Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Kandungan
Hypothesis Testing
Comparing two variables using Parametric Tests
‐ Two Independent Samples
‐ Two Related Samples
Comparing two variables using Non‐Parametric Tests
‐ Two Independent Samples
‐ Two Related Samples
Comparing more than two variables using Parametric Tests
‐ Testing for More than Two Independent Samples
‐ Testing for More than Two Related Samples
Comparing more than two variables using Non‐Parametric Tests
‐ Testing for More than Two Independent Samples
‐ Testing for More than Two Related Samples
Chi Square
Correlation and Regression
• Hypothesis testing is a
technique to help determine
whether a specific treatment
has an effect on the
individuals in a population.
MPPU1034 MPPU1034
3 4
Hypothesis Testing
MPPU1034 MPPU1034
5 6
Hypothesis Testing (cont.)
MPPU1034 MPPU1034
7 8
Hypothesis Testing (cont.)
MPPU1034 MPPU1034
9 10
Example 2:
Dependent
Ho2 : Tidak terdapat perbezaan ranking pemilihan universiti antara
Independent
negeri.
Ho3 : Tidak terdapat perbezaan skor yang signifikan antara Ujian Fizik
Ho : Tidak terdapat perbezaan / hubungan dan Ujian Kimia
yang signifikan antara XXXX dan YYY.
dependent dependent
MPPU1034 MPPU1034
MPPU1034 MPPU1034
13 14
Step 1
15 16
Step 2
MPPU1034 MPPU1034
17 18
Step 3 Step 4
If mean difference is large enough to be significant
Compute the test statistic. in the critical region, we conclude that the REJECT
difference is significant or that the the null hypothesis
The test statistic (in this chapter a z-score) forms a treatment has a significant effect. In this
ratio case we reject the null hypothesis.
19 20
Learning Check 1 Learning Check 1
An automobile manufacturer claims that a newly introduced model will µ=45 and σ=2. Find Z?
average µ=45 miles per gallon with σ =2. – M=42 for s sample of n=4 ?
A sample of n=4 cars is tested and averages only M=42 miles per =
gallon. =
Is this sample mean likely to occur if = 1
the manufacturer’s claim is true?
(Assume that the distribution of the millage
?
Is this sample mean likely to
occur if the manufacturer’s
scores is normal and significant level α=0.05) claim is true?
MPPU1034 MPPU1034
21 22
Some Examples of Null Hypothesis
H0 : 1 = 2
Some Examples of H0 : =
NULL HYPOTHESIS The three population means are equal to each other
H0 : 1 = 2 = 2
MPPU1034 MPPU1034
23 24
Eg. 1:
Directionality
one-tailed two-tailed
MPPU1034 MPPU1034
25 26
• Just because the sample mean (following treatment) is • Because the hypothesis test relies on sample data, and
different from the original population mean does not because sample data are not completely reliable, there is
necessarily indicate that the treatment has caused a
change. always the risk that misleading data will cause the
• You should recall that there usually is some discrepancy hypothesis test to reach a wrong conclusion.
between a sample mean and the population mean simply • Two types of error are possible.
as a result of sampling error.
MPPU1034 MPPU1034
27 28
MPPU1034 MPPU1034
29 30
Directional Tests Two Types of Hypotheses
The Research Hypothesis or
Alternative Hypothesis (H1)
• When a research study predicts a specific direction for the The hypothesis which states that the
treatment effect (increase or decrease), it is possible to The Null Hypothesis (Ho)
difference between the sample means
incorporate the directional prediction into the hypothesis The hypothesis which states or between the sample mean and the
test. that there is no real difference population mean is real (i.e. the
between the sample means or difference does exist!)
• The result is called a directional test or a one-tailed test. between the sample mean and
H1 : 1 2
A directional test includes the directional prediction in the the population mean
or
statement of the hypotheses and in the location of the H0 : 1 = 2
critical region. H1 : 1 < 2
or
H1 : 1 > 2
Strong Notes: H0 and H1 may involve statistical indexes other than means (refer to examples)!!
MPPU1034 MPPU1034
31 32
Measuring Effect Size
Some Examples of Alternative Hypothesis
Running time in the 100-metre dash is lower for men than women
• A hypothesis test evaluates the statistical significance of
H1 : males < females
the results from a research study.
• That is, the test determines whether or not it is likely that
Practice improves test scores the obtained sample mean occurred without any
H1 : pg > cg contribution from a treatment effect.
