Skip to main content icon/video/no-internet

Survey

A survey is a data-collection method in which individuals answer specific questions about their behavior, attitudes, beliefs, or emotions. Surveys are frequently used by multiple disciplines, including social and behavioral sciences, political sciences, public health, and business. Surveys are commonly used in nonexperimental (correlational) research but might also be incorporated into experiments. Nonexperimental surveys can be either cross-sectional (a single data collection), repeated cross-sectional (two or more data collections with different samples), or panel surveys (multiple data collections from the same sample). However, unless an experimental design is used, survey research does not allow for the drawing of causal inferences.

Types of Survey Questions

Surveys are composed of multiple questions assessing the constructs of interests. Typically, questions regarding demographic characteristics of participants (e.g., age, sex, race, marital status, income, or education level) are also included. If new questions are developed, they should be first evaluated in a pilot study to ensure that they are clearly worded and correctly understood, reliable, and valid. The main types of survey questions include open-ended and closed-ended questions. The latter type can be divided into dichotomous, nominal, rank-ordered, ordinal, and continuous, depending on the characteristics of the response options.

Open-Ended versus Closed-Ended Questions

Open-ended questions require participants to formulate answers in their own words. Examples of such questions are “What is the main stressor in your life?” and “What do you think should be done to control crime in inner cities?” Space is typically provided for participants to write their answers. Although open-ended questions typically elicit more complete and deeper answers than closed-ended questions, they are much more difficult to analyze quantitatively. The answers need to be first coded into categories that are either predetermined or developed through content analysis of the obtained responses. The development of the scoring system and coding of individual answers should be performed by multiple “experts” who are knowledgeable in the subject field and trained to perform these tasks. High inter-rater agreement in coding responses is necessary to achieve reliability of scoring. Compared with closed-ended questions, there is a higher risk of participants misunderstanding open-ended questions and providing irrelevant answers or choosing not to answer these questions. For these reasons, most survey questions are closed ended or only partially open ended.

Closed-ended questions list several response options from which the respondent needs to choose. For instance, the question “What is the main stressor in your life?” might be followed by these response categories:

—Finances

—Work

—Relationship with spouse

—Other relationships

—Children

Because closed-ended questions provide a limited number of specific responses, they are much easier to summarize and analyze. These answers are also more relevant and comparable than open-ended responses. However, the participant is forced to choose from alternatives that might not provide the most accurate answer. This problem is addressed in partially open-ended items that provide a set of closed-ended responses (as in closed-ended questions), as well as an open-ended option, such as

—Other (Specify):___________________

Types of Closed-Ended Questions

Depending on the format of the response categories, closed-ended questions might be dichotomous, nominal, ordinal, continuous, or rank ordered. Dichotomous questions offer only two possible answers (e.g., yes/no, true/false). Nominal questions include three or more response options that cannot be ordered. For example, the question on the main stressor given in the previous example is a nominal question. In contrast, ordinal items include three or more response categories that can be ordered on a continuum ranging from low to high levels. A set of such ordered responses is typically called a rating scale. Examples include the

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Sage Recommends

We found other relevant content for you on other Sage platforms.

Loading