Our Benchmarks

  • Precise always accurate analysis
  • Professional expert statistical solutions
  • Custom only customized approach
  • Affordable low prices&special discounts

We Guarantee

Ordinal and Nominal Data Coding Variables

Most reports of polls in the news media describe some form of behavior or opinion, and these will include both ordinal and nominal data points. They generally feature reports on how people view some particular proposal or issue. Studies focus on the “what” sorts of questions of social science research. They all feature what the American public tends to think about some kind of issue. In all cases the variables these studies generate can be coded. It just takes a moment to figure out the best way to do so.

statistical data analysis schema

 Coding Variables for Polls

Nominal variables are those that featured unordered categories that just feature some particular choice. Religion, which is V145 in the dataset, is coded in the following manner in the software:

  • Mainline Protestant
  • Evangelical Protestant
  • Catholic
  • Jewish
  • Other
  • None
  • N/A

The coding of this could go in any other. Catholics could be 1 and Jews could be 2, and the basic information would never change since the numbers assigned to it don’t actually rank anything. They’re merely added for convenience, and don’t put one religious belief over another.

Ordinal scales have some sort of underlying order that requires users to maintain it at all times. For instance, everyone is familiar with the dataset V137 question regarding school vouchers. This uses a scale as follows:

  1. Strongly Favor
  2. Favor
  3. Oppose
  4.  Strongly Oppose
  5. N/A

When coding variables like this, it can be extremely important to keep these in the same order that they’ve always been included in. Messing around with them could invite problems into the whole picture. There’s a third category that needs to be coded as well, and possibly even more depending on the type of data being worked with and the particular aims of the research study in question.

coding variables

Interval level scales have some sort of underlying order in which individual intervals are the same. Pretend that years of education were being coded. In this case, a score of 10 would mean that the respondent had 10 years of formal education. However, the score of the values matters just as much as the order because each year adds one increment to the pre-existing level of education. That’s why ordinal and nominal data doesn’t work in these kinds of scenarios.

Different Types of Data Points

Those who often code nominal variables will find even other situations they might come across when performing a statistical data analysis. Relationships are often spoken of in terms of dependent and independent variables. Independent variables cause dependent variables to move, but they themselves are not caused by anything that the study concerns itself with. Most researchers start with a dependent variable and find independents to explain why that dependent is doing whatever its doing. These variables can be interval based, nominal or ordinal.

You can also learn how to use minitab by browsing our website.

Get help with coding variables today!