Different types of data can give you different insights into your customers or what they think about your products and services. Some types of data help you categorize your respondents. Some types of data help you understand the feelings of your customers for your products. Other data gives us an idea of the gap between what you think your customers think about your products and what they actually think about your product. By the end of this lesson you'll be able to describe the importance of scales, describe the insights scales can provide, as well as the limitations, and list three important types of scales: nominal, ordinal, and interval. Now before we go into the [INAUDIBLE] of the data, it's important to recognize that the variables are measured on different scales. So the way we are going to analyze the question should account for the different types of scales. The type of analysis being descriptive, being inferential, being causal, should be based on the scale that you use in the questionnaire. Scales can be used to describe something. For example, whether or not you are likely to use something, or what is your gender, what is your age, what is your highest level of education and so on. The questions asked to get this answer are important in that they describe a particular entity. A question can also be measured to get a sense of rank ordering. So for example, what are the brands that you like the most? Can you rank various brands from most favorite to least favorite? What are the most important factors that influence your decision to buy a product? The last characteristic that is important is the idea of distance. For example, is your intention to buy a product higher or lower than someone in a different region. In other words, what is the distance between those two intentions to buy. And so those characteristics are useful to classify the data in one category or another and therefore it's important to understand which scale has been used. There are three types of scales that we are going to cover here, nominal scale, ordinal scale, and interval scale. Based on what we've just discussed, you can already see that these three characteristics emerge that will be used to describe scales. Can this scale be descriptive in some way, yes or no? Does a scale provide a sense of ordering between different options? And the last characteristic, does the scale provide a sense of distance between different answers? All right, let's look at some examples. The nominal scale is used to categorize or provide labels on different entities. One example could be, what is highest level of education, high school level, college, post-graduate, or which brand of toothpaste did you last purchase? So as you can see such scales are very good in describing and categorize something. However, there is really no sense of ordering, right? Which toothpaste brand you last purchased doesn't give you a sense of the preferences for brand A over brand B over brand C for example. It doesn't give you either a sense of distance between brand A and brand B, right? So it's the same thing with education level or gender. They're just used to describe customer and not used to provide a sense of rank ordering of distance between customers. That's where ordinal scales come into play. These scales are used to inform about the rank ordering of some entity. One example might be, rank the following four automotive brands in order of preference, Toyota, Honda, Ford, and Chevrolet. So these types of scales have two characteristics. First, they provide a description of an entity and/or they help to categorize different answers. The second characteristic of such scale is a sense of preferences through rank ordering. If I prefer Toyota over Honda, that's going to appear where the question has been asked. And this information would be useful for marketing decisions. However, ordinal scales do not provide a sense of distance. So when a respondent says, for example, I prefer Toyota over Honda, it does not really tell you by how much. Does this person prefer Toyota tens times more that Honda, five times more than Honda, two times more than Honda? You can't discern that from an ordinal scales, because a sense of distance doesn't exist. To get a sense of distance, you should use and analyze interval scales. Interval scales use numbers to rank objects or entities, or categorize those entities and provide a sense of distance or intensity between these items. One example might be measuring customer satisfaction. Could you rate United Alliance in terms of customers service from poor to very good? As you can see the distance between poor and very poor can take any values. Usually people use something called a Likert scale, usually from one to five where one is quite poor and five is very good. If someone says three, that means that this person is basically a yes slightly, but not so great. Someone else can decide that customer service United Airlines is going to be very good in which case this person is going to give a five. So contrary to nominal scales and ordinal scales, interval scales possess three characteristics. They allow someone to describe or categorize different entities. They allow someone to provide a sense of ordering between objects. And on top of that, they provide a sense of distance between those objects.