meeting 15 Associative Hypothesis Test
5/17/2011
Basic Principles
Associative Hypothesis Test:
Correlation And Regression
Reduction
Population
parameter
ρ
Presented by:
Mahendra AN
Statistic
r
Generalization = correlation from
between two or more variables
Basic Concepts
Basic Concepts
r=0
Positive correlation
r=0.5
r=1
Negative correlation
Statistics Techniques
Parametric Statistics for Correlations
Data Types
Correlation techniques
1. Product moment
Nominal
1. Contingency coefficient
Ordinal
1. Rank Spearman
2. Tau Kendal
Interval or ratio
1. Product moment Pearson
2. Multiple correlation
2
3. Partial correlation
z For searching of correlation
between two variables for
interval or ratio data from the
same source and normal
distributed
r
xy
=
∑ xy
∑x y
2
2. Multiple correlation
Vs partial
correlation
2
z Alternative significance test
t=
r n−2
1− r
2
© Mahendra Adhi Nugroho, M.Sc, Accounting Program Study of Yogyakarta State University
For internal use only!
1
5/17/2011
Parametric Statistics for Correlations
Correlation Coefficient Interpretation
Multiple correlation
Coefficient Interval
Correlation
0.00 – 0.199
Very low
2
R
y.x1x2
=
Partial correlation
1
1
1
1−r x x
1
2
1 2
2
R
y.x1x2
1
2
1−r x x
2
1
)
−
1 2
1−r y x
2
2
2
z Alternative
significance test
2
R /k
1 − R 2 / (n − k − 1)
(
=
2
z Alternative
significance test
Fh =
r yx − r yx .rx x
2
r yx + r yx − 2r yx r yx rx x
t=
r
p
0.02 – 0.399
Low
0.40 – 0.599
Medium
0.60 – 0.799
Strong
0.80 – 1.000
Very strong
n−3
1− r p
2
Non parametric statistics for correlations
Non parametric statistics for correlations
1. Contingency coefficient
2. Rank Spearman
3. Tau Kendal
z For nominal data
z Strong connection with
chi square
• For ordinal data from
different data sources
and not normal
di t ib t d
distributed
z For ordinal or rank data
z For more than 10 data
z Can be used for
searching partial
correlation
χ
C=
2
2
N +χ
ρ = 1−
2
(OPij + Eij )
2
r
k
χ = ∑∑
2
i =1 j =1
EP
ij
6∑ b i
n(n2 − 1)
• Alternative significance
test
ρ
n−2
zh = 1
t=r
1− r2
n −1
Regression
τ=∑
A−∑B
N ( N − 1)
2
z Significance test
z=
τ
2(2 N + 5)
9 N ( N − 1) )
Linear Regression Lines
z Correlation search direction and strength
of symmetric , causal or reciprocal
relationships between two variables while
regression predict changes of dependent
variables when independent variables are
change
© Mahendra Adhi Nugroho, M.Sc, Accounting Program Study of Yogyakarta State University
For internal use only!
2
5/17/2011
Linear Regression
1. Simple Linear
Regression
2. Multiple linear
Regression
Y =a+b1X1+b2X2 +...+bnXn +ε
Y = a + bX + ε
b=r
Simple Linear Regression Line
s a = Y − bX
s
y
x
(∑Y )⎛⎜⎝ ∑ X i ⎞⎟⎠ − (∑ X )(∑ X Y )
n∑ X i − (∑ X i )
2
a=
b=
i
i
2
i
i
2
n∑ X i Y i − (∑ X i ) − (∑Y i )
(
n∑ X i − ∑
2
X i)
2
Simple Linear Regression Test Steps
2
z Linearity test F = s
s
z Count a and b
z Build regression equation and draw
regression
i liline
z Significance test F = s
s
z Correlation hypothesis test
TC
2
G
Kita membangun langit dengan segala daya yang ada
kemudian merasa langit kita sudah cukup tinggi
bahkan mengungguli langit lain.
Tetapi akan selau ada langit di atas kita.
2
reg
2
== MAHENDRA
ADHI NUGROHO,
2008 ==
sis
r=
n∑ X i Y i − (∑ X i )(∑ Y i )
(n∑ X i −(∑ X i) ) (n∑ Y i −(∑ Y i ) )
2
2
2
2
© Mahendra Adhi Nugroho, M.Sc, Accounting Program Study of Yogyakarta State University
For internal use only!
