Scribd
Upload
40 views2 pages

Chapter 14: Nonparametric Statistics and The Chi-Square Test of Independence

Stat

Uploaded by

Nene
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
40 views2 pages

Chapter 14: Nonparametric Statistics and The Chi-Square Test of Independence

Stat

Uploaded by

Nene
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Chapter 14: Nonparametric Statistics and the Chi-Square Test of Independence

I. Overview

a. Most statistical procedures in this book assume that the data to be analyzed are normally
distributed.

b. In the real world, distributions of data are often not normally distributed, so there is
another class of statistics called nonparametric statistics.
i. Do not rely on the assumption of a normal distribution.

c. There are a wide variety of nonparametric statistics. In this chapter, we focus on one:
The chi-square test of independence.

II. Chi-Square Test of Independence in Depth

a. Purpose: To determine whether cases are distributed among the categories of two
variables in the numbers one would expect by chance alone.

b. Types of variables: Two categorical variables


i. e.g., gender and level of mathematics class enrolled in.

c. Procedure: A contingency table is produced that shows how many cases in each group on
one variable fall into each of the categories on the second variable. These numbers are
then compared to the number of cases one would expect to see in each category by
chance alone.

i. e.g., Suppose that in one high school, there are 100 students in eleventh grade.
Half of these students are girls, the other half are boys. Half of these 100
students are taking advanced math, and half are taking basic math.

1. Based on the number of boys and girls in the sample, and the number of
students in each level of math, we would expect there to be 25 students in
each cell of this contingency table, by chance alone. So the numbers in
the table below are the expected values.

Basic math Advanced math

Girl 25 25

Boy 25 25

ii. Now suppose that I examine the actual number of cases that appear in each cell
of the table. The table below contains the observed number of students in each
cell:

Basic math Advanced math

Girl 15 35
Boy 35 15

iii. As we can see, the observed values in the cells look quite different from the
expected values. There are more girls and fewer boys than we would expect by
chance alone in the advanced math and vice-versa in the basic math.

iv. Now we need to determine whether the difference between the observed and
expected values is statistically significant.

1. To do this, we calculate the χ2 value by subtracting the expected value


from the observed value in each cell, squaring that, and dividing by the
expected value for the cell. I follow this procedure for each cell and then
add up these values for each cell. This produces the observed chi-square
value. The formula is as follows:

χ2 =

2. Once we have calculated the χ2 value we look in Appendix E to


determine whether the value is statistically significant.

a. To use Appendix E, use (R – 1)(C – 1) to calculate the degrees of


freedom.

i. R represents the number of rows in the contingency


table, C represents the number of columns.\

v. Interpret the results.

1. If the χ2 value that you calculated is statistically significant, this tells you
that the observed values in your contingency table differ from the
expected values. It does not tell you exactly what is causing this
difference, but you can usually get a pretty good idea by looking at the
values in your contingency table.

III. Summary

a. Distributions of data are not always normally distributed.

b. There are several nonparametric statistics researchers can use when their data are not
normally distributed.

c. One of these is the chi-square test of independence (χ2).


i. This statistic is used to determine whether expected frequencies differ from
observed frequencies on two categorical variables.

You might also like