Chi-square Test for

Independence
Perform chi-square test for independence to determine the
significant relationship between the variables involved in the study
Chi-square Test for Independence
USED IF:

•The objective of the study is to identify if there is a

significant relationship between the variables

•The data is non-parametric

•The data does not assume a normal distribution

 Example:
Many parents allow their underage children to drink alcohol in limited situations
when an adult is present to supervise. The idea is that teens will learn
responsible drinking habits if they first experience alcohol in a controlled
environment. Other parents take a strict no-drinking approach with the idea
that they are sending a clear message about what is right and what is wrong.
Recent research, however, suggests that the more permissive approach may
actually result in more negative consequences (McMorris et al., 2011). In the
study, teens who were allowed to drink with their parents were significantly
more likely to experienced alcohol-related problems than teens who were not
allowed to drink. In an attempt to replicate this study, researchers surveyed a
sample of 150 students each year from ages 14-17. The students were asked
about their alcohol use and about alcohol-related problems such as binge
drinking, fights, and blackouts The results are shown in table on the next page.
Do the data show a significant relationship between the parents' rules about
alcohol and subsequent alcohol-related problems
 Step 1:
State the hypothesis & select a level of significance
•THE NULL HYPOTHESIS FOR THE CHI-SQUARE TEST FOR INDEPENDENCE STATES THAT THE TWO
VARIABLES BEING MEASURED ARE INDEPENDENT

•THIS GENERAL HYPOTHESIS CAN BE STATED IN TWO DIFFERENT WAYS

Version 1
Ho : In the general population, there is no relationship between parents' rules
in for alcohol use and the development of alcohol-related problems
Version 2
Ha: In the general population, the distribution of alcohol-related problems has
the same proportions for teens whose parents permit to drink and for those
whose parents do not

For this test we will use

a = .05.
 Step 2:
Determine the degree of freedom and locate the critical region

df= (C-1) (R-1)

df= (2-1) (2-1) = 1

 df= 1
a= .05
critical value for chi-square= 3.84
 Step 3:
Determine the expected frequencies and compute for the chi-square statistics

OBSERVED FREQUENCIES ( fo ) : THE ACTUAL FREQUENCY THAT IS OBTAINED

FROM THE EXPERIMENT/RESEARCH
EXPECTED FREQUENCIES ( fe ) : THE HYPOTHETICAL DISTRIBUTION THAT IS IN
PERFECT AGREEMENT WITH THE NULL HYPOTHESIS
•FORMULA FOR EXPECTED FREQUENCIES
160/200 OR 80%: DID NOT EXPERIENCE ALCOHOL-RELATED PROBLEMS
40/200 OR 20%: EXPERIENCED ALCOHOL-RELATED PROBLEMS

FOR 80 TEENS WHO WERE NOT ALLOWED TO DRINK, WE EXPECT:

80% OF 8O: 64 EXPECTED TO EXPERIENCE PROBLEMS
20% OF 8O: 16 EXPECTED NOT TO EXPERIENCE PROBLEMS

FOR 120 TEENS WHO WERE ALLOWED TO DRINK, WE EXPECT:

80% OF 12O: 96 EXPECTED TO EXPERIENCE PROBLEMS
20% OF 12O: 24 EXPECTED NOT TO EXPERIENCE PROBLEMS
•THE CHI-SQUARE TEST FORMULA
(observed frequencies) (expected frequencies)

(chi-square value)
 Step 4:
Make a decision regarding the null hypothesis and the outcome of the study

Chi-square value: 6. 381

Critical value: 3.84

The obtained chi-square value (6.381) exceeds the critical value (3.84). Therefore, the
decision is to reject the null hypothesis. In the literature, this would be reported as a
significant result with x 2(1, n = 200) = 6,381, p <.05. According to version 1 of this
means that we have decided that there is a significant relationship between parents' rules
about alcohol and sub sequent problems. Expressed in terms of version 2 of H, the data
show a significant differ ence in alcohol-related problems between teens whose parents
allowed drinking and those whose parents did not. To describe the details of the significant
result, you must compare the original data (1st table) with the expected frequencies (in
2nd table). Looking at the two tables, it should be clear that teens whose parents allowed
drinking experienced more alcohol-related problems than would be expected if the two
variables were independent
 Learning check
1. If a chi-square test for goodness of fit has df = 1
then the individuals are classified
into ____ categories and if a chi-square test for
independence has df = 1 then the individuals are
classified into ____ categories.

A.) 2,2
B.) 2,4
C.) 1,4
D.) 1.1
 Learning check
2. A sample of 100 people is classified by gender (male/female)
and by whether or
not they are registered voters. The sample consists of 60 females,
of whom 50 are registered voters, and 40 males, of whom 25 are
registered voters. If these data were used for a chi-square test for
independence, the expected frequency for registered males would be

A.) 15
B.) 25
C.) 30
D.) 45
 Learning check
3. A chi-square test for independence is conducted
for data from a 2 X 2 data matrix
with a total of n = 40 participants. If the data
produce x2= 4.00, then what decision should be
made?

