Maximum Difference Scaling involves the researcher making the participants of a survey choose between two items in order of importance. In this article, we will explore when, why, and how to use the MaxDiff questioning technique.

What is a Max Diff Survey?

Maximum Difference Scaling (MaxDiff analysis) is defined as an analytic approach used by researchers to measure the preference score of their survey participants on different items. MaxDiff which is also known as the best-worst scaling is similar in features with conjoint analysis. 

In this survey scaling method, the participants are asked to select the most and least important factors in the given options in a set by the researcher. Hence, it is better to add an easy question where you can get the important features in your survey. 

MaxDiff is a survey that can help you understand what your audience wants. It’s a great way to ask questions and gather data. The MaxDiff method, first developed by the UK government, asks respondents to rank items in order of preference in a matrix of comparisons. For example, you could ask customers to compare two or three different prices for a product or service.

Read: Survey Methods: Definition, Types, and Examples

A MaxDiff survey is particularly effective when the options are complex or abstract, or when you want to find out how much people value different attributes of an experience or product. It’s also ideal for asking about innovative products or services that don’t yet exist.

The MaxDiff survey allows respondents to choose from multiple choices in a grid format. Respondents are asked to pick their preferred option from each pair of choices and then are ranked overall at the end.

MaxDiff is a simple and easy-to-use tool. It is very comprehensive when you’re analyzing critical research situations because it involves analyzing the ‘Best’ and ‘Worst’ scales in a given set with no consideration for any other data.

Standard MaxDiff Terms

There are two Standard MaxDiff terms – they are attributes and set. Before we move to the uses of MaxDiff survey questions, let’s review them first.

Attribute: Attribute is an individual or single statement that researchers like to rank in a set of items.

Set: Set is a group of attributes. The number of characteristics and size present in the set displayed to the participants are decided by the researchers.

So, standard MaxDiff terms are the set of attributes displayed to the participants by the researchers. The participants are expected to pick the most and the least items in the features displayed to them.

Read: Survey Data Collection: Methods, Types, and Analysis

Uses of MaxDiff

MaxDiff survey is commonly used by researchers in customer satisfaction, product features, and brand preferences. MaxDiff analysis is used to measure possible outcomes. Other ways in which MaxDiff survey is used are

• Risk Management
• Asset management
• Security Policy
• Cloud computing
• Access control
• Privacy
• Organizational Security
• HR Security
• Physical and Environmental Security
• Communications and Operations Management
• Documentation
• Incident Event and Communications Management
• Compliance
• Business Continuity and Disaster Recovery

Read: The 4,5, and 7 Point Likert Scale + [Questionnaire Examples]

Examples of Maxdiff Surveys

Let’s consider these examples to properly understand how MaxDiff surveys work.

Example 1:

For example, let’s say you’re running a MaxDiff survey to figure out which features are most important for customers when deciding whether or not to buy a certain type of smartphone. If you ask your participants “Which features do you prefer?” and most of them can name at least one feature, then it suggests that customers consider this feature when making this type of purchase. E.g

• Long-lasting battery
• Great camera
• Warp-speed processing
• Crystal-clear display
• Multiple windows
• Plenty of storage space
• Infrared remote control

Try to limit the number of features to three to eight for your participants to choose from and for you to easily generate the best data. 

Read: Guttman Scale: Definition, Interpretation, Examples

Example 2:

if you want to test out brands of coffee versus tea, one question would ask a participant to rate how much they like the smell of coffee versus the smell of tea. The second question could ask how many cups of each beverage a participant drinks per day on average. By combining these two pieces of information into one question, you can get a better idea of how much coffee is consumed per day and how much people enjoy the smell vs. how much tea is consumed and how much people enjoy the taste.

When to Use MaxDiff & Why

The primary purpose of MaxDiff is to be used in identifying preferences among given items. MaxDiff is preferably used for single-level preference data while Conjoint analysis is the best option when gathering multilevel preference data. 

This is because a MaxDiff reduces the number of choices the participants have to make. Therefore MaxDiff is most suitable if you want to determine the value of two attributes.

MaxDiff has been proven to provide statistically significant results for even the most difficult of questions, such as “Which car should I buy?” or “Which presidential candidate do you support?” MaxDiff is particularly useful for determining the relative value of something. For example, which of these three flavors of ice cream do you like best?

Identifying preferences among small differences, such as 2% milk versus whole milk. If you are torn between 2% and whole milk, try answering this question: Which would you choose if you had to drink 2% for the rest of your life?

Read: Thurstone Scale: Definition, Examples, Pros & Cons

You can also make use of MaxDiff when you do not have a large marketing budget. For example, if you are planning a vacation trip, you can decide activities based on the most important attribute to your guests. Would your guests rather go on local wildlife or shopping? 

Rating scales often fail in this area as they are not good measures of preferences. So if what you’re looking for as a researcher is a tradeoff between various service or product attributes, MaxDiff is your best bet.

MaxDiff also easily gathers discrete preferences from a large set of attributes by asking the participants which attribute is most important to them. Therefore make use of MaxDiff if you want to understand the level of importance in items such as benefits, menu, features, etc. 

