© aSup -2007 Statistics II – SPECIAL CORRELATION 1 SPECIAL CORRELATION
© aSup -2007
Statistics II – SPECIAL CORRELATION
2
The SPEARMAN Correlation The Pearson correlation specially
measures the degree of linear relationship between two variables
Other correlation measures have been developed for nonlinear relationship and of other types of data
One of these useful measures is called the Spearman correlation
© aSup -2007
Statistics II – SPECIAL CORRELATION
3
The SPEARMAN Correlation Measure the relationship between
variables measured on an ordinal scale of measurement
The reason that the Spearman correlation measures consistency, rather than form, comes from a simple observation: when two variables are consistently related, their ranks will be linearly related
© aSup -2007
Statistics II – SPECIAL CORRELATION
4
INTRODUCTION
Pearson product-moment coefficient is the standard index of the amount of correlation between two variables, and we prefer it whenever its use is possible and convenient.
But there are data to which this kind of correlation method cannot be applied, and there are instances in which can be applied but in which, for practical purpose, other procedures are more expedient
© aSup -2007
Statistics II – SPECIAL CORRELATION
5
Person X Y Rank X Rank Y
A
B
C
D
E
3
4
8
10
13
12
10
11
9
3
1
2
3
4
5
5
3
4
2
1
rs =SP
√(SSx) (SSy)
SP = ΣXY(ΣX)(ΣY)
n
SSX = ΣX2(ΣX)2
n
XY
5
6
12
8
5
© aSup -2007
Statistics II – SPECIAL CORRELATION
6
The COMPUTATION
1. Rank the individual in the (two) variables
2. For every pair of rank (for each individual), determine the difference (d) in the two ranks
3. Square each d to find d2
© aSup -2007
Statistics II – SPECIAL CORRELATION
7
Person X Y Rank X Rank Y
A
B
C
D
E
3
4
8
10
13
12
10
11
9
3
1
2
3
4
5
5
3
4
2
1
rs = 1 -6 Σ D2
n(n2 – 1)
D2
16
1
1
4
16
© aSup -2007
Statistics II – SPECIAL CORRELATION
8
Spearman’s Rank-Difference Correlation Method
Especially, when samples are small It can be applied as a quick substitute
when the number of pairs, or N, is less than 30
It should be applied when the data are already in terms of rank orders rather than interval measurement
© aSup -2007
Statistics II – SPECIAL CORRELATION
9
INTERPRETATION OF A RANK DIFFERENCE COEFFICIENT
The rho coefficient is closely to the Pearson r that would be computed from the original measurement.
The rρ values are systematically a bit lower than the corresponding Pearson-r values, but the maximum difference, which occurs when both coefficient are near .50
© aSup -2007
Statistics II – SPECIAL CORRELATION
10
To measure the relationship between anxiety level and test performance, a psychologist obtains a sample of n = 6 college students from introductory statistics course. The students are asked to come to the laboratory 15 minutes before the final exam. In the lab, the psychologist records psychological measure of anxiety (heart rate, skin resistance, blood pressure, etc) for each student. In addition, the psychologist obtains the exam score for each student.
LEARNING CHECK
© aSup -2007
Statistics II – SPECIAL CORRELATION
11
Student
Anxiety
Rating
Exam Scores
A 5 80
B 2 88
C 7 80
D 7 79
E 4 86
F 5 85
Compute the Pearson and Spearman correlation for the following data.
