The Guttman Scale: Definition and Basics

The Guttman scale – An overview

The Guttman scale was developed during the 20^th century to be used in education research. It was named after the American mathematician, sociologist, and Professor of Social and Psychological Assessment.

He initially developed the scale to determine how students faired in an examination. His idea was to find a way to predict the number of questions the students got right in a paper just from the final score.

He began by arranging the questions in an ascending order based on the level of difficulty. He counted down the list until they got the final score. This score determined the number of questions they got right.

Guttman then applied this hypothesis about the scale to social sciences, leading to developing a unidimensional, cumulative scaling technique used when a continuum of respondent opinions.

Using this scale, the scientist can infer based on the cumulative weight of the data collected. The surveyor predicts the number of statements the respondent agreed or disagreed with.

For example, if a respondent scores a 3 on a Guttman scale, the respondent agreed with 3 the first three statements. Similarly, if they score a 10 on a 10-point Guttman scale, they agreed with all 10.

In addition to being unidimensional and deterministic, the Guttman scaling technique includes questions or statements that are ordinal in nature. This means that they are arranged from the minimum to the maximum in terms of importance.

The number of times the respondent agrees to statements, the more they agree with a statement.

Guttman scale examples

The Guttman scale is typically designed so that if the respondents agree with the first few statements and then disagree with one, it’s likely they’ll disagree with the remaining one in the order.

Let’s consider an example derived from the Bogardus Social Distance Scale with statements focusing on people’s opinions on immigrants.

The options provided are dichotomous (yes/no):

Using this version of the Guttman scale, you can determine your chosen population’s opinions on immigrants’ social contact.

The Guttman scale can also be structured in the following way. In this example, let’s consider the study is researching people’s opinions on dining in Izakaya-style Japanese restaurants.

Please check statements that you agree with:

• I like to eat out
• I like going to restaurants
• I like going to restaurants serving international cuisines
• I like going out to restaurants serving Japanese-style food
• I like going out to restaurants serving Izakaya-style Japanese restaurants

If you’ve noticed in both examples, we’ve listed the statements in a way that their specificity increases gradually. This makes determining how extreme a view is easier.

Why use the Guttman scale?

One of the main reasons market research experts use the Guttman scale instead of other cumulative scales is that it is hierarchically structured. The way statements are included in the survey doesn’t impact the survey’s productivity. If the respondent stops answering questions after a certain statement, it can be assumed that it was because they disagreed with it. And subsequently, they stop agreeing with subsequent statements or questions.

The Guttman scale is also more intuitive than other scaling techniques based on its ability to determine the extremeness of people’s opinions on certain topics. The scale also allows the surveyors to quantify qualitative data by assigning a score and ranking the responses.

Developing a Guttman scale for your survey

While it might sound difficult, developing the Guttman scale for your survey is considerably easier.

Define the focus of your research

Just like any other scaling technique, you begin by defining the focus of your research statement. If you want to develop a cumulative scale that determines people’s opinion on immigrants, you need to define:

• Whether the type of immigrants come through legal or illegal ways;
• Where the immigrants are coming from;
• And the reason for immigration.

The more particular you are in your survey topic, the more granular the results of your research.

Develop survey items

Survey items refer to statements or questions you plan to ask respondents. You can either choose to develop the scale over dichotomous yes/no or agree/disagree statements or the format illustrated in the second example mentioned above.

The scaling method then requires you to develop a large set of statements that reflect your topic. Typically, you would develop at least 80 to 100 statements that you think would be related to immigration.

Rating the items

To determine whether the items you’ve developed work with the survey’s focus, you can engage a panel or group with subject matter experts. This panel can determine how favorable the statements are.

Scoring the items

The best way to simplify the Guttman scale’s result analysis is by assigning a score to each statement. Assign a score of 1 to “yes/agree” and 0 to “no/disagree.”

Once you’ve finalized the scale items, all that’s left is focusing on your target population, choosing a sampling methodology, and administering the survey.

If you are conducting online research, you might be using:

• a list for your sample frame or,
• leverage a market research panel provider

Analyzing the Guttman scale

To analyze the survey’s results, you have to record the responses and construct a table or matrix. To develop a cumulative scale out of the responses, list respondents that agreed to more statements at the top and those that agreed to fewer statements at the bottom.

For respondent data with the same number of agreements, sort the statement with the most agreements on the left side and do gown from there. This is an example of a subset of items.

To develop other subsets, you need to undertake a scalogram analysis. Other statistical techniques for developing a cumulative technique can be used to determine how good the scale is and assign final scores to the respondents.’

The ones with a higher score are likely more in support of the opinion you’re surveying, while respondents with average scores have more supportive opinions.