Now here are some reasons why MaxDiff is a great research tool that you can consider.

• Efficient Setup: MaxDiff is suitable for evaluating participants’ choices of the most and least preferred in a subset of 4 or 5 attributes.
• Eliminate Bias: MaxDiff compulsorily ensures that a participant picks what they consider as the least important and most important. This helps the researcher collect data that are truly relevant and important.
• Statistical Modeling: MaxDiff statistical models are created by the output of a MaxDiff question, unlike other question types. These models make it possible to quantify or measure the preferences of the participants and easily understand what they truly value.

Read: What is a Survey Rating Scale? + [Types & Question Examples]

MaxDiff Analysis Vs Standard Rating Scale Questions

Studies revealed that rating questions can have user scale bias, scale meaning bias, and lack of discrimination. This is in addition to other ranking questions problems such as difficulty in evaluation by the participants, flawed data rejecting ties, and limitation in testing number of items. 

These have led researchers to see that participants engage in response strategies when presented with constant sum questions, to make the evaluating of all the items easier. This is why researchers prefer to use MaxDiff surveys. 

While the MaxDiff survey will show the differences or bias between items, and the responses for them, the standard rating scale wouldn’t. MaxDiff survey eliminates the issue of bias as participants are asked to choose their preference, unlike the standard rating scale.

Data from the participants are placed on a standard scale of ‘0’ to ‘100’ points and summed up to a total of 100 in the MaxDiff survey unlike in a standard rating scale.

Maxdiff Analysis vs Conjoint Analysis

MaxDiff and conjoint analysis are both techniques that researchers and marketers can use to determine how to best position products in their market. Both techniques have been shown to be highly effective at helping companies understand where to focus their marketing efforts and resources.

The conjoint analysis involves the use of a product description in market research, followed by asking respondents which one of the multiple versions of the product they would most likely buy. Researchers can then compare the responses from different groups of respondents to determine which version of the product is most likely to sell. 

This technique allows companies to determine how much they need to increase sales, as well as understand how changes to their product will affect potential sales. MaxDiff analysis also uses a product description, but instead of asking respondents which product version they would buy, it asks them which two products they prefer out of a set of products. 

This allows researchers and marketers to narrow down choices for consumers, making it easier for them to identify their preferred choice. The MaxDiff survey has been shown to be highly effective at helping companies understand consumer preferences for large numbers of similar products in a targeted market segment. 

In addition, MaxDiff makes it possible to compare things that are not necessarily comparable. For example, if you were looking for customers’ preferences for candy bars and you wanted to know which ones they preferred the most, you could create two groups: one with chocolate-flavored options and another with non-chocolate options and ask customers to choose their favorite from each group. 

You would then compare the answers from these two groups and see how much stronger people’s preference was for chocolate bars versus non-chocolate bars.

How to Create a MaxDiff Survey

To create your own MaxDiff survey, first list all possible features for your product or service. Researchers must first determine how many choices need to be presented (typically three) and then create an even number of items (usually two). Then create two groups that include the best features and worst features.

You can create the MaxDiff survey form on Formplus using the form builder. If you do not have an account, quickly sign up and gain access to over 1000 preset survey templates.

Once you’re logged into your account, use the form builder features to create your MaxDiff form. It’s quite easy to do.

How To Analyze MaxDiff Data

The MaxDiff system is used in marketing research to measure preference for products or services. There are four types of analysis that are offered in the MaxDiff system:

• Counting Analysis (for basic, topline results): Counting Analysis provides basic, topline results. It is the easiest way to use MaxDiff, but it is less accurate than Individual-Level Score Estimation.
• Individual-Level Score Estimation (recommended): Individual-Level Score Estimation uses the results from Counting Analysis to run a regression analysis that outputs individual-level scores. This is more accurate than Counting Analysis, and it gives user respondents the opportunity to provide more nuanced responses.
• Aggregate Score Estimation via Logit (advanced): Aggregate Score Estimation via Logit allows the researcher to estimate aggregate scores for items using a logistic regression model (advanced). It provides an estimation of scale values for each item. This is useful for identifying which items are most important without any consideration of individual-level responses. While this method is more accurate than Counting Analysis or Individual-Level Score Estimation, it requires more time and resources.
• Latent Class Estimation (advanced): Latent Class Estimation uses a latent class model to estimate latent classes of respondents. This type of analysis allows for richer segmentation of respondents into meaningful groups. It is useful for understanding patterns in the data that may not be fully understood using other rating scale analyses.

Conclusion

The MaxDiff questions can be much more nuanced and targeted, which is the reason they’re used by so many businesses to elevate above the competition.

Typically, you won’t hear of a consumer product with MaxDiff questions because it only lends itself well to business-to-business sales. If you’re trying to gain traction for yourself as an entrepreneur and have a new company, we recommend using MaxDiff as your survey questions to target a niche audience and get answers that are directly related to your product or service.

Additionally, if you have an existing consumer product line but want to know what features will drive incremental customer purchases, we recommend using MaxDiff questions for your survey.