CAPI is an established abbreviation for Computer Aided Personal Interviewing.
Gaining ground in International Development as organisations move to replace pen and paper interviewing, abbreviated as PAPI for surveys and routine data collection.
Moving to an app or mobile-based system has a number of advantages over a paper-based system.
With CAPI it is much quicker and easier to collate, analyse and present data. Data is digitised on collection so there is no need to spend time typing up data. Analysis can be set up in advance. For example, data on indicators already set up is automatically calculated on completion of a survey. Additional analysis can be easily completed by running queries on the data set.
There is a reduced chance of error when inputting data. Input fields can be flagged as only containing a certain type of data (e.g. numbers within a certain range, drop-down options), and an error/guidance message can be programmed to appear if the wrong answer is entered. An error message can also be programmed to appear if a question is missed out.
Using CAPI It is quicker to collect data using an app. The app can be programmed to only show questions that the enumerator should complete and programmed to jump questions where needed.
A mobile phone or tablet is portable. It is easier to transport than folders of paper which are susceptible to getting wet in a boat or in the rain. Similarly, there is less need to pre-plan data collection. Enumerators do not need to spend time printing off forms before planning to go into the field, just ensure they take a charged mobile phone/tablet with them.
Photos and GPS locations can be automatically incorporated into the survey, avoiding the need for an additional camera or GPS device which later has to be linked to the results. This also allows for GPS recording of where the interview took place.
Data once in the cloud can be presented as a “dashboard” and available to all with user access as they require.
When we look at comparing costs of PAPI versus CAPI we do find that CAPI is more cost effective. This becomes more apparent the bigger the survey.
There are however some limitations/issues to bear in mind when using an app/mobile system.
The initial financial outlay for the phones/tablets and software (programme) is expensive. This can be reduced this cost where possible, by using free software. There may be a need to develop capacity and skills in using apps and providing assistance with programming. In some locations there might be limited mobile phone coverage. To overcome this, you need to use an app/system that either stores data when there is no coverage, for submission later when there is coverage.
There are numerous CAPI platforms that allow mobile phones or apps to be used to collect data and transmit data digitally to a server.Some use open source and therefore free to use and others have a subscription where support is included.
There are many and some are mentioned below.
SurveyBe is an efficient data collection system that is easy to set up. It has two parts – designer and implementer. Designer is where you build the questionnaire and download the data, and implementer is where the questionnaire is completed. They provide training and support but there is a cost associated with this. There is also a cost to use the app.
Epi Info is open source software used in public health. It has the provision for easy, customised data entry and database construction, and data analyses with epidemiologic statistics, maps, and graphs.
Magpi works well and is easy use and set up. However, it does incur fees.
AKVO works well but has a fee for using but does bring support.
mWater provides an accessible, easy to use data collection and analysis mobile app and commonly used in the WASH sector. It has extensive mapping capability. Support options are available from the team behind the app, but there is a cost involved with this. Otherwise it is free to use.
Open Data Kit Collect is a very popular open source CAPI program in which questionnaires are implemented on mobile smartphones. The ODK questionnaires are written in .xml format and can be created manually or generated automatically using an online interface. The ODK software suite consists of several different programs including ODK Collect and ODK Aggregate. The ODK Collect program is installed on a smartphone and the questionnaires are subsequently saved to the phone’s memory, where they can be accessed and completed without wireless connectivity. ODK can keep data from multiple surveys saved to the phone, so in the absence of mobile network, enumerators continue for a few days and then return once in a signal‐accessible area to send several completed questionnaires all at once.
ODK Collect enables users to ask questions not only in a linear sequence but also with a predetermined if‐then logic system, relying on answers to previous questions. The program also supports the incorporation of GPS points, photos, videos and sound recordings as attachments to surveys or as the basis of the questionnaire responses. Survey results are sent to a server hosting ODK’s Aggregate tool via Wi-Fi or a mobile internet connection. ODK is compatible with smartphones running the Android operating system.
Kobo is a free open source data collection system. The advantage over ODK is that hosting is provided by Kobo and is free. Features are more advanced and easier to use than ODK.
A final output is often a dashboard that displays the results. Many of the CAPI platforms have the ability to create a web-based dashboard.
A dashboard should have a user-friendly interface to make it easy for the user to access information and download reports. And the ability to import & export data from multiple sources with seamless reporting making it possible to create custom reports tailored to all stakeholders.
If you are looking to implement CAPI or create monitoring dashboards and need some help do reach out to David Hearle who will be able to help.
David has access to UDACAPI which is a very cost-effective system for surveys and can be used for routine monitoring with results displayed on a web-based dashboard.