Help Topics & Tipsheets
Introduction to the Survey Process
Before starting a project, researchers should be familiar with the steps involved in the survey process. Good surveys start with clear theory, then the translation of theory into sensible survey questions, pre-testing and revision of questions, development and pre-testing of the survey instrument, sampling, data gathering, and, finally, data analysis.
The resources below provide more thorough introductions to the general survey process:
- Background on the Survey Process (Rand.org)
- What is a Survey? (ASA booklet with basic overview of the survey process)
- Fink, Arlene G. 2012. How to Conduct Surveys: A Step-by-Step Guide. Thousand Oaks: Sage Publications.
- Fowler, Floyd J. 2013. Survey Research Methods. Thousand Oaks: Sage Publications.
- Groves, Robert M., Floyd J. Fowler, Jr., Mick P. Couper, Kames M. Lepkowski, Eleanor Singer, and Roger Tourangeau. 2009. Survey Methodology. New York: Wiley Press.
- Rea, Louis M. and Richard A. Parker. 2014. Designing & Conducting Survey Research: A Comprehensive Guide. San Francisco: Jossey-Bass.
- Salant, Priscilla and Don A. Dillman. 1994. How to Conduct Your Own Survey. New York: Wiley.
Instrument Design and Development
Designing effective survey questions is among the most important elements of survey research, but it can often be a challenging one. Survey response psychology is complex and there are many potential sources of measurement error that poorly designed surveys and survey questions can introduce. The resources below provide more information about constructing effective survey questions and minimizing measurement error.
Surveys can be conducted in many different modes, including online, telephone, mail, and face-to-face. Each mode has unique advantages and disadvantages that researchers should consider before selecting the best format to implement a survey. The resources below provide more information about survey modality. Most Duke affiliates interested in conducting their own surveys will likely be interested in web survey formats. The Duke Initiative on Survey Methodology offers training and consultation in how to construct web surveys.
One of the most important aspects of survey research is selecting a population of interest and developing a method to sample units from that population to survey. There are a number of common sampling techniques, and researchers should carefully consider which approach best suits their needs. The tipsheet below offers more information about sampling methods and how those methods affect data analysis.
DISM Tipsheet: Sampling
One common source of error in surveys is nonresponse. It comes in two forms: 1) unit nonresponse in which sampled units do not answer the survey and 2) item nonresponse in which participants return the survey but leave individual questions unanswered. The tipsheet below provide advice on how to deal with nonresponse error.
DISM Tipsheet: Nonresponse Error
Cross-Cultural Survey Research
Many researchers want to conduct surveys in multiple languages, but surveys of this type can raise a number of concerns, including issues of translation and cultural norms. Here are some useful links with tips on conducting cross-cultural surveys.
DISM Tipsheet: Cross-Cultural Survey Research