Quantitative Approaches

Model for Conceptualizing

Quantitative Research
Overall purpose or
objective
Research literature
Research questions
and hypotheses
Selecting appropriate
methods
Validity and reliability
of the data

2
 Levels of Measurement
• Nominal
• Ordinal
• Interval
• Ratio
 Nominal-Scale Measurement
• Lowest of the four levels of measurement
• Categories that are not more or less, but are
different from one another in some way
• Mutually exclusive and exhaustive categories
• Named categories
 Examples of Nominal Data
• Gender
1 = Male
2 = Female

• Satisfied by the treatment

 Yes
 No
 Ordinal-Scale Measurement
• Order/ranking imposed on categories
• School grades
 Very good
 Good
 Satisfying
 Sufficient
 Failed
 Interval-Scale Measurement
• Numerical distances between intervals
• Absence of a zero point
• Likert scale scores
1 = Strongly disagree
2 = Disagree
3 = Neutral
4 = Agree
5 = Strongly agree
 Ratio-Scale Measurement
• Highest for measurement
• Continuum of values
• Absolute zero point
 Example of Ratio Data
• Test scores
– 1 = Lowest third percentile
– 2 = Middle third percentile
– 3 = Top third percentile
 The Four Major Levels of Measurement
(Davis and Cosenza, 1988)
Basic Typical Typical Statistics
Level Description Empirical Usage
Operation Descriptive Inferential

Nominal Uses numerals to Determination Classification Percentages/ Nonparametric

identify objects, of equality/ mode ( X2 )
individuals, events, or inequality
groups.
Ordinal In addition to Determination Rankings/ Median Nonparametric
identification, the of greater or Ratings (mean and (parametric)
numerals provide less variance)
information about the
relative amount of some ( Rank-order,
characteristics Correlation )
possessed by an event,
object, etc.
Interval Possesses all the Determination Preferred Mean/ Parametric
properties of nominal of equality of measure of Variance/
and ordinal scales plus intervals complex Standard
the intervals between concepts/ ( t, F )
deviation/Co
consecutive points are constructs efficient of
equal. variation
Ratio Incorporates all the Determination When Mean/ Parametric
properties of nominal, of equality of precision Variance/
ordinal, and interval ratios instruments Standard
scales plus it includes are available ( t, F )
deviation/Co
 Population & Sample

Population Random Sample

$
$
$
$ $
$
$
Some Considerations in Determining Sample
Size According to Roscoe (1975)

 The sample size is greater than 30 and less than

500 is enough for most research.

 If the sample will be split into sub-samples, the

minimum number of samples in each category is 30.

 In the multivariate study, the sample size should be

several times (preferably 10X or more) than the
number of variables in the study.
 Sample Size According Krejcie and Morgan
(1970) for Multivariables
------------------------------------------------------------------------------------------
N: 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
------------------------------------------------------------------------------------------
S: 10 14 19 24 28 32 36 40 44 48 52 56 59 63 66
------------------------------------------------------------------------------------------
N : population size ; S: sample size
------------------------------------------------------------------------------------------
N: 85 90 95 100 110 120 130 140 150 160 170 180 190
------------------------------------------------------------------------------------------
S: 70 73 76 80 86 92 97 103 108 113 118 123 127
------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------
N: 200 210 220 230 240 250 260 270 280 290 300 320 340
------------------------------------------------------------------------------------------
S: 132 136 140 144 148 152 155 159 162 165 169 175 181
------------------------------------------------------------------------------------------
N : population size ; S: sample size
------------------------------------------------------------------------------------------
N: 360 380 400 420 440 460 480 500 550 600 650 700 750
------------------------------------------------------------------------------------------
S: 186 191 196 201 205 210 214 217 226 234 242 248 254
------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------
N: 800 850 900 950 1000 1100 1200 1300 1400 1500 1600
------------------------------------------------------------------------------------------
S: 260 265 269 274 278 285 291 297 302 306 310
------------------------------------------------------------------------------------------
N : population size ; S: sample size
------------------------------------------------------------------------------------------
N: 1700 1800 1900 2000 2200 2400 2600 2800 3000 3500
------------------------------------------------------------------------------------------
S: 313 317 320 322 327 331 335 338 341 346
------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------
N: 4000 4500 5000 6000 7000 8000 9000 10000 15000
------------------------------------------------------------------------------------------
S: 351 354 357 361 364 367 368 370 375
------------------------------------------------------------------------------------------
N : population size ; S: sample size
------------------------------------------------------------------------------------------
N: 20000 30000 40000 50000 75000 1000000
------------------------------------------------------------------------------------------
S: 377 379 380 381 382 384
------------------------------------------------------------------------------------------
Source : Sekaran, Uma (2000). Research Methods for Business: A Skill Building
Approach. Third Edition. John Wiley & Sons, Inc.
 Sample Size- Finite Population (where the
population is less than 50,000)
Slovin Formula (1960)

