In the wake of the COVID-19 pandemic, telephone surveys have been used extensively for carrying out studies on health knowledge, morbidity, and mortality surveillance. In order to understand the extent of different sources of non-observation errors in telephone surveys, we compare the distributions of units covered in the sampling frame and survey respondents with those who were excluded from the sampling frame and survey nonrespondents, respectively. The distributions are compared with respect to key socio-economic and demographic characteristics, which are often associated with most health outcomes for two different study designs, viz., panel surveys and repeated cross-sectional surveys.