Survey Issues

Avoid Survey Fatigue

Survey fatigue happens when a certain population (like the WSU student body) is surveyed so often that people get tired of taking surveys and response rates drop. This is annoying for the population being studied and compromises all researchers’ ability to collect good data because no one will respond to their surveys. A main cause of survey fatigue is people completing an unnecessary census (trying to survey every single student rather than taking a sample). In most cases, a sample will provide just as much information as a census, but costs less and does not risk causing survey fatigue on the WSU campus. Click here for more information.

Survey fatigue can also be minimized by:

  1. Avoid duplicate survey efforts on campus. Many people survey populations within WSU. Sometimes multiple surveys fielded around the same time ask many of the same questions- in cases like this the two groups of researchers should try to collaborate, share data, and avoid asking the same question of the same people multiple times.
  2. Avoid conflicting timeline with other WSU efforts. When faculty, staff and/or students have to engage in university-wide activities or some important university-wide survey is being conducted, it is very likely that they are preoccupied with those activities and they may not have time to respond to your survey during those times. When conducting your surveys, take consideration of the following time periods.
  • Mid-term exams and final exams periods
  • When some important university-wide survey is being conducted, such as GPSA census survey or other surveys
  • When conducting your surveys, some other time periods should be also taken into consideration as it may hard to reach targeted WSU population.
  • University holiday and public holiday periods
  • Spring break, summer break and winter break
  • The beginning and the end of a semester

Four Sources of Survey Error

An important objective when conducting survey research is to mitigate survey error at all phases of the research endeavor.  The following is a list of four commonly described sources of survey error.

  1. Coverage Error: An error in coverage occurs when the sampling frame (or list of sample members) does not accurately represent the population on the characteristic(s) one wants to study. High-quality samples require that every member of the population of interest has a known, nonzero probability of being sampled. Coverage error is measured as the difference between the estimates produced when a sample list is inaccurate compared to what they would have been when using an accurate sample.
  2. Sampling Error: An error in sampling occurs anytime we decide to survey only some, rather than all, members of a sampling frame. Sampling error is measured as the difference between the estimates produced when only a sub-sample of units on the frame is surveyed compared to when the estimates are produced when every unit on the frame are surveyed.
  3. Nonresponse Error: An error in nonresponse occurs when those who do not respond to a survey are systematically different from those who do respond. Nonresponse error is measured as the difference between the estimates produced when only some of the sample responds compared to when all of them respond.
  4. Measurement Error: An error in measurement occurs when respondents in your sample are unable or unwilling to provide accurate answers. Measurement error is measured as the difference between the estimates produced and the TRUE value because respondents provided inaccurate answers to the survey questions.