You are on page 1of 19

CHI-SQUARE TEST

CHAPTER 8
Introduction

• There are two types of chi-square test:


– Goodness of fit test
– Independence test (not covered by our syllabus)

• There are two types of Goodness of Fit Test:


– Equal expected frequencies
– Unequal expected frequencies
• Test Statistics:

 2

O  E
2

• O -- observed frequency in a particular category.


• E -- expected frequency in a particular category.
• Critical Value:

 ,k 1
2

– k is the number of categories.


Equal Expected Frequencies

• To compare an observed distribution to an


expected distribution.

• Refer to Page 222, Example 3.


Refer to Page 222, Example 3.

• Test that the number of books


borrowed is uniformly (equal Day No. of
Books
by proportion) distributed over
Monday 1325
the week. Tuesday 1250
Wednesday 1115
Thursday 1305
Friday 1100
Saturday 1455
TOTAL 7550
STEP 1 : HYPOTHESES

• H0 : There is no difference in the number of


books borrowed by day over the week.

• H1 : There is a difference in the number of


books borrowed by day over the week.
STEP 2 : TEST STATISTICS
Day No. of Books (O) E (O-E)2/E
Monday 1325 1258.33 3.5324
Tuesday 1250 1258.33 0.0551
Wednesday 1115 1258.33 16.326
Thursday 1305 1258.33 1.7309
Friday 1100 1258.33 19.9224
Saturday 1455 1258.33 30.7384
TOTAL 7550 1258.33 72.3052

E
7550 • Test value = 72.3052
5
 1258.33
STEP 3 : CRITICAL VALUE

• Critical Value:


df = k – 1
2
0.05, 6 1 k = 6  no. of days

 2
0.05, 5  11.0705
STEP 4 : DECISION

• If Test Value > Critical Value  Reject H0.

• Test Statistics = 72.3052


• Critical Value = 11.0705

• Decision : Reject H0.


STEP 5 : CONCLUSION

• There is a difference in the number of books


borrowed over the week.
Unequal Expected Frequencies

• Refer Page 224, Q5.


• From experience: 62% highly satisfied, 20%
satisfied, 8% dissatisfied and 10% highly
dissatisfied.

• From 500 new customers: 320 highly satisfied,


80 satisfied, 50 dissatisfied and 100 highly
dissatisfied.
STEP 1 : HYPOTHESES

• H0 : The percentage of customer’s satisfaction


level are still the same in the following year.

• H1 : The percentage of customer’s satisfaction


level has changed in the following year.
STEP 2 : TEST STATISTICS

• To find E :
– Highly satisfied: E = 62% x 500 = 310
– Satisfied : E = 20% x 500 = 100
– Dissatisfied : E = 8% x 500 = 40
– Highly dissatisfied : E = 10% x 500 = 50
STEP 2 : TEST STATISTICS

Education O E (O – E)2/E
Highly 320 310 0.3226
Satisfied
Satisfied 80 100 4.0
Dissatisfied 50 40 2.5
Highly 100 50 50
Dissatisfied
Total 56.8226
STEP 3 : CRITICAL VALUE

• Critical Value:

 2
0.05, 4 1

 2
0.05, 3  7.81473
STEP 4 : DECISION

• If Test Value > Critical Value  Reject H0.

• Test Statistics = 56.8226


• Critical Value = 7.81473

• Decision : Reject H0.


STEP 5 : CONCLUSION

• The percentage of customer’s satisfaction


level has changed in the following year.
The End. Thank you

You might also like