Mobile Data Collection

(The following is an excerpt from “Benchmarking of Mobile Data Collection Solutions – 2017” by CartONG, produced for UNHCR)

 

Mobile Data Collection (MDC) can help improve the quality of data, information, analysis and decision making. By using one of the MDC platforms described in this report, organizations can collect data faster and with fewer errors than on paper. The sharp decline in hardware costs for mobile phone also means that MDC is often cheaper than doing a survey on paper.

Advantages of Mobile Data Collection

Mobile data collection has many advantages over paper-based alternatives:

  • Fewer errors: “Garbage In – Garbage Out” is the mantra of many evaluators. MDC is able to reduce the amount of “garbage” significantly by eliminating or reducing two potential sources of errors:
  • Inconsistent/impossible/missing data: [Many] products (…) include internal checks that highlight impossible or inconsistent data during the data capture phase in the field so that it is not possible to enter 114-year-old children or pregnant men. In many cases, this type of automatic plausibility check also includes errors of summation, for example during household survey when expenditure for individual household items don’t add up to the total amount. Since the software can alert the enumerator to issues on the spot, the data can be corrected immediately. The same goes for fields that on a paper survey might be accidentally skipped or omitted. This is particularly relevant since research has shown that errors in the data capture phase are not random, but biased towards households with distinct characteristics, which might skew the whole data set.
  • Re-keying errors: At some point, data from all paper-based surveys has to be entered into a computer. During this process errors invariably happen. Because data collected digitally does not need to be re-entered, this source of errors is effectively eliminated.
  • Faster data collection: Data collection via mobile devices tends to be faster than on paper, partially because of built-in functions that can automatically skip questions based on previous answers. For example: if a household does not have any children, questions related to the children can be skipped automatically. The time savings increase with complexity and length of the survey. For example, Fitzgerald et al. found that by using MDC with skip-logic, they were able to save close to one hour per household during an in-depth household survey in Ethiopia and Malawi that, on paper, ran to 50 pages. Given that most households were subsistence farmers, the researchers also found that respondents were more likely to answer all questions when the survey took less time.
  • Faster analysis: Because the data doesn’t have to be manually entered, it is also much faster to run simple analyses on the data, even while the survey is still underway. All applications and services tested for this report include at least a basic tool to visualise data out of the box. In addition to providing NGO staff with answers more quickly, this can also be an important feedback tool for communities that have been surveyed.
  • Better quality control: Many MDC applications are able to capture the GPS coordinates where an interview takes place, as well as the time the interviewer took to complete the interview. The GPS coordinates allow supervisors to ensure that staff have visited the right location and facilitates repeat visits which might be necessary for monitoring. The duration can help to identify enumerators who are either extremely fast or extremely slow, either of which might be an indicator of quality issues.
  • Costs: Costs are frequently listed as one of the areas where paper-based data collection has an advantage over digital data collection. However, this depends heavily on the individual case especially on the number of surveys conducted and the number of submissions and length of each survey. While MDC have higher initial costs for software, development and hardware and capacity building, they do have cost benefits in other areas. Changes to digital surveys, for example, can be rolled out easily to all enumerators, while paper survey forms might have to destroyed and reprinted. Also, the costs for data entry are completely removed with MDC. In a 2015 study in Thailand and the Philippines, Oxfam found that MDC was cheaper than paper unless new mobile phones had to be bought for the survey and were not used for anything else afterwards. In most scenarios, this is rather unlikely as staff will use their phone before and after data collection, which eliminates the cost benefit of paper. The cost benefits of MDC grow even further when the same survey is run multiple times, for example for projects where changes over time are being tracked with the same questions.
  • Tool collection: Smartphones are much more than a touchscreen to enter data. Organizations can use a variety of different tools and features to enrich the data by collecting GPS points without a stand-alone GPS receiver, taking photos without bringing along a separate camera, scan barcodes, record audio and video etc. A lot of times this additional data automatically integrated into your survey without any extra efforts.

In short: through mobile data collection, NGOs are able to get more accurate information faster and at a lower cost than with paper.

 Common Challenges of Mobile Data Collection

  • Survey design: As described above, MDC surveys can prevent enumerators from entering impossible data or omitting questions. However, other risks are only encountered in mobile data collection. Sometimes the excitement about a new technology leads to an increased focus on the technical aspects of a survey, at the expense of design of the survey itself. Often, such a shift in focus means that creating a complex form logic is perceived as the key to a good survey while other important elements, such as defining the goals of the survey or questioning the ethics of questions are neglected.
  • Hardware failure: Mobile phones can break, run out of electricity and their batteries are particularly sensitive to high or low temperatures. Replacement devices, paper forms as backup, car chargers and battery packs can mitigate these issues, but in many cases, a severe hardware failure will mean that an enumerator cannot continue her/his work until s/he has returned to the office. In some cases, the data stored on the device might be lost as well.
  • Connectivity: Some MDC solutions require an active online connection to save data. (…)
  • Familiarity with the technology: While smartphone literacy is increasing steadily, this is not the case across all demographic groups and geographic zones. Smartphone literacy can be an issue especially in surveys where, for example, enumerators should be older because the survey is aimed at the elderly.
  • Languages: Many MDC solutions provide the user interface for their server and for the analysis module only in very few languages – sometimes just in English. (…)
  • Security and privacy: Surveys often collect personal information. Based on the right to privacy, recognised in most international human rights treaties, such data is protected. It is the responsibility of the organization collecting the data to ensure that the collection, storage, analysis and publication of data conforms to security and privacy standards and do not pose threat to the individual or its rights. Depending on the type of survey, different levels of security can be acceptable. For example, data on the health of a patient requires a very high level of security and care should be taken that any public visualisation of health data can never be traced back to an individual. A key informant interview on general needs in a camp, on the other hand, might require much less protection.

MDC specific guidance and tools are available in modules 1 to 5, which can be downloaded here.