How to Construct Survey Scales

Surveys are used to make valid and reliable comparisons of attitudes expressed by those filling them out. Surveys can be used in a wide variety of applications, ranging from investigations of important social issues to determinations of customer preferences in market research. Survey scales are simply the formats used to collect response data. The most commonly used scale is the Likert scale, named after psychologist Rensis Likert. It measures degree of agreement (strongly agree to strongly disagree). Semantic differential scales, also common, require respondents to rate a statement against two adjectives with opposite meanings (good-bad, tasty-bland, easy-hard). Self-rating scales are direct measures of the respondent's attitude about the survey item (describes me very well-does not describe me at all.) All these scales provide numerical data and lead to straightforward analysis. Many survey designers supplement numerical scales with open-ended questions in order to add additional depth to the data.

Things You'll Need

• Clear identification of issues
• Survey administrative procedures

Instructions

1. Construct a five-point Likert scale. Develop a series of statements that are either favorable or unfavorable toward the object or issue under investigation, such as, "Menu variety is an important factor for me in choosing a restaurant." Respondents answer on a five-point scale: (5) strongly agree, (4) agree, (3) neither agree nor disagree, (2) disagree and (1) strongly disagree. Test the items to determine which provide sufficient information.

2. Construct a seven-point semantic differential scale. Require respondents to rate the item of interest in respect to two adjectives with opposite meanings. Here is an example item: "I find the food at fast-food restaurants to be _____. " Respondents rate the statement against the bipolar adjectives, "healthy" and "unhealthy." To explore attitudes further, add additional bipolar pairs, such as "tasty or bland," "a good value or overpriced," and "natural or processed." Limit the number of items, because respondents find long semantic-differential surveys tedious.

3. Construct a seven-point self-rating scale. Items should require respondents to answer concerning their attitude about the subject under question. Here is an example item: "How favorably or unfavorably do you feel toward capital punishment?" The respondent answers on a seven-point continuum between "very favorable" and "very unfavorable." Another variant makes a general statement, such as "I support capital punishment." The respondent answers on a continuum between "Describes Me Very Well" and "Does Not Describe Me At All." Self-rating scales are easy to construct, and items require the least amount of preparation and screening by the researcher.

4. Include open-ended items. These items directly capture the respondents' own words and provide perspective beyond that gained by quantitative scales. Responses may add significant nuance, detail or explanatory information. They may also reveal entirely new information. Exercise care when reporting data from open-ended questions. Scrub the data if it could compromise the anonymity of survey respondents.

Tips & Warnings

• Assure respondent anonymity.

• To increase reliability and validity, combine several types of questions.

• Pretest survey items.

• Conduct a pilot survey.

References

• "Methods of Research in Social Psychology;" J. Merrill Carlsmith, Phoebe D Ellsworth and Elliot Aronson; 1976
• "Strategies of Social Research;" Herman W. Smith; 1991