Data come in four varieties. In this video, learn how to distinguish these data types and how to use them.
- [Instructor] Let's turn our attention to different types of data. Data can appear in one of four types. We can also talk about four types of scales so I'll use data and scale interchangeably. One type is nominal in which numbers are just identifiers with no real meaning. Another is ordinal in which numbers begin to take on meaning, a bigger number means more of some property than a smaller number, but the exact difference between the numbers are meaningless. The third is interval. The differences between numbers are meaningful but the value of zero is arbitrary and has no real meaning. Last but not least, ratio. The numbers include a meaningful zero-point. The total absence of a measured property. Here's a closer look at nominal data. With nominal data, numbers are just names. A number identifies an individual but that's all. A great example is a number on a team jersey. It just identifies a player but that's all it does. You wouldn't include team jersey numbers in any arithmetic operations. That wouldn't make any sense. Nominal data is also called categorical data. Numbers can identify categories. If you have data in categories what does make sense is to talk about the amount of individuals in each category. Let's turn to ordinal data. An ordinal data, a number is more than just an identifier. With ordinal data, a number begins to mean something. A higher number, that means more of a particular property. A lower number, that means less of a property. An often used example is the Mohs scale of hardness of minerals. A lower number means less hardness, a greater number means more hardness. With this type of data however the differences between numbers don't mean anything. With interval data, the differences between numbers start to mean something. Because the data have equal interval units of measure. A common thermometers are a good example. A 10 degree difference is the same in any point on a thermometer. 20 degrees to 30 degrees is the same differences as 50 degrees to 60 degrees. What doesn't make sense is ratio statements. This is because zero degrees Fahrenheit is just an arbitrary value. It doesn't indicate a total lack of heat. So 50 degrees Fahrenheit is not twice as hot as 25 degrees Fahrenheit. With ratio data, you can make statements like twice as much as or half as much as. Those kinds of statements makes sense. A ratio scale includes the complete absence of a property. In other words a meaningful zero-point. Four inches is really twice as long as two inches. Continuing with the temperature example, one type of thermometer does have a meaningful zero-point. The Kelvin scale. That zero-point is called absolute zero. Physicist tell us that, that means no vibrational, molecular motion. The complete absence of heat. So on the Kelvin scale, 200 K is really twice as hot as 100 K. Here's a table that sums it all up. An a nominal scale numbers just provide identity. An ordinal scale does that but also numbers provide magnitude. On an interval scale, the numbers provide identity and magnitude, and can represent equal intervals. A ratio scale adds a meaningful zero-point so you can make meaningful statements like twice as much as and half as much as. Here's another. On a nominal scale, numbers are identifiers, arithmetic on numbers makes any sense, and you can work with frequencies of categories. An ordinal scale, a number signifies amount of a property, but really don't have units of measure, and the differences among numbers don't make sense. An interval scale provides units of measure, and the differences among numbers make sense, but there is no meaningful zero-point. A ratio scale has all the features of interval data, plus a meaningful zero-point, and statements like twice as much as makes sense.
- Using Excel functions and graphics
- Data types and variables
- Calculating probability
- Mean, median, and mode
- Calculating variability
- Organizing and graphing distributions
- Visualizing normal distributions
- Sampling distributions
- Making estimations
- Testing hypotheses: mean, z- and t-testing, and more
- Analyzing variance
- Performing repeated measure testing
- Regression testing
- Hypotheses testing with correlation