Survey

Survey is a ubiquitous term that is used and understood differently depending on the context. Its definition is further complicated because it is used interchangeably as a synonym for other topics and activities listed under the broad classification of survey research or survey methods. There are multiple uses of the term survey that are relevant to particular linguistic applications. Survey is used as a noun when it refers to a document (e.g. fill out this survey) or a process (e.g. to conduct a survey); as an adjective (e.g. to use survey methods); or as a verb (e.g. to survey a group of people). Survey is used interchangeably with and strongly associated with the terms poll and public opinion polling, and survey methods are used to conduct a census—an enumeration of the total population. In addition to being used to identify a type of research tool, survey research is a subject that can be studied in an educational course or workshop, and it is an academic discipline for undergraduate and graduate degrees. This entry focuses on the basic definition from which these other uses of the term survey have evolved by outlining the essential elements of this term as it relates to the scientific study of people.

A survey is a research method used by social scientists (e.g. economists, political scientists, psychologists, and sociologists) to empirically and scientifically study and provide information about people and social phenomena. A survey is scientific because there is an established process that can be followed, documented, and replicated. This process is rigorous and systematic. The typical steps in the survey process are (a) problem formation, (b) hypothesis development, (c) research design, (d) sample design and selection, (e) questionnaire development, (f) data collection, (g) data analysis, (h) reporting and dissemination, and (i) application of information. Underscoring the complexity of a survey, each of these steps in the process also has a set of essential and accepted practices that is followed, documented, and replicated, and specifie professional training is required to learn these practices. The documentation that accompanies a survey provides the information necessary to evaluate the survey results and to expand the understanding and analysis of the information provided from the survey. Sampling error and nonsampling error are two dimensions of survey error. Most scientific surveys could more accurately be called “sample surveys” because probability theory is used to scientifically select subgroups of the population to study, and there is a body of knowledge on acceptable statistical procedures for sampling and for calculating sampling error. Quantifying the nonsampling error associated with other steps in the survey process (e.g. question wording, interviewer effects, and item and unit non-response) is more challenging. Total survey error incorporates the error possible in any of the steps in the survey process. When the results of a survey are reported and, particularly, when people use survey results to make decisions, it is very important to review the documentation that describes how the survey was conducted so the quality of the data can be assessed.

...