Test the correlation with α = .05
© aSup -2007
Statistics II – SPECIAL CORRELATION
12
The BISERIAL Coefficient of Correlation
The biserial r is especially designed for the situation in which both of the variables correlated are continuously measurable, BUT one of the two is for some reason reduced to two categories
This reduction to two categories may be a consequence of the only way in which the data can be obtained, as, for example, when one variable is whether or not a student passes or fails a certain standard
© aSup -2007
Statistics II – SPECIAL CORRELATION
13
The COMPUTATION The principle upon which the formula
for biserial r is based is that with zero correlation
There would no difference means for the continuous variable, and the larger the difference between means, the larger the correlation
© aSup -2007
Statistics II – SPECIAL CORRELATION
14
AN EVALUATION OF THE BISERIAL r Before computing rb, of course we need to
dichotomize each Y distribution. In adopting a division point, it is well to
come as near the median as possible, why? In all these special instances, however, we
are not relieve of the responsibility of defending the assumption of the normal population distribution of Y
It may seem contradictory to suggest that when the obtained Y distribution is skewed, we resort the biserial r, but note that is the sample distribution that is skewed and the population distribution that must be assumed to be normal
© aSup -2007
Statistics II – SPECIAL CORRELATION
15
THE BISERIAL r IS LESS RELIABLE THAN THE PEARSON r
Whenever there is a real choices between computing a pearson r or a Biserial r, however, one should favor the former, unless the sample is very large and computation time is an important consideration
The standard error for a biserial r is considerably larger than that for a Pearson r derived from the same sample
© aSup -2007
Statistics II – SPECIAL CORRELATION
16
The POINT BISERIAL Coefficient of Correlation
When one of the two variables in a correlation problem is genuine dichotomy, the appropriate type of coefficient to use is point biserial r
Examples of genuine dichotomies are male vs female, being a farmer vs not being a farmer
Bimodal or other peculiar distributions, although not representating entirely discrete categories, are sufficiently discontinuous to call for the point biserial rather than biserial r
© aSup -2007
Statistics II – SPECIAL CORRELATION
17
The COMPUTATION A product-moment r could be
computed with Pearson’s basic formula If rpbi were computed from data that
actually justified the use of rb, the coefficient computed would be markly smaller than rb obtained from the same data
rb is √pq/y times as large as rpbi
© aSup -2007
Statistics II – SPECIAL CORRELATION
18
POINT-BISERIAL vs BISERIAL When the dichotomous variable is
normally distributed without reasonable doubt, it is recommended that rb be computed and interpreted
If there is little doubt that the distribution is a genuine dichotomy, rpbi should be computed and interpreted
When in doubt, the rpbi is probably the safer choice
© aSup -2007
Statistics II – SPECIAL CORRELATION
19
TETRACHORIC CORRELATION
A tetrachoric r is computed from data in which both X and Y have been reduced artificially to two categories
Under the appropriate condition it gives a coefficient that is numerically equivalent to a Pearson r and may be regard as an approximation to it
© aSup -2007
Statistics II – SPECIAL CORRELATION
20
TETRACHORIC CORRELATION The tetrachoric r requires that both X and Y
represent continuous, normally distributed, and linearly related variables
The tetrachoric r is less reliable than the Pearson r.
It is more reliable whena. N is large, as is true of all statisticb. rt is large, as is true of other r’sc. the division in the two categories are near the medians
© aSup -2007
Statistics II – SPECIAL CORRELATION
21
THE Phi COEFFICIENT rФ related to the chi square from 2 x 2 table
When two distributions correlated are genuinely dichotomous– when the two classes are separated by real gap between them, and previously discussed correlational method do not apply– we may resort to the phi coefficient
This coefficient was designed for so-called point distributions, which implies that the two classes have two point values and merely represent some qualitative attribute
© aSup -2007
Statistics II – SPECIAL CORRELATION
22
DEFINITION
A partial correlation between two variables is one that nullifies the effects of a third variable (or a number of other variables) upon both the variables being correlated
© aSup -2007
Statistics II – SPECIAL CORRELATION
23
EXAMPLE
The correlation between height and weight of boys in a group where age is permitted to vary would be higher than the correlation between height and weight in a group at constant age
The reason is obvious. Because certain boys are older, they are both heavier and taller. Age is a factor that enhances the strength of correspondence between height and weight
© aSup -2007
Statistics II – SPECIAL CORRELATION
24
THE GENERAL FORMULA
r12.3 =r12 – r13r23
√ (1 – r213)(1 – r2
23)
When only one variable is held constant, we speak of a first-order partial correlation
© aSup -2007
Statistics II – SPECIAL CORRELATION
25
SECOND ORDER PARTIAL r
r12.34 =R12.3 – r14.3r24.3
√ (1 – r214.3)(1 – r2
24.3)
When only one variable is held constant, we speak of a first-order partial correlation
© aSup -2007
Statistics II – SPECIAL CORRELATION
26
THE BISERIAL CORRELATION
WhereMp = mean of X values for the higher group in the
dichotomized variable, the one having ability on which sample is divided into two subgroups
Mq = mean of X values for the lower groupp = proportion of cases in the higher groupq = proportion of cases in the higher groupY = ordinate of the unit normal-distribution curve at
the point of division between segments containing p and q proportion of the cases
St = standard deviation of the total sample in the continously measured variable X
rb =Mp – Mq
St
Xpqy
© aSup -2007
Statistics II – SPECIAL CORRELATION
27
THE POINT BISERIAL CORRELATION
WhereMp = mean of X values for the higher group in the
dichotomized variable, the one having ability on which sample is divided into two subgroups
Mq = mean of X values for the lower groupp = proportion of cases in the higher groupq = proportion of cases in the higher groupSt = standard deviation of the total sample in the
continously measured variable X
rpbi =Mp – Mq
St
pq
© aSup -2007
Statistics II – SPECIAL CORRELATION
28
THE TETRACHORIC CORRELATION
rcos-pi =ad - bc
yy’N2
© aSup -2007
Statistics II – SPECIAL CORRELATION
29
THE GENERAL FORMULA
r12.3 =r12 – r13r23
√ (1 – r213)(1 – r2
23)
When only one variable is held constant, we speak of a first-order partial correlation