Scribd
Upload
7 views19 pages

Chi Square X

Uploaded by

maexi3601
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
7 views19 pages

Chi Square X

Uploaded by

maexi3601
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 19

CHI-SQUARE (X2)

• The chi-square (Greek letter chi, X2) is the most


commonly used method of comparing
proportions.
• It is particularly useful in tests evaluating an
association/relationship between nominal or
ordinal data.
• Typical situations or settings are cases where
persons, events, or objects are grouped in two or
more nominal categories such as “Yes – No”
responses, “Favor-Against-Undecided”or “A, B, C,
or D”.
• Chi-square analysis compares the observed
frequencies of the response with the expected
frequencies.
• It is a measure of actual divergence of the
observed and expected frequencies.
• Formula:
• Ʃ(F0-Fe)2
• X= Fe

• Where: F0 = observed number of cases


• Fe = expected number of cases
Chi-square and Goodness of Fit: One Sample
Case

• There are research problems where responses fall


in any one of a number of categories.
• The data are expressed in frequencies and the
observed frequencies (F0) are compared to the
frequencies expected (Fe) on the basis of some
hypothesis
• If the differences between the observed and the
expected frequencies are small, X2 will be small.
• The greater the difference between the observed
and expected frequencies under the null
hypothesis, the greater or larger the X2 will be.
• If the difference between the observed and
expected values are so large collectively as to
occur by chance only, say 0.05 or less, when the
null hypothesis is true, then the null hypothesis is
rejected.
example

• Consider the nomination of three (3) presidential


candidates of a political party, A, B, and C.
• The chairman wonders whether or not they will
be equally popular among the members of the
party.
• From this the hypothesis of equal preference, a
random sample of 315 men were selected and
interviewed which one of the three candidates
they prefer.
• Results of the survey:

• Candidates Frequency
• A 98
• B 115
• C 102

• Are you going to reject the null hypothesis that


equal members of men in the party prefer each of
three candidates?
• Or are you going to accept the null hypothesis of
equality or preference?
Calculating the X2 value
• F0 Fe F0-Fe (F0-Fe)2
• A 98 105 -7 49
• B 115 105 10 100
• C 102 105 -3 9
• 158
• Ʃ(F0-Fe)2
• X2= Fe

• X2= 158/105
• X2= 1.5047619048 / 1.505
• X2 =1.505
• In order to test the significance of the
computed X2 value using a specified criterion
of significance, the obtained value is referred
to a table with appropriate degrees of
freedom which is equal to k-1,
• Where k is equal to the number of categories
of the variable.
• In this problem df = 3-1 = 2
• Therefore, for the X2 to be significant at the
0.05 level, the computed value should be
more than (>) the tabular value which is
5.991
summarizing
• Level of significance = 5%
• df = k-1 (number of categories minus 1)
• = 3-1
• =2
• Critical value: X2 (.05) = 5.991
• Decision rule: Reject H0 if X computed >
5.991, otherwise do not reject H0
• Conclusion:
• Since 1.505 < 5.991, do not reject H0.
• There is no sufficient evidence or reason to
reject the null hypothesis that the frequencies
in the population are equal.
Chi-square as a Test of Independence: Two
Variable Problem
• Chi-square can also be used to test the
significance of the relationship when data
are expressed in terms of frequencies of joint
occurrence.
• For this kind of problem, a two-way
contingency table with rows and columns are
set-up.
• In the case of two variable problems, the
expected frequencies are those predicted on
the independence of the two variables.
Formula:
• Fe = [(rowtotal)(columntotal)]
• N
• Example:
• Suppose, one wants to know if there is a
relationship between gender and school
choice. A sample 100 female and 100 male
freshman students were asked individually
for their school choice..
• Test the null hypothesis of no significant
relationship between the students’gender
and school choice at 5% level of
significance.
Formula:
• Fe = [(rowtotal)(columntotal)]
• N
• Male:
• Public = 65
• Private = 35

• Female:
• Public = 42
• Private= 58
Formula:
• Fe = [(rowtotal)(columntotal)]
• N

School Male Female Total


Choice
Public 65 C1 42 C3 107

Private 35 C2 58 C4 93

Total 100 100 200


Computational Procedure:
• 1. Compute the expected frequency of each
cell
• Fe = [(rowtotal)(columntotal)]
• N
• C1 = 100x107/200
• Fe = 53.5
• C2 = 100x93/200
• Fe = 46.5
• C3 = 100x107/200
• Fe= 53.5
• C4 = 100x93/200
• Fe= 46.5
Computational Procedure:
• 2. Present in table form
Fo Fe Fo-Fe (Fo-Fe)2 (Fo-Fe)2/
Fe

65 53.5 11.5 132.25 2.47

35 46.5 -11.5 132.25 2.84

42 53.5 -11.5 132.25 2.47

58 46.5 11.5 132.25 2.84

•
X2=10.62
Computational Procedure:
• 3. Prepare the hypothesis
• Null hypothesis: No significant relationship
exist between gender and school choice.
• Level of Significance: 5%
• Degrees of Freedom: (c-1)(r-1)
• c= no. of columns
• r = no. of rows
• df = (2-1)(2-1) = (1)(1) = 1
• CV = 3.841
• 4. Decision rule
• Since the computed value X2 = 10.62 is greater
than the critical value (3.841), we reject the null
hypothesis.
• Thus, the two variables of gender and school are
related: females tend to prefer private schools
while males tend to prefer to study in public
schools.
Thank You

Kingsoft Office
ake Presentation much more fun

You might also like