The data collection procedures raise two further issues. The length of the
questionnaire likely discouraged respondent participation, contributing to
response rates of less than two-thirds of the eligible respondents at both
survey waves. Scattering the multiple items (presented in Appendix A)
throughout the questionnaire likely lowered the reliability of several measures,
creating threats to measurement validity. Including additional indicators
may have improved reliability and validity for some measures, but this
step would mean increasing the length of an already long questionnaire.
Finding a balance between greater reliability (and, hence, validity) and lower
non-response bias is likely to remain an imp