Scribd
Upload
26 views5 pages

Tests for Independence in Statistics

The document discusses tests for independence, including parametric and nonparametric tests, and provides examples for correlation tests and Spearman rank correlation. It explains how to calculate test statistics for correlation and independence using contingency tables, along with the necessary formulas. Additionally, it illustrates the process of determining expected frequencies and calculating the chi-square test statistic to assess independence between two categorical variables.

Uploaded by

Julia Rolin
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
26 views5 pages

Tests for Independence in Statistics

The document discusses tests for independence, including parametric and nonparametric tests, and provides examples for correlation tests and Spearman rank correlation. It explains how to calculate test statistics for correlation and independence using contingency tables, along with the necessary formulas. Additionally, it illustrates the process of determining expected frequencies and calculating the chi-square test statistic to assess independence between two categorical variables.

Uploaded by

Julia Rolin
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

Quantitative Methods

Tests for Independence

Tests for Independence

Parametric and Nonparametric Tests


Parametric tests are based on assumptions about population
distributions and population parameters (e.g., mean = 3,
variance = 124.5).
Nonparametric tests make few, if any, assumptions about the
population distribution and test things other than parameter
values (e.g., runs tests, rank correlation tests).

1
© Kaplan, Inc.

1
Tests for Independence

Test Statistics: Correlation


Test of whether the population correlation coefficient equals zero.

r n2
t-stat = with n – 2 degrees of freedom,
1 r2
where:
r = sample correlation coefficient

2
© Kaplan, Inc.

Tests for Independence

Test of Correlation: Example


Sample correlation = 0.35, sample size = 42
0.35 42  2
t-stat = = 2.363
1  0.352
Level of Significance for Two-Tailed Test Reject H0:
because 2.363 > 2.021
df 0.10 0.05
38 1.686 2.024 Conclude that
correlation is not
39 1.685 2.023 equal to zero
40 1.684 2.021

3 -4
© Kaplan, Inc.

2
Tests for Independence

Nonparametric Test: Correlation


Spearman rank correlation test: whether two sets of ranks
are correlated
n
6 di2
Rank correlation r  1  i1
n(n  1)
2

where: n = sample size


di = difference between two ranks

If n > 30, may use t-table with n – 2 degrees of freedom


4 -1
© Kaplan, Inc.

Tests for Independence

Spearman Rank Correlation: Example


Candidate IQ Mock Rank IQ Rank di di2
score % mock
1 100 65 1 2 -1 1
2 120 80 4 4 0 0
3 104 71 2 3 -1 1
4 105 59 3 1 2 4
6
n
6 di2 66 36
r  1  1  0.4
r  1 i 1
4(4  1)
2
60
n(n  1)
2

5 -6
© Kaplan, Inc.

3
Tests for Independence

Tests of Independence: Example


Contingency Table: Observed Values
Earnings Dividend Yield
Growth Low Medium High Total
Low 28 53 42 123
Medium 42 32 39 113
High 49 25 14 88
Total 119 110 95 324

Determine what the table values should be if the two categories are
independent: expected observations in cell i,j =
total for row i × total for column j
total for all columns and rows 6
© Kaplan, Inc.

Tests for Independence

Tests of Independence: Solution


Expected Frequencies If Earnings Growth and
Dividend Yield Are Independent
Earnings Dividend Yield
Growth Low Medium High Total
Low 45.2 41.8 36.1 123
Medium 41.5 38.4 33.1 113
High 32.3 29.9 25.8 88
Total 119 110 95 324
Observed frequency in cell i,j
 Oi,j  Ei,j 
2
r c
χ 2  
Test statistic: Expected frequency
in cell i,j
i=1 j=1 Ei,j
© Kaplan, Inc. Degrees of freedom = [(r – 1)(c – 1)] 7 -2

4
Tests for Independence

Tests of Independence: Solution


(Observed – Expected)2 / Expected
Earnings Dividend Yield
Growth Low Medium High Total
Low 6.545 3.001 0.964 10.510
Medium 0.006 1.067 1.052 2.125
High 8.634 0.803 5.397 14.834
Total 27.469
Degrees of freedom = [(3 – 1)(3 – 1)] = 4 Test statistic

Critical value from chi-square table (df = 4) = 9.488


27.469 > 9.488 → reject hypothesis that earnings growth and
© Kaplan, Inc.
dividend yield are independent 8

You might also like