In surveys, nonresponse is considered a source of possible bias, which increases with the size of the nonresponding group. Nonresponse bias was investigated in 3 samples of offending drivers who were required to respond to an online questionnaire before taking a driver improvement course, creating an initial 100% response rate. The next 2 questionnaire waves were voluntary, and response rates were much lower. Results (means, internal consistency, correlations, etc.) in the first wave were compared between those who responded twice or thrice and those who responded only to the first wave. No substantial differences were found. Compared to common method variance, the effects of nonresponse are very small.
Wåhlberg et al. (Mon,) studied this question.