• The hypothesis test is influenced not only by the size of the
The relationship between income (X) and education (Y) for treatment effect but also by the size of the sample.
Adsians is different from those live in European communities • Thus, even a very small effect can be significant if it is
H1 : XYAsians XYEuropeans observed in a very large sample.
MPPU1034 MPPU1034
33 34
Measuring Effect Size
MPPU1034 MPPU1034
35 36
Power of a Hypothesis Test Power of a Hypothesis Test
MPPU1034 MPPU1034
37 38
MPPU1034 MPPU1034
39 40
Power of a Hypothesis Test
Interpretation
1. Cohen’s d = 0.2 ( small effect)
2. Cohen’s d = 0.5 ( medium effect)
Cohen’s d
Percentage of variance accounted , 3. Cohen’s d = 0.8 ( Large effect)
MPPU1034 MPPU1034
41 42
How to write hypothesis?
MPPU1034 MPPU1034
43 44
test your knowledge.. test your knowledge..
We always test a null hypothesis against an
1 alternative. The alternative hypothesis states that there is no
3 difference/no effect.
A. True
B. False A. True
ANSWER: A B. False
4 If our sample value is close to the null/population value, Type I error occurs when we
we conclude that
5
A. correctly fail to reject a false null hypothesis.
A. we can neither accept or reject the null. B. correctly reject a false null hypothesis.
B. something happened in the study; there is a C. incorrectly reject a false null hypothesis.
significant effect. D. incorrectly reject a true null hypothesis.
C. nothing happened in the study; there is no effect. ANSWER: D
D. something happened in the study, but the effect is
very small.
MPPU1034 MPPU1034
MPPU1034 MPPU1034
49 50
Example: Example:
1
2
MPPU1034 MPPU1034
51
MPPU1034
1 2
Parametric VS Non-Parametric
Mean Median
Pearson Spearman
3 4
One Sample T‐Test Measurement scales
Dependent Sample T‐Test Normality
Independent Sample T‐Test Homogeneity of variance
The t Test
One Sample T‐Test
- Known
- Compare with sample
vs 119,108,134,140,98,87,88, 97, ……
̅ 110 /
Populasi
/ Adakah terdapat
perbezaan halaju sampel
berbanding populasi?
7 8
7 8
The Estimated Standard Error and
The t Test For One Sample
the t Statistic
• The hypothesis test allows researchers to evaluate the
mean difference between known population mean
(variance remain unknown) using the data from a single
sample.
11 12
11 12
t = 2.5
4. Make decision
(b) ???
Example : Manual
Mean, M 71.3
n= 10 Standard Deviation, s 15.853
67, 54, 78, 95, 56, 78, 44, 78, 76, 87 Standard Error, SM 5.013
Standard Error ?
15 16
15 16
Result Result
1. State the hypothesis 2. Find the critical value with α = 0.05 and df= n ‐ 1
H0 : = 65 df = 9
H1 : 65 From Table
‐ 2.262 + 2.262
Result Result
3. Compute t 4. Make decision
Fail to Reject hypothesis Null
t and
Compute, t = 71.3 – 65
71.3 – 65
‐ 2.262 + 2.262
5.013
5.013
= 1.256
1.256
19 20
19 20
Test/ Computer..
Using SPSS for population mean of 65 at alpha
0.05
Spss view
‐4.677
Result : Reject H0
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
23 24
23 24
SPSS
The t Test For
One Sample
27 28
27 28
Independent Sample T‐Test
47%
53% Male Dependent Sample T‐Test
Female
3 4
7 8
Two Independent Samples Two Independent Samples
2. Find the critical value with α = 0.05 and df= df1 + df2
df = df1 + df2 = (8-1) + (5-1) = 7 + 4
df = 11
71.3 – 65 + 2.201
5.013
t = 2.117
71.3 – 65 2.201
5.013
2.117
11 12
SPSS…
= 9-6 (2.117)2
=
2.4494922 + 2.507132
(2.117)2 + 11
2
= 0.289
= 1.216
15 16
cth
1. Cohen’s d = 0.2 ( small effect)
2. Cohen’s d = 0.5 ( medium effect) SPSS
3. Cohen’s d = 0.8 ( Large effect)
t-TEST
1. r=0.01 (Small effect) Independent sample
2. r=0.09 (Medium effect)
3. r=0.25 (Large effect)
• Jantina
– Independent
• Markah Fizik
– dependent
• Markah Math
– dependent
19 20
Independent Samples Test
Levene's Test for Equality of
Variances t‐test for Equality of Means
95% Confidence Interval of
UJIAN FIZIK Equal variances
.575 .456 ‐.406 23 .688 ‐3.567 8.779 ‐21.728 14.595
assumed
Equal variances not
‐.412 20.347 .685 ‐3.567 8.658 ‐21.607 14.474
assumed
UJIAN MATEMATIK Equal variances
.236 .631 .561 23 .580 5.233 9.332 ‐14.072 24.538
assumed
Equal variances not
.577 21.187 .570 5.233 9.076 ‐13.630 24.097
assumed
3 4
Hypothesis tests with the Hypothesis tests with the
repeated‐measures repeated‐measures
The repeated‐measures t statistic allows
researchers to test a hypothesis about the In this situation it is possible to compute a
population mean difference between two difference score for each individual:
treatment conditions using sample data from a difference score = D = X2 – X1
repeated‐measures research study.
