Skip to content
ConductExperience
Menu

Likert scale questions and scoring

A Likert scale is a set of statements that people answer on one ordered scale, such as strongly disagree to strongly agree. Each answer counts as a number, and a person's numbers are added up, or averaged, into one score. Rensis Likert described the method in 1932.

One statement on such a scale is a Likert-type item. ConductExperience asks it as a rating scale, a grid or a single choice, and works out the score for you.

Try a Likert scale

Both questions are drawn by the screen a participant answers on. Answer them, and the table under each shows the row the download holds.

A rating scale: 7 points, a label at each end

The sign-up was easy to follow.
Strongly disagreeStrongly agree

This is the question as a participant sees it. Nothing you answer here is saved or sent.

The row it writes

One row of the answers table, without its first columns, which name the person and the times. A blank cell means not answered.

easyThe sign-up was easy to follow.
blank

A grid: 5 points, every point labeled, 3 statements, one score

How much do you agree with each statement?
I enjoy my first coffee of the day.
I could easily skip my morning coffee.
Coffee helps me feel ready for the day.

This is the question as a participant sees it. Nothing you answer here is saved or sent.

The row it writes

One row of the answers table, without its first columns, which name the person and the times. A blank cell means not answered.

coffee__1I enjoy my first coffee of the day.coffee__2I could easily skip my morning coffee.coffee__3Coffee helps me feel ready for the day.likingCoffee likingliking_bandCoffee liking band
blankblankblankblankblank

One item or a scale

  • An item is one statement and its answers. A scale is several items on the same answers, whose numbers are added up.
  • Likert's own scale gave each of five answers a number from 1 to 5, and a person's numbers were added up.
  • One item is enough to ask how easy something was. A scale suits something that no single item holds, such as an attitude.

Five points or seven

Five and seven are the usual numbers of points. More points give people more room to answer, and fewer are quicker. Use one number for the whole scale, and the same labels on every item. An odd number gives a middle answer, and an even number does not.

A rating scale in ConductExperience takes 3 to 11 points.

Labels at the ends or on every point

LabelsQuestion typeWhat it gives
At the endsRating scaleA rating of 3 to 11 points with a label at each end.
On every pointGrid or single choiceA grid: rows that share one set of labeled answers, one answer per row, up to 300 rows. A single choice: the labeled answers for one statement.

A label on every point lets everyone read the points the same way. End labels take less room.

Reverse an item and score the scale

  • Some statements point the other way, such as “I could easily skip my morning coffee.” Turn them upside down in the Scores section, so that a high score always means more of what you measure.
  • Turning a question upside down counts its lowest answer as its highest: on 5 points, 1 becomes 5 and 5 becomes 1.
  • The Scores section adds up or averages the questions you choose. In the example, 3 statements make Coffee liking, from 3 to 15, and 2 bands name where a score falls: Low (3 to 9) and High (10 to 15).
  • A score is blank until every item in it is answered. Participants never see a score. A score is a research measure, not a diagnosis.

What the download holds

Each questionnaire has an answers table with a row each time a person takes it. After the columns that name the person, the part and the time, it holds:

  • One column for each rating scale, named by its key, such as easy. It holds the number of the point chosen, from 1 up to the last point, or from 0 for a rating numbered from 0. A single choice has a column too, holding the value of its answer.
  • One column for each row of a grid, named <key>__<number>, such as coffee__1 for the first row. It holds the number of the answer chosen in that row.
  • One column for each score, named by its key, such as liking, and <key>_band for the band it falls in, such as liking_band.
  • A blank cell means the question was not answered or not shown. data_dictionary.csv says what every column holds, as the words under the column names in the example do.
The whole download

What Results shows

  • For a rating scale or a choice question: the count and share of each answer.
  • For a grid: the same for each row.
  • For a score: its mean, median, lowest and highest, and the share of people in each band.
  • Filtered by version, group, status and start date.
  • Excluded people are left out unless you include them, and one line gives the counts for the paper: started, finished, excluded by reason and analyzed.

Build a Likert scale survey

  1. Start a study on the survey template.
  2. Add a questionnaire part. Add a Rating scale for one statement, or a Grid for several statements that share their answers.
  3. Write the statements and the labels. For a rating scale, set the points, from 3 to 11.
  4. Open the Scores section. Add a score, choose its questions, and turn upside down the ones to reverse. Name bands if you want them.
  5. Preview the study as a participant, publish it and share the link.
The editor and its other question types

Limits

  • A rating scale takes 3 to 11 points. A grid takes up to 300 rows.
  • Results shows a single rating as counts and shares. A mean comes from a score.
  • A scale joins the library only when its holder's terms allow it. Any other scale you write or load needs your own right to use it.

Source

Likert, R. (1932). A technique for the measurement of attitudes. Archives of Psychology, 22(140), 1–55.

Related

Planning a study?

Start one now, or tell us what you want to measure.