Determinants of attrition in NIDS-CRAM Waves 1 & 2

Reza Daniels, Kim Ingle and Tim Brophy
Additional contact information
Kim Ingle: Southern Africa Labour and Development Research Unit (SALDRU)
Tim Brophy: Southern Africa Labour and Development Research Unit (SALDRU)

No 271, SALDRU Working Papers from Southern Africa Labour and Development Research Unit, University of Cape Town

Abstract: This paper investigates the determinants of attrition between Waves 1 and 2 of the National Income Dynamics Study – Coronavirus Rapid Mobile Survey (NIDS-CRAM). The number of successfully interviewed respondents reduced from 7073 in Wave 1 to 5676 in Wave 2, which represents almost 20 percent of the sample. We fit probit regression models to predict the determinants of attrition and estimate marginal effects for four different specifications of the model. A useful finding is that attrition appears to be random across all four regression models based on the observed covariates, when measured by standard goodness of fit statistics. However, one of the most important findings is that respondents who underwent Covid-19 tests are 3 percent more likely to drop out of the survey. While this rate is low, it is a worrying trend that must be closely monitored in future Waves because it will negatively affect the efficacy of the survey to track Covid-19 testing behaviour. More generally, we find that attrition in NIDS-CRAM is not based on the same observable characteristics as its predecessor NIDS, which showed clear evidence that it was correlated with higher income households. Attrition is also not influenced by how often respondents previously participated in NIDS. It is affected by language of the interviewer, the sample batch the respondent was in during Wave 1, and contact effort by the survey organization. The most important factors influencing attrition are therefore related to survey operations rather than respondents. That said, researchers still need to conduct their own investigations about whether attrition on observable characteristics of respondents affect their estimation samples.

Date: 2020
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.opensaldru.uct.ac.za/handle/11090/994 Full text (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:ldr:wpaper:271

Access Statistics for this paper

More papers in SALDRU Working Papers from Southern Africa Labour and Development Research Unit, University of Cape Town Contact information at EDIRC.
Bibliographic data for series maintained by Alison Siljeur ().