3
Basic Principles
Associative Hypothesis Test:
Correlation And Regression
Reduction
Population
parameter
ρ
Presented by:
Mahendra AN
Statistic
r
Generalization = correlation from
between two or more variables
Basic Concepts
Basic Concepts
r=0
Positive correlation
r=0.5
r=1
Negative correlation
Statistics Techniques
Parametric Statistics for Correlations
Data Types
Correlation techniques
1. Product moment
Nominal
1. Contingency coefficient
Ordinal
1. Rank Spearman
2. Tau Kendal
Interval or ratio
1. Product moment Pearson
2. Multiple correlation
2
3. Partial correlation
z For searching of correlation
between two variables for
interval or ratio data from the
same source and normal
distributed
r
xy
=
∑ xy
∑x y
2
2. Multiple correlation
Vs partial
correlation
2
z Alternative significance test
t=
r n−2
1− r
2
© Mahendra Adhi Nugroho, M.Sc, Accounting Program Study of Yogyakarta State University
For internal use only!
1
5/17/2011
Parametric Statistics for Correlations
Correlation Coefficient Interpretation
Multiple correlation
Coefficient Interval
Correlation
0.00 – 0.199
Very low
2
R
y.x1x2
=
Partial correlation
1
1
1
1−r x x
1
2
1 2
2
R
y.x1x2
1
2
1−r x x
2
1
)
−
1 2
1−r y x
2
2
2
z Alternative
significance test
2
R /k
1 − R 2 / (n − k − 1)
(
=
2
z Alternative
significance test
Fh =
r yx − r yx .rx x
2
r yx + r yx − 2r yx r yx rx x
t=
r
p
0.02 – 0.399
Low
0.40 – 0.599
Medium
0.60 – 0.799
Strong
0.80 – 1.000
Very strong
n−3
1− r p
2
Non parametric statistics for correlations
Non parametric statistics for correlations
1. Contingency coefficient
2. Rank Spearman
3. Tau Kendal
z For nominal data
z Strong connection with
chi square
• For ordinal data from
different data sources
and not normal
di t ib t d
distributed
z For ordinal or rank data
z For more than 10 data
z Can be used for
searching partial
correlation
χ
C=
2
2
N +χ
ρ = 1−
2
(OPij + Eij )
2
r
k
χ = ∑∑
2
i =1 j =1
EP
ij
6∑ b i
n(n2 − 1)
• Alternative significance
test
ρ
n−2
zh = 1
t=r
1− r2
n −1
Regression
τ=∑
A−∑B
N ( N − 1)
2
z Significance test
z=
τ
2(2 N + 5)
9 N ( N − 1) )
Linear Regression Lines
z Correlation search direction and strength
of symmetric , causal or reciprocal
relationships between two variables while
regression predict changes of dependent
variables when independent variables are
change
© Mahendra Adhi Nugroho, M.Sc, Accounting Program Study of Yogyakarta State University
For internal use only!
2
5/17/2011
Linear Regression
1. Simple Linear
Regression
2. Multiple linear
Regression
Y =a+b1X1+b2X2 +...+bnXn +ε
Y = a + bX + ε
b=r
Simple Linear Regression Line
s a = Y − bX
s
y
x
(∑Y )⎛⎜⎝ ∑ X i ⎞⎟⎠ − (∑ X )(∑ X Y )
n∑ X i − (∑ X i )
2
a=
b=
i
i
2
i
i
2
n∑ X i Y i − (∑ X i ) − (∑Y i )
(
n∑ X i − ∑
2
X i)
2
Simple Linear Regression Test Steps
2
z Linearity test F = s
s
z Count a and b
z Build regression equation and draw
regression
i liline
z Significance test F = s
s
z Correlation hypothesis test
TC
2
G
Kita membangun langit dengan segala daya yang ada
kemudian merasa langit kita sudah cukup tinggi
bahkan mengungguli langit lain.
Tetapi akan selau ada langit di atas kita.
2
reg
2
== MAHENDRA
ADHI NUGROHO,
2008 ==
sis
r=
n∑ X i Y i − (∑ X i )(∑ Y i )
(n∑ X i −(∑ X i) ) (n∑ Y i −(∑ Y i ) )
2
2
2
2
© Mahendra Adhi Nugroho, M.Sc, Accounting Program Study of Yogyakarta State University
For internal use only!
3