Likert scale
The Likert scale (after Rensis Likert , pronounced Lick-ert) is a method of measuring personal attitudes . The scales consist of several items of the Likert type. These are statements to which the respondents agree more or less strongly on a given multi-level answer scale or which they can reject. The point values of the individual answers are added unweighted and thus result in the value of the scale. The term Likert scale should not be confused with the response scale of a single Likert-type item.
construction
Basic idea
The methodological consideration is as follows: One is interested in the attitude of the test person with regard to a certain object. All related items are formulated as strictly positive or negative statements. The Likert scale is based on the idea that the more the test person rejects the statement of an individual item, the further their attitude deviates from the formulation of the item. As a whole, the answers are then mapped to the degree of attitude. It is hoped that this procedure will result in a methodically stable measurement of the attitude.
In the concrete execution of a questionnaire, the answer options for an item reflect the degree of approval or disapproval of the respondent to the statement of the item. The possible answers are coded as natural numbers and arranged in ascending order.
Answers to the individual items of a Likert scale are formally ordinal - or scaled (and not cardinal), since it cannot necessarily be assumed that test participants perceive the various answer options as equidistant. The result for a Likert scale item can therefore be summarized using the median or mode as a location parameter. There are different opinions as to whether the mean value can also be used. If the Likert scale is formulated symmetrically, the values are quasimetric and can be treated like an interval scale.
The aim in constructing the items and possible answers is to approximate the equidistance requirement (equal intervals between the answer levels). The possible answers are therefore usually formulated symmetrically (completely agree, tend to agree, neither disagree nor disagree, tend to disagree, completely disagree). In addition, if necessary, the perception of an equidistant scale is supported by an illustration of an item scale with the answer options drawn.
The sums of the Likert items are often used as interval scaled.
Even or odd scales?
An even number of possible answers is just as common as an odd number. The latter is used when a medium value (neutral or “neither nor”) makes sense. An even number forces a decision on one side.
Scale construction
It is common to test a large number of items (statements) in a quantitative pilot study . Items are considered unsuitable if they show too few differences between the respondents. It is a deficiency, for example, when more than 80% of the respondents maximally agree with an item ( ceiling effect ) or maximally reject it ( floor effect ). Such items are often sorted out.
Further items can be sorted out due to insufficient or negative correlation with the overall test value in order to arrive at a scale that is as one-dimensional as possible (see Cronbach's alpha ).
application areas
Likert scales are used in questionnaire surveys, particularly in empirical social , market and election research and in psychology .
example
Let's say attitudes towards off-road cars are to be measured. Based on theoretical considerations, the following items, among others, are used:
- Item 1 : Off-road cars are mostly driven by show-offs.
- Item 2 : Women feel protected in off-road cars.
Answer options (assigned codes):
- Variant 1: applies (1), applies somewhat (2), partly-partially (3), rather does not apply (4), does not apply (5)
- Variant 2: applies (1), rather applies (2), rather does not apply (3), does not apply (4)
A statement and the number selected on the answer scale thus represent an indicator of the attitude. The aim is to create a consistent and precise final scale or set of items with which the most valid (valid) result possible for the question examined can be achieved.
See also
literature
- Jürgen Rost: Textbook test theory, test construction. Huber, Bern a. a. 1996, ISBN 3-456-82480-7 .
- Naresh K. Malhotra: Marketing Research. Prentice-Hall, Upper Saddle River NJ 1999, ISBN 0-13-242801-6 .
- Donald R. Lehmann, Sunil Gupta, Joel H. Steckel: Marketing Research. Addison-Wesley, Reading MA et al. a. 1998, ISBN 0-321-01416-2 .
- Rensis Likert : A technique for the measurement of attitudes . In: Archives of Psychology , 22, 140, 1932, p. 55.
Web links
Individual evidence
- ↑ How Do You Pronounce Likert? What is a Likert Scale? East Carolina University Department of Psychology on May 29, 2015, accessed December 14, 2015.
- ↑ James Carifio, Rocco J. Perla: Ten Common Misunderstandings, Misconceptions, Persistent Myths and Urban Legends about Likert Scales and Likert Response Formats and their Antidotes . In: Journal of Social Sciences , 3, 2007, pp. 106-116.