Where X1 is the person’s score in the first
treatment and X2 is the score in the second
treatment.
Or
Where X1 is the person’s score in the pre‐test
(before treatment) and X2 is the score in the post‐
test (after treatment).
Repeated‐Measures Repeated‐measures
H0: μD = 0
Repeated‐measures Repeated‐measures
In words, the null hypothesis says that there is no
consistent or systematic difference between the two
treatment conditions.
11 12
Hypothesis Tests with the Example Of Repeated‐measures
Repeated‐Measures Designs
• Effect size for the independent‐measures t is The following data demonstrate these points:
measured in the same way that we measured
Subject X1 X2 D
effect size for the single‐sample t and the
independent‐measures t. A 9 16 ?
• Specifically, you can compute an estimate of B 25 28 ?
Cohen’s d to obtain a standardized measure of the
mean difference, or you can compute r2 to obtain a C 31 36 ?
measure of the percentage of variance accounted D 58 61 ?
for by the treatment effect.
E 72 79 ?
Samples Samples
1
2. Find the critical value with α = 0.05 and df= n‐1
4. Make decision
Samples Samples
df = 4
From Table
SS = 141 – 252 SMD = 0.894
5
‐ 2.776 + 2.776 SS = 16 t= 5–0
0.894
MD t = 5.593
‐ 2.776 + 2.776
5.593
4 Make decision..
Reject hypothesis Null
‐ 2.776 + 2.776
5.593
19 SPSS 20
Example of Repeated‐ The t Test for Two Related Samples
Measures Designs
• Notice that the original data show big differences
from one treatment to another. 1
4
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
21 22
Pastikan anda dapat menterjemah
setiap nilai yang ditunjukkan…
Kesimpulan?
1 p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
Ambil
2
3 Perhatian p < α Ho ditolak, Terdapat Perbezaan yg signifikan
p = Sig.
• p = 0.005
4 5 6 • p < α
11 • @ p < 0.05 , maka Ho ditolak.
Terdapat perbezaan yang
9
7
8 10
signifikan antara skor X1 dan
X2
23
Mean Median
Pearson Spearman
Parametric : Data
Interval/ Independent t‐test Mann‐Whitney test
ratio
Normal One‐way independent ANOVA Kruskal‐Wallis test
distribution Non Parametric
Paired t‐test Data : Ordinal/Wilcoxon test
Not Normal
One‐way repeated ANOVA distribution Friedman’s test
3 4
Hipotesis ..
1 = Lelaki
2 = perempuan
Latihan1Q3.sav
7 8
1 = Lelaki
6 2 = perempuan
7 9
2
8
3
Output Kesimpulan?
p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
2
1 3 p < α Ho ditolak, Terdapat Perbezaan yg signifikan
p = Asymp.Sig.
Ambil
Bagi s1;
Perhatian • p = 0.772
• p > α
• @ p > 0.05 , maka Ho GAGAL
4
ditolak. Tidak terdapat
perbezaan yang signifikan
antara pilihan yang dibuat.
5
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
11
3 4
1
Contoh
Kajian dilakukan
untuk melihat
adakah terdapat
perbezaan pilihan
bagi calon yang
memilih UTM 2
berdasarkan pilihan
3
mereka sebelum dan
selepas taklimat
diberikan.
4
Latihan1Q5.sav
5
3 4 5
7 5
Ambil
Perhatian
6 6
9
7
7 8
Kesimpulan?
p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
p < α Ho ditolak, Terdapat Perbezaan yg signifikan
p = Asymp Sig.
