CHAPTER 9

To get you started, we'll go through four popular statistical tests: chi-squared, t-test, correlation, and regression, as well as a method for extracting the data you need. My new Tauntaun dataset must be loaded into R for all of the analyses in this exercise. . Remember that we're working with data that has been "super-cleaned, combined, etc." and is saved as an.RDS file.Our first test, the chi-squared test, is a straightforward method for analyzing categorical results with any number of categories. For example, you can use this test to see if Tauntaun fur length (short, medium, long) and fur color (grey, white) are related (or dependent) on one another. That is to say, the probability of a specific Tauntaun fur color depends on the length of its fur; and vice versa, the likelihood of a particular fur color depends on the length of its fur.Now, the purpose of the chi-squared test is to determine if gender and party are INDEPENDENT of each other. The chi-squared test of independence will compare the observed numbers in box a (our data) with the numbers in box d.The key argument here is that there are several different chi-squared distributions. A single parameter called degrees of freedom determines the form of the distribution.R has many functions for working with different kinds of statistical distributions, and these functions generally start with the letter d, p, q, or r, followed by the distribution name. statistical tests work this same way (sometimes),Conduct a statistical test, which returns a test statistic (here, our observed test statistic was 30.1).Find where this test statistic falls on a specific theoretical distribution (here, we are looking at the theoretical chi-square distribution with 2 degrees of freedom).Draw inferences (conclusions) based on this information.The "frequentist method of hypothesis testing," as it is called in science, was the dominant method of statistical testing in the twentieth century.We call a value "unlikely" to have happened by chance alone if it occurs at the extreme end of a theoretical distribution, and we call it statistically important. We've already established that the chi-square distribution is made up of summed, squared random z scores.Working through the statistical research aid folders, learning the context of the data inputs and test outputs, and analyzing the test results were all covered in this chapter.i can generalize this process to work through most problems in R. And remember, R’s strength is in its open-source community.