After compiling the data, nonresponse bias was tested using a t-test to
examine differences between early and late respondents on all of the
measures. Respondents who responded after the initial message were
termed ‘early,’ while those who responded after receiving a second message
were termed ‘late’ respondents