ABSTRACT
There is widespread recognition that survey research today faces significant methodological challenges. People are harder to reach and, when reached, they are less willing to answer questions. Survey results will be biased if those who do not respond are systematically different from those who do. Bias can also be introduced when some members of a study population are excluded, either intentionally or unintentionally, such as the Random Digit Dial telephone surveys that exclude households without landline telephones (i.e., cell phone-only households). At the same time, new technologies have made it less expensive to conduct a survey-especially using opt-in internet samples-so we now find the academic landscape dotted with more original surveys than ever before. While this explosion of data sources offers exciting potential for new research, too often these surveys are marred by less rigorous and less transparent methodological standards that have the potential to undermine any substantive conclusions.
