Efficiency in any operation is critical for the evaluation of outputs. In research, in general, and socioeconomic research in particular, quality of data and the speed at which they are processed largely determine the efficiency of a program. It is not unusual to get the baseline report of a project in the middle of its execution or even towards its end. This is not necessarily a failure of project managers but part of the constraints they face due to the data collection tools available to them. Thus, besides the project timeline, data collection tools play an essential role in effectively executing project activities.
Traditionally, data was obtained using Pen-and-Paper Personal Interviewing (PAPI) tools. PAPI methods often require almost no programming skills to implement the survey and give project managers a great deal of flexibility in implementing the survey design, particularly for qualitative surveys that require open-ended or qualitative responses in local languages. However, data collection using PAPI tools is also prone to survey errors. All of these investigations, triangulations, and corrections take time and cause delays, particularly to produce baseline indicators, monitor, and track changes in performance indicators efficiently as well as in delivering project outputs.
As part of bigger efforts to improve research quality and efficiency, IITA social scientists have switched from the traditional PAPI method of data collection to CAPI (Computer-Assisted Personal Interviewing) method that relies on electronic gadgets. The CAPI method has several features that make it different from PAPI.
CAPI requires programming skills and to transfer questionnaire designs into an electrotonic format that are not very difficult to acquire. The use of CAPI also requires the purchase of gadgets with costs that have been decreasing over time. Overall, the benefits of using CAPI, in terms of better data quality, availability, and survey costs, outweigh the investment costs to acquire the technology, particularly for complex large-scale surveys.
Switching from PAPI to CAPI takes care of the data quality and efficiency challenges of PAPI. CAPI helps to ensure high-quality data by facilitating logic checks and real-time validations during the interview. In addition, by programming consistency checks for values, CAPI tools can reduce the number of implausible values for some variables of interest such as income, age, land size, etc. For instance, if the age of respondents is expected to fall within a certain range (e.g.,18-80 years), automatic consistency checks can be programmed into CAPI tools to ensure that the age of respondents is limited between the lower and upper bound values.
|MODULE A: Household composition|
|AGE CHECK: the age of the farmer is between 18 and 80 years?|
Enable if: current.A03_Age BETWEEN 18 and 80
Figure 1. Consistency check for implausible values
Similarly, to ensure data quality, logic checks or validations can also be directly programmed into CAPI tools. For instance, if a farmer grows three different cassava varieties on a given plot, it is logical to assume that the sum of the proportion of area allocated to each type adds up to 100 percent. This could be automatically programmed for validation during the survey.
|Does the proportion of different varieties add up to 100?|
|Enable if :(current.D06_1_prop_var1) =100 OR (current.D06_1_prop_var1 + current.D06_1_B_prop_var2)=100 OR (current.D06_1_prop_var1 + current.D06_1_B_prop_var2 + current.D06_1_C_Prop_var3)=100 OR|
Figure 2. Logical check to ensure data quality
CAPI tools also allow project leaders and survey supervisors to monitor the activities of enumerators by capturing the interview start and end times as well as GPS locations of respondents. In addition, since data are collected in electronic format, it can be uploaded in real-time to a central server for consistency and data quality checks by supervisors, making the data readily available for analysis right after data collection is completed. This dramatically reduces the time required to transfer data from the questionnaire to electronic format and, by extension, delays in data availability. CAPI tools also minimize data loss due to missing questionnaires or poor legibility.
The socioeconomics team at IITA has successfully collected data using CAPI tools several projects such as Cassava Monitoring Survey (CMS), Stress Tolerant Maize for Africa (STMA) survey, and Nigeria Baseline and Varietal Monitoring Survey (NIBAS). Through CAPI tools, the team has been able to capture survey information in near real-time, undertake data quality checks on a daily basis, and monitor enumerator tasks. Prior to each survey, all enumerators were trained on how to use tablets. To ensure data quality, during the enumerator training and questionnaire pretest stages, the survey supervisors and project programming teams undertook quality checks for programming and translation errors. In addition, the supervisors often provided real-time guidance to all enumerators during the data collection. These large-scale surveys by IITA were successfully conducted using the Surveybe software (https://surveybe.com). Other popular CAPI tools include CSPro (https://www.census.gov/population/international/software/cspro/) and Open Data Kit (ODK) (https://opendatakit.org).
Authors: Tahirou Abdoulaye, IITA-Niger/Mali; Arega Alene, IITA-Malawi; Tesfamicheal Wossen, IITA-Kenya; and Victor Manyong,
IITA-Tanzania (Dar es Salaam).