• p = 0.008
• p < α
• @ p < 0.05 , maka Ho ditolak.
Terdapat perbezaan yang
signifikan antara pilihan
sebelum dan selepas taklimat
ANOVA
Analysis of Variance
3 4
The Logic and the Process of
Analysis of Variance
• ANOVA can be used in situations where there
are two or more means being • ANOVA allows researcher to evaluate all
compared, whereas the t tests are limited to
only two means are involved.
ANOVA vs t-test
of the mean differences in a single
• These situations would require a series of hypothesis test using a single α‐level
several t tests to evaluate all of the mean and, thereby, keeps the risk of a Type I
differences. (t test can compare only 2 means error under control no matter how many
at a time.) different means are being compared.
• Although each t test can be done with a specific
α‐level (risk of Type I error), the α‐levels
accumulate over a series of tests so that
the final experimentwise α‐level can be quite
large
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
5 6
7 8
The structure and sequence of calculations for the ANOVA.
11 12
Example Steps
One experiment call examining learning performance 1. State the hypothesis
under three temperature condition (GroupA1, GroupA2,
Ho: µ1= µ2= µ3
Group A3). The dependent variable is the number of
problem solved correctly. H1: At least one of the treatment means is different.
Temperature conditions 2. Find the critical value with α = 0.05 and F(between, within)
Is there any
A1 ‐ 50o A2 ‐70o A3‐ 90o significant
0 4 1 different between
three groups at
1 3 2
alpha 0.05?
3 6 2
1 3 0
4. Make decision
0 4 0
Source SS df MS F 15 16
Draw
Anova table ?
Sum of Sum of
Squares df Mean Square F Sig. Squares df Mean Square F Sig.
Between Groups Between Groups
30.00 2 15.00 11.25 .002
Within Groups Within Groups
16.00 12 1.33
Total Total
46.00 14
19
Analysis of Variance 20
23 24
Measuring Effect Size for an Measuring Effect Size for an
Analysis of Variance Analysis of Variance
• To supplement the hypothesis test, it is • For the t statistics, this percentage was identified
recommended that you calculate a measure of as r2, but in the context of ANOVA the percentage is
effect size. identified as η2 (the Greek letter eta, squared).
• For an analysis of variance the common technique • The formula for computing effect size is:
for measuring effect size is to compute the
percentage of variance that is accounted for by the SSbetween treatments
treatment effects. η2 = ───────────
SStotal
8438.641
η2 = ─────────── = 0.183 Treatment
46122.166 A B C
2 3 7
6 7 5
2 6 4
6 4 8
27 Source SS df MS F 28
Example 2
Between SSbetween dfbetween
Treatment Within
subject
SSwithin
subject
dfwithin
A B C treatment
Total
31 32
SPSS OUTPUT
ANOVA
ANSWER
Sum of Mean
Squares df Square F Sig.
Between
Groups 8 2 4 1 0.405
Within Groups
36 9 4
Total
44 11
ANOVA TABLE
ANOVA TABLE
ANSWER ANSWER
Sum of Mean Sum of Mean
Squares df Square F Squares df Square F
Between Between
Groups Groups 468.7 2 234.35 28.53
Within Groups Within Groups
345 42 8.21
Total Total
813.7 44
35 36
Latihan1Q2.sav
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
37 38
1
2
3 5
4
6
39 40
1 2
10
Ambil
Perhatian
4
5 6 7
3
Kesimpulan?
p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
p < α Ho ditolak, Terdapat Perbezaan yg signifikan
• p = 0.000 p = Sig.
• p < α
• @ p < 0.05 , maka Ho ditolak. Terdapat perbezaan
markah pensyarah yang signifikan antara
program. Perbezaan ini wujud antara program
‐ SPA dengan SPE
‐ SPA dengan SPS
‐ etc
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
2
The Logical Background for a
Repeated‐Measures ANOVA
Analysis of Variance
• The repeated‐measures design eliminates
individual differences from the
between‐treatments variability because the same
subjects are used in every treatment condition.
• The result is a test statistic similar to the
independent‐measures F‐ratio but with all
Repeated‐Measures individual differences removed.
Example 1
11 12
Steps
1. State the hypothesis
Ho: µ1= µ2= µ3
H1: At least one of the treatment means is different.
