Market researchers use comparative and non-comparative scales to collect data that is valid and reliable. This lesson will characterize these scales and discuss how they can be used for different purposes.
Airlines & Comparative Scales
”We know you have a choice when you fly, and we appreciate you choosing XYZ airlines for your travel today.”
If you’ve been on a commercial flight in the last 25 years, you’ve probably heard that message from the pilot or a flight attendant. The acknowledgement that you have a choice when flying is an example of what marketing researchers call a comparative scale. Comparative scales are directly competing products or services. When collecting this type of market research, study subjects are presented with two or more options and asked to choose the one they prefer most.
Types of Comparative Scales
In cases where a company needs more detail, researchers sometimes use paired comparison scales, asking customers to select their preference from items presented two at a time according to specific criteria. Since it would be a bit ridiculous to ask a customer to fly airline A from Paris to New York, and then airline B on the same route, comparative scaling is effective only with less expensive and more manageable stimulus objects.
A better option for collecting airline preference data would be a rank order scale. This scale type is commonly used to measure how customers feel about a particular brand or brand attribute. To measure this, researchers place a group of stimulus objects in front of a customer, give them criteria, and ask them to rank the items. In the airline preference example, researchers might place the names of 7 airlines in front of a customer and ask them to order the list based on their perception of the speed of the check-in process.
The primary limitation of a rank order is that it only measures a single variable, for example the speed of the check-in process. Since an airline knows that check-in speed is only one of many reasons a customer might choose it, a constant sum scale can help researchers account for more than a single attribute. In this type of scale, customers are given a limited quantity of something of value (like points, chips, or money) and asked to distribute it between objects based on certain criteria. In the airline preference example, a market researcher might give each study subject 50 chips and ask them to assign each airline brand a certain number of chips based on their preference.
Salsa & Non-Comparative Scales
If you’ve shopped at warehouse giant Costco, you might have observed a business practice that is relatively unique to their industry. As customers walk through the store, they’ll likely encounter a half-dozen or more associates handing out free samples of various food items in an effort to capture business. However, their sample tray usually contains only one brand. Customers aren’t being asked to choose one brand over another. Instead, they’re trying to capture customers on their own merits alone rather than introducing a competing product into the mix. Non-comparative scales are the inverse of comparative scales. In a non-comparative scale, study subjects are asked to evaluate only one product or service.
In the Costco example, we identified food samples distributed to customers as a non-comparative scale. Like comparative scales, there are several sub-types of non-comparative scales. In a non-continuous scale known as a Likert scale, a customer is asked to view a statement and then indicate their agreement or disagreement with the statement. In the Costco example, a researcher distributing food samples might ask a customer to try a particular brand of salsa and subsequently answer whether they agree or disagree with the statement, ”XYZ Salsa is mild in flavor.”
A similar non-continuous scale is the semantic differential scale. Semantic scales replace the agree/disagree scale with two significantly contrasting descriptors like hot/cold or intense/mellow. Applied to the Costco example, a customer might be asked to try the salsa and then mark on a 7-point scale whether the salsa was hot or mild.
A less frequently used non-comparative, non-continuous scale is the staple scale. In this type of scale, a researcher places a single descriptive word in the center of a scale containing even numbers of either a positive or negative integer. In the Costco food sampling example, a researcher might ask a customer to sample the salsa and then view a scale with the word ”spicy” in the center of a scale with a +/- range of 2, 4, 6, 8, 10.
Trusting the Numbers
All of these scales are designed with one goal in mind: help researchers collect valid and reliable data. In research, valid means that the data being collected is actually relevant and genuinely indicative of the entity being measured. A reliable study is a study whose findings are indeed indicative of the population as a whole and not merely the cohort. In other words, a study involving the taste of salsa is only reliable if it can be proven that, even though the number of people studied was only 72, the results are applicable to a much larger related population.
In market research, data is collected and measured on either a comparative scale or a non-comparative scale. A comparative scale asks customers to evaluate one product in direct comparison to others. A non-comparative scale evaluates a single product by itself. Paired comparison scales provide two items at a time and ask the customer to choose one over the other. In a rank order scale, subjects are asked to place a group of items in the order of preference. Constant sum scales are created by giving study subjects a limited number of point, money, or chips and asking them to distribute them between the products based on certain criteria.
Non-comparative scales include several types of non-continuous scales. Likert Scales ask customers to read a statement and then indicate that they agree/disagree. A semantic differential scale is very similar, but instead of agree/disagree, the semantic scale uses polar opposite adjectives like hot/cold or wet/dry. In the staple scale, a descriptive term is placed in the middle of a scale marked by positive or negative even numbers.
Functioning properly, all market research should be both reliable and valid. Reliable data is data that is consistent in collection and reporting. Valid data is data that actually reflects the data point that the researcher is attempting to collect.