument is already on paper. Forms come back to a data-entry step. We key them, check them, and document the edits in the codebook. The deliverable is the keyed file, the codebook and the field report.

## Tools we will actually name

- **myVoice field collector**, for offline CAPI as described above. It is our platform. Details of the online Research Hub are on [how myVoice works](/how-myvoice-works/).
- **CSPro**, used for MICS 2022, where we were the prime field contractor for UNICEF Jamaica.
- **SurveyCTO**, available when the client's programme specifies it. Our managing director works in SurveyCTO and CSPro. Past jobs used the package the programme named.

In 2014 the team was trained by the Statistical Institute of Jamaica on its eDACS tablets. That was training on a statistical office system. It is not the tool we propose today.

Analysis files are handled in SPSS, Stata and R. Programming and cleaning are described on the [data management](/data-management/) page.

## Quality controls on every mode

Whatever the mode, the field standard is the same shape.

- Interviewers and supervisors are trained on your instrument, then piloted.
- A supervisor is accountable for a named set of interviewers, not a floating pool with no lead.
- Consent is read before the interview. The wording is yours when you are the data controller. Our [data protection](/data-protection/) page states the script we use when we are the controller.
- Callbacks are logged. On the field collector they sit on the case.
- GPS and per-question timing apply to myVoice CAPI jobs. A CSPro or SurveyCTO job uses the checks that package and your programme specify, and we still supervise daily.
- Edits are written down. The dataset, codebook and field report move together.

Training rooms in Kingston and Manchester are available on demand, and vehicles are arranged for the job. Coverage is an island-wide interviewer and supervisor pool, recruited and trained for the survey in front of us.

MICS 2022 is the large CAPI reference: more than 60 interviewers trained, all 14 parishes, 7,118 households completed of 7,401 occupied. Read the [case study](/case-studies/mics-2022/) before you treat any other job as the same scale.

## At a glance

- Modes: CAPI, CATI, PAPI
- CAPI tools: myVoice field collector, CSPro, SurveyCTO
- Devices: Android tablets and phones we own
- GPS flag: Supervisor flag if more than 500 m from the address
- Field app price: 10,000 myVoice points a month; interviewers free


## Questions

**Can you run a survey on tablets and by phone in the same study?**

Yes. The 2019 Child Protection Compact baseline used CAPI with some CATI. The 2020 DBJ, Lumin and UWI MSME line-of-credit study used paper and phone. MICS 2022 was CAPI only, in CSPro.

**What does the myVoice field collector check?**

It captures offline and syncs later. Each case stores contact attempts, a household roster and random (Kish) respondent selection. GPS is taken at the start and the finish, with a supervisor flag when an interview is more than 500 metres from the address. Each question stores timing. It runs on Android.

**Will you use our CAPI package?**

Yes, when the programme names it. We fielded MICS 2022 in CSPro. SurveyCTO is also a tool our project manager uses. On new work we lead with the myVoice field collector unless your programme requires another package.

**Do you still use paper?**

Yes, when the instrument or the setting calls for paper. Paper interviews are keyed as a separate data-entry step, with a codebook, and delivered with the field report.


Source: https://balcostics.com/capi-cati-surveys/