A Likert /ˈlɪkɜt/scale (devised in 1932 by organisational psychologist Rensis Likert) and also known as a ‘summative scale’, is a bipolar psychometric scale used in qualitative research to record responses along a range which captures intensity of feeling about attitudes to a given issue. It is regarded as a balanced method of data collection as it features an equal number of positive and negative responses, usually separated by a neutral response in mid-position. However, some researchers prefer to produce a ‘forced choice’ by omitting the middle option. Generally, five possible responses are set along a horizontal line (although some practitioners use as many as seven, or even nine, which gives more scope to respondents who like to avoid extremes!)
A typical five-item response line is:
Strongly disagree - Disagree - Neither agree nor disagree - Agree - Strongly agree
As well as levels of agreement, Likert scales can also be used to record other variable responses:
Very frequently - Frequently - Occasionally - Rarely - Never
Very important - important - moderately important - of little importance - Unimportant
Almost always true - Usually true - Occasionally trues - Usually not true - Almost never true
Strictly speaking, a Likert scale is the sum of responses to a number of statements (‘Likert items’) and refers to the range of potential scores. So, in a 5-point range like the one below, if scores of are distributed in a range of 1-5, the Likert scale is 5-25:
Strongly disagree 1
Neither agree nor disagree 3
Strongly agree 5
To report on a Likert scale, the values for each separate option should be summed and a score created for each respondent. Scores can then be used to create a chart showing the distribution of opinion across the target population. Scores are very often plotted and reported using diverging stacked bar charts (see Robbins & Heiberger 2011). For results to be meaningful, all the items selected should belong to a similar category, so that the summed score produces a reliable measurement of the particular behaviour or attitude being investigated.
The advantage of Likert scales is that they provide quantitative data about personal attitudes whilst allowing for degrees of opinion (or no opinion). Possible drawbacks are ‘central tendency bias’, where respondents avoid the extremes, ‘acquiescence bias’, where they simply agree with the statement presented, and ‘social desirability bias’, where they give the response that they think represents them in the most positive light. Another potential disadvantage is that few options are on offer, and respondents may not easily be able to align themselves with any of them. There may also be a problem within sets of items, whereby respondents are influenced by their own answers to earlier questions, either remaining consistent out of habit, or deliberately breaking the pattern. These issues can be resolved at the design stage by means of carefully designed and sequenced questions.
"Example Likert Scale" by Nicholas Smith http://commons.wikimedia.org/wiki/File:Example_Likert_Scale.svg#mediaviewer/File:Example_Likert_Scale.svg
Bertram, D. “Likert scales…are the meaning of life” http://poincare.matf.bg.ac.rs/~kristina/topic-dane-likert.pdf
Bryman, A. 2012. Social Research Methods. 4th ed Oxford: Oxford University Press
Denscombe, M. 2014 The Good Research Guide: For Small Scale Research Projects 2014Maidenhead: Open University Press
Likert, R. 1932. “A Technique for the Measurement of Attitudes” Archives of Psychology, No.140.
Robbins, N. B. & M. R. Heiberger. 2011. “Plotting Likert and Other Rating Scales” https://www.amstat.org/sections/SRMS/Proceedings/y2011/Files/300784_64164.pdf
Uebersax, J.S. “Likert scales: dispelling the confusion” http://www.john-uebersax.com/stat/likert.htm