2. Find the critical value with α = 0.05
4. Make decision
Example 1
15 16
Source SS df MS F Source SS df MS F
Between SSbetween dfbetween Between
Total
Total SStotal dftotal
Source SS df MS F Measure:MEASURE_1
19 20
Example 2 Example 2
Depression Scores
Sample Before Therapy After Therapy 6 Month Person
Depression Scores Follow up Total
Sample Before Therapy After Therapy 6 Month A 71 53 55 179 n=5
Follow up
B 62 45 44 151 K=3
A 71 53 55
C 82 56 61 199 N=15
B 62 45 44
D 77 50 46 173 G=892
C 82 56 61
E 81 54 55 190
D 77 50 46
T=373 T=258 T=261
E 81 54 55
SS=273.2 SS=73.2 SS=198.8
Source SS df MS F Source SS df MS
Between SSbetween dfbetween Between
Total
Total SStotal dftotal
23 24
Measuring Effect Size for the
ANOVA table Repeated‐Measures Analysis of
Variance
Source SS df MS The formula for computing effect size for a repeated‐measures
Between 1718.5 2 859.25 69.6 ANOVA is:
Latihan3Q1.sav
27 28
Kesimpulan?
p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
p < α Ho ditolak, Terdapat Perbezaan yg signifikan
p = Sig.
• p = 0.000
• p < α
@ p > 0.05 , maka Ho ditolak. Terdapat
perbezaan skor pencapaian pelajar dalam
ujian‐ujian yang diduduki.
31
31
Lebih dari 2 1 data ordinal/
kumpulan / Dependent
independent
Latihan1Q6.sav
3 4
Kesimpulan?
p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
p < α Ho ditolak, Terdapat Perbezaan yg signifikan
p = Asymp Sig.
• p = 0.470
• p > α
• @ p > 0.05 , maka Ho GAGAL
ditolak. Terdapat perbezaan
pilihan yang signifikan antara
kumpulan
7
7
SPSS‐Ordinal data
3 4
Sampel
8
7 8
Kesimpulan?
p > α Gagal menolak Ho, Tidak terdapat perbezaan yang signifikan
p < α Ho ditolak, Terdapat Perbezaan yg signifikan
p = Asymp Sig.
• p = 0.000
• p < α
• @ p > 0.05 , maka Ho ditolak.
Perdapat perbezaan pilihan
yang dibuat antara penyataan
soalan 1‐ 5.
Compare three or more unmatched
One‐way ANOVA Kruskal‐Wallis test
groups
Compare three or more matched
Repeated‐measures ANOVA Friedman test
groups
Quantify association between two
Pearson correlation Spearman correlation
variables
3 4
Independent Dependent
variables variables
YES NO
GROUP A 16 34
GROUP B 7 43
12
11 12
Goodness‐of‐Fit Goodness‐of‐Fit
15 16
15 16
Goodness‐of‐Fit Goodness‐of‐Fit
Goodness‐of‐Fit
CHI SQUARE
Goodness‐of‐Fit
4. Make decision
Reject hypothesis Null
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
19 20
SPSS
Keptusan?
• p = 0.044
• p < α
• @ p < 0.05 , maka Ho
ditolak. Terdapat
perbezaan pilihan view
oleh pembeli rumah.
23 24
Data Structure On
Spss
27 28
SPSS
Output
Decision
Independence Independence
The second chi‐square test, the chi‐ A frequency distribution showing willingness to
square test for independence, can use mental health service according to gender
be used and interpreted in two for a sample of 150
different ways: Willingness to use Mental Health Service
a) Testing hypotheses about the Probably No Maybe Probably yes
relationship between two Male 17 32 11
variables in a population, or Female 13 43 34
Version 1:
Independence
A frequency distribution showing willingness to
Ho : In the general population, there is no use mental health service according to gender
relationship between gender and for a sample of 150
willingness to use mental health service.
Willingness to use Mental Health Service
Probably No Maybe Probably yes
Version 2
Male 17 32 11 60
Ho : In the general population, the Female 13 43 34 90
30 75 45 N=150
distribution of reported willingness to use
mental health service is the same for male and
female.
36
35 36
Willingness to use Mental Health Service
Probably No Maybe Probably yes
Willingness to use Mental Health Service
Male 17 (12) 32 (30) 11 (18) 60
Probably No Maybe Probably yes
Female 13 (18) 43 (45) 34 (27) 90
Male 17 (12) 32 (30) 11 (18) 60
30 75 45 N=150
Female 13 (18) 43 (45) 34 (27) 90
20% 50% 30%
30 75 45 N=150
20% 50% 30%
=30/150*100
=75/150*100
Version 1:
Independence R=Row = (2‐1)(3‐1)
C=Column
Ho : In the general population, there is no 2
1
relationship between gender and
willingness to use mental health service.