A.) Reject with a = 0.5 but not with a = .01.

B.) Reject H, with a = 0.1 but not with a = .05.
C.) Reject with either a = 0.5 or a = .01.
D.) Fail to reject H, with a = 0.5 or with a = .01.
Chi-Square and the Pearson Correlation
Special Applications of the Chi-Square Tests
 Learning Objective

The similarities and differences between the Chi-square test for independence and the
Pearson Correlation

It used to test the relationship between two categorical variables.

There are 2 variables, and both are measured as categories, usually at the nominal
level.
correlation is about the linear relationship between two variables.
Chi-square is usually about the independence of two variables.
the similarities and the differences between the chi-square test for
independence and the independent-measures t test or ANOVA.

In statistics, an ANOVA is used to determine whether or not there is a

statistically significant difference between the means of three or more
independent groups.
Use Chi-Square Tests when every variable you’re working with is
categorical.
Use ANOVA when you have at least one categorical variable and one
continuous dependent variable.
 Chi-Square and the Pearson Correlation
The chi-square test for independence and the Pearson correlation
are both statistical techniques intended to evaluate the
relationship between two variables. The type of data obtained in a
research study determines which of these two statistical
procedures is appropriate. Suppose, for example, that a
researcher is interested in the relationship between self-esteem
and academic performance for 10-year-old children. If the
researcher obtained numerical scores for both variables, the
resulting data would be similar to the values shown in Table
17.11(a) and the researcher could use a Pearson correlation to
evaluate the relationship. On the other hand, if both variables
are classified into non-numerical categories as in Table 17.11(b),
then the data consist of frequencies and the relationship could be
evaluated with a chi-square test for independence
 ■ Chi-Square and the Independent-Measures t and ANOVA
Once again, consider a researcher investigating the relationship between self-esteem and
academic performance for 10-year-old children. This time, suppose the researcher
measured academic performance by simply classifying individuals into two categories,
high and low, and then obtained a numerical score for each individual’s self-esteem.
The resulting data would be similar to the scores in Table 17.12(a), and an independent-
measures t test would be used to evaluate the mean difference between the two groups
of scores. Alternatively, the researcher could measure self-esteem by classifying
individuals into three categories: high, medium, and low. If a numerical score is then
obtained for each individual’s academic performance, the resulting data would look like
the scores in Table 17.12(b), and an ANOVA would be used to evaluate the mean
differences among the three groups.
The point of these examples is that the chi-square test
for independence, the Pearson correlation, and tests for
mean differences can all be used to evaluate the
relationship between two variables. One main distinction
among the different statistical procedures is the form of
the data. However, another distinction is the
fundamental purpose of these different statistics. The
chi-square test and the tests for mean differences (t
and ANOVA) evaluate the significance of the relationship;
that is, they determine whether the relationship observed
in the sample provide
enough evidence to conclude that there is a corresponding relationship in
the population. You can also evaluate the significance of a Pearson
correlation, however, the main purpose of a correlation is to measure
the strength of the relationship. In particular, squaring the
correlation, r2 , provides a measure of effect size, describing the
proportion of variance in one variable that is accounted for by its
relationship with the other variable.
The Median Test for Independent Samples
The median test provides a nonparametric alternative to the independent-measures t test (or
ANOVA) to determine whether there are significant differences among two or more independent
samples. The null hypothesis for the median test states that the different samples come from
populations that share a common median (no differences). The alternative hypothesis states
that the samples come from populations that are different and do not share a common median.
The logic behind the median test is that whenever several different samples are selected from
the same population distribution, roughly half of the scores in each sample should be above the
population median and roughly half should be below. That is, all the separate samples should
be distributed around the same median. On the other hand, if the samples come from
populations with different medians, then the scores in some samples will be consistently higher
and the scores in other samples will be consistently lower
*Learning Check
1. Which of the following accurately describes the chi-square test for independence?

a. It is similar to a single-sample t test because it uses one sample to test a hypothesis about one
population.
b. It is similar to a correlation because it uses one sample to evaluate the relationship between two
variables.
c. It is similar to an independent-measures t test because it uses separate samples to evaluate the
difference between separate populations.
d. It is similar to both a correlation and an independent-measures t test because it can be used to
evaluate a relationship between variables or a difference between populations.
2. One version of the null hypothesis for the chi-square test for independence states
that there is no relationship between the two variables. With this version of the null
hypothesis, the chi-square test is most similar to what other statistical procedure?

a. An independent-measures t test.
b. A repeated-measures t test.
c. A single-sample t test.
d. A Pearson correlation.
 Free Resources
Use these free recolorable icons and illustrations in your Canva design.
 Free Resources
Use these free recolorable icons and illustrations in your Canva design.