Random News

eriinfo

Africa: Increases in Afrobarometer Interview Duration: Trends and Causes

eriinfo.com09 mins
Africa: Increases in Afrobarometer Interview Duration: Trends and Causes


Key findings

  • Between Round 7 (2016/2018) and Round 10 (2024) of Afrobarometer surveys, both average interview duration and average question duration have increased significantly.
  • The average time taken to complete an interview rose from 63.0 minutes in Round 7, to 78 minutes in Round 9 and 79.9 minutes in Round 10.
  • The average time taken to ask and answer a question in Afrobarometer interviews increased from 14.0 to 17.1 seconds.
  • The vast majority of single-country samples showed significant increases in average interview duration over the period under consideration, and almost all were above Afrobarometer’s 45-minute target.
  • Morocco (38.7 minutes) and Zimbabwe (45.7) were the only countries near or below the 45-minute target in Round 7.
  • Most countries displayed large increases in duration across the three rounds, with Côte d’Ivoire, Namibia, Sierra Leone, and Tanzania displaying continuous increases from Round 7 to Round 9.
  • Four countries showed decreases between rounds 7 and 9: Liberia (-13.3 minutes), Eswatini (-10.4), Ethiopia (-9.1), and Nigeria (-8.2).
  • Questionnaire sections – sets of 10 questions each for which timestamps are available – vary significantly in the number of survey items they contain, their average duration, and their average question duration. “Forced choice” questions are associated with a longer duration.
  • Key sources of variation in interview duration are interview language, interview end-times, and interviewer-respondent dyad matching.
  • Among the five most common interview languages, Arabic was associated with the lowest average interview duration (59.8 minutes), followed by English (74.1), Kiswahili (80.8), Portuguese (85.6), and French (89.0).
  • Interviewer and respondent fluency in the chosen interview language contributed to faster interviews, as did having interviewer and respondent who were both native or both non-native in the interview language.
  • Interviews completed in the morning (mean=74.9, SD=20.3) exhibited a significantly lower mean duration compared to those completed in the afternoon (mean=79.1, SD=27.7)
  • Interviews completed by respondents with similar demographic characteristics to interviewers were shorter on average than those in which respondents and interviewers had different demographic profiles, a pattern that was especially strong when respondents and interviewers were both young.

The time taken to complete a survey interview represents both an opportunity cost for potential survey participants and a cost for Afrobarometer. It can determine whether people agree to be interviewed, what kind of answers they give, and whether they choose to terminate an interview early (Gummer & Roßmann, 2014). These factors affect the quality and quantity of the responses received. Additionally, the longer a single interview takes to complete, the more resources (financial and human) are spent on collecting data.


Follow us on WhatsApp | LinkedIn for the latest headlines

One potential mechanism through which longer interview duration can cause less reliable data is “satisficing,” where respondents put less effort into answering questions as they tire, including by choosing the first response, answering questions at random, or giving “don’t know” responses (Krosnick, 1991). Interview length can also affect data validity if people “break off” before its completion. This is particularly concerning if some population groups are more likely to do so than others, as this will lead to their underrepresentation in the sample. For example, Korinek, Mistiaen, and Ravallion (2005) uncovered a negative relationship between income and survey compliance, which suggests that higher-income households may place a higher opportunity cost on their time. Interestingly, research has found that survey break-off is lower when a participant is initially told a longer estimated interview duration (Yan, Conrad, Tourangeau, & Couper, 2011).

Afrobarometer produces high-quality data on public attitudes across Africa. As the demand for insights on African citizens’ opinions grows, the organisation faces a trade-off between lengthening surveys to gather more information and keeping interview duration reasonable to avoid compromising survey validity and reliability.

This methods note analyses patterns in interview duration across the four most recent survey rounds and provides recommendations on how the questionnaires could be adjusted to optimise Afrobarometer survey methodology. Afrobarometer gives respondents an estimated interview duration of approximately 45 minutes. However, we find that the mean interview duration is much longer than this estimate and has increased from 63.0 to 79.9 minutes over the past decade. We also find that question complexity may be contributing to longer interview duration. Lastly, our results indicate that differences in interview language, interview end-times, and respondent-interviewer dyad matching show the most consistent differences in duration.