3 Compute
H1: In the general population, there is a consistent
predictable relationship between gender and
willingness to use mental model service.
39 40
Table
5.99 Calculation
8.23
SPSS
CHI SQUARE
Reject H0
4 Or
There is a significant relationship for Independence
between gender and willingness to use
mental model service.
1 1 17
1 2 32
1 3 11
2 1 13
…
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
43 44
47 48
Keputusan?
Correlation
Y
Y = mX + c ??
3 4
Correlations: Measuring and
Describing Relationships
• A correlation is a statistical
method used to measure and
describe the relationship between
two variables.
• A relationship exists when
changes in one variable tend to be
accompanied by consistent and
predictable changes in the
other variable.
• A correlation typically
evaluates three aspects
of the relationship:
the
degree
the
form
the
direction
7 8
Correlations: Measuring and Correlations: Measuring and
Describing Relationships Describing Relationships
relationships that are not linear
11 12
Correlations: Measuring and
Describing Relationships
a perfect negative ??
correla on, −1.00 no linear
trend, 0.00
a relatively weak
negative correlation,
approximately −0.40
a strong positive
relationship,
approximately
+0.90
www.utm.my innovative ● entrepreneurial ● global www.utm.my innovative ● entrepreneurial ● global
15 16
Examples of positive and
negative relationships.
19 20
Student X Y
A 0 4
B 2 1
C 8 10
D 6 9
E 4 6
23 24
Reporting SPSS
A 11 14
B 6 7
C 16 15
D 4 7
E 1 3
F 10 9
G 5 9
H 3 8
27 28
SPSS
Correlations
XX YY
XX Pearson Correlation
1 .904**
Sig. (2‐tailed)
.002
N
8 8
YY Pearson Correlation
.904** 1
Sig. (2‐tailed)
.002
N
8 8
**. Correlation is significant at the 0.01 level (2‐tailed).
31 32
The Point‐Biserial Correlation The Point‐Biserial Correlation
and the Phi Coefficient and the Phi Coefficient
• In situations where one variable
With either one or two dichotomous variables the is dichotomous (nominal or
point-biserial correlation
SPSS
CORRELATION
35 36
39
Linear Regression
• The Pearson correlation measures the
degree to which a set of data points form
a straight line relationship.
• Regression is a statistical procedure that
determines the equation for the straight
line that best fits a specific set of data.
Introduction to Regression
Introduction to 3
Introduction to 4
Introduction to 7 8
Linear Regression
Example
X Y
2 3
6 11
0 6
4 6
5 7
7 12
5 10
3 9
SPSS
SPSS
Regression
11 12
15 16
Example (pg#582)
Person Y(dependent) X1(independent) X2(independent)
A 11 4 10
B 5 5 6
C 7 3 7
D 3 2 4
E 4 1 3
F 12 7 5
G 10 8 8
H 4 2 4
J 8 6 10
J 6 2 3
19 20
Example
Person Y(dependent) X1(independent) X2(independent)
A 11 4 10
B 5 5 6
C 7 3 7 SPx1y=52
D 3 2 4
E 4 1 3 SPx2y=47
F 12 7 5
G 10 8 8 SPx1x2=35
H 4 2 4
J 8 6 10
J 6 2 3
My=7 Mx1=4 Mx2=6
SSy=90 SSx1=52 SSx2=64
SPSS
Multiple Regression
23 24
SPSS
example
X1 X2 Y
2 4 3
6 3 11
0 4 6
4 5 6
5 6 7
7 4 12
5 6 10
3 5 9
27 28
Coefficientsa
ANOVAb
Standardize
Unstandardized d
Coefficients Coefficients Sum of Mean
Model Squares df Square F Sig.
Model B Std. Error Beta t Sig. 1 Regression
1 (Constant)
37.147 2 18.573 3.458 .114a
5.726 4.123 1.389 .224 Residual
X1 26.853 5 5.371
1.011 .387 .758 2.612 .048
X2 Total
‐.382 .827 ‐.134 ‐.462 .663 64.000 7
a. Dependent Variable: Y
a. Predictors: (Constant), X2, X1
b. Dependent Variable: Y
Y=1.011X1 ‐ 0.382X2 + 5.726
31 32
35