N
n  2
1  Ne

N = Population
n = sample size
e = sampling error (2%, 3% , 5%....)
Example :

N = 2000
e = 5% or 0.05

N
n
1  Ne 2

2000
n 2
 385
1  2000(0.05)
 Sample Size- Infinite Population (where the
population is greater than 50,000)

2
z p (1  p )
n 2
e
Z = confident level ( CL = 90% z=1.65; CL=95% z=1.96; CL=99% z=2.58)
e = sampling error ( 2%; 3% etc)
P = population variance or proportion estimation for one group/attribute in
population
Example :

• Sampling error 2%, confident level = 95% or z

= 1.96
• Population variance 50% p=50%

1.96 2 0.50(1  0.50)

n 2
 2401
0.02
 Sample Size Population is known
2
z [ p(1  p)]N
n 2
z [ p(1  p)]  ( N  1)e 2
Z = confident level ( CL = 90% z=1.65; CL=95% z=1.96; CL=99% z=2.58)
E = sampling error ( 2%; 3% etc)
P = population variance
N= population
• Sampling error 3%, confident level = 95% or z
= 1.96
• Population variance 50% p=50%
• N = 50.000

2
1.96 [0.50(1  0.50)]50000
n 2 2
 1045
1.96 [0.50(1  0.50)]  (50000  1)0.03
 Questionnaires
• Administration
– In person/on phone
– Self-administered
– Mail
 Validity and Reliability Test

All questions in questioner must be tested the value of Validity and Reliability
a.Validity Test
Measurement of construct validity using correlation coefficients between each of
the statements in the questionnaire with a total score.
For statements that measure positive values​​, any such answer given value (score)
as follows:
Statement Score
Strongly Agree 5
Agree 4
Neutral 3
Disagree 2
Strongly Disagree 1
 An Analogy to Validity and Reliability
(Babbie, 1992)

Reliable but not valid Valid but not reliable Valid and reliable
 Example:
No of Statement
Respondent
1 2 3 4 5 6 7 8 9 10 Total

A 5 4 5 4 5 4 4 5 4 5 45
B 4 5 4 5 4 5 5 4 5 4 45
C 3 4 4 3 4 2 4 5 4 4 39
D 5 5 4 3 4 5 4 3 4 1 38

E 3 3 4 4 3 4 4 4 4 1 34
F 2 1 2 1 2 2 2 2 1 5 21
G 2 3 2 3 2 2 3 3 1 5 26
H 1 1 1 2 1 2 2 1 2 5 16
I 4 4 4 5 5 5 4 4 4 1 40
J 3 2 2 2 2 3 2 2 2 5 24
Respondent X Y X2 Y2 XY
A 5 45 25 2025 235
B 4 45 16 2025 180
C 3 39 9 1521 117
D 5 38 25 1444 190
E 3 34 9 1156 102
F 2 21 4 441 42
G 2 26 4 676 52
H 1 16 1 256 16
I 4 40 16 1600 160
J 3 24 9 576 72
Total (N =10) 32 328 118 11720 1166
• Where: X : score of statement No. 1
Y : Total score

N ( X iYi )  ( X i . Y i)
rXY 
{N  X  ( X i ) }{N  Yi  ( Yi ) }
i
2 2 2 2

(10)(1166)  (32)(328)
rXY   0,884
{(10)(118)  (32)(32)}{(10)(11720)  (328)(328)}
• The results of the calculations for the items (statements
which are as follows:
Statement No. Correlation Coefficient

1 0,884
2 0,893
3 0,931
4 0,811
5 0,920
6 0,705
7 0,827
8 0,893
9 0,867
10 0,564
• Statistically, the number of correlations obtained should be compared
with the correlation table of critical points on the line N-2 with a
significance level of 5% or 1% (Ancok, 1989)

Degree of freedom (df ) 5% 1%

1 .997 1.000
2 .950 .990
3 .878 .959
4 .811 .917
5 .754 .874
6 .707 .834
7 .666 .798
8 .632 .765
9 .602 .735
10 .576 .708
Degree of freedom(df) 5% 1%
11 .553 .684
12 .532 .661
13 .497 .623
14 .497 .623
15 .482 .606
16 .468 .590
17 .456 .575
18 .444 .561
19 .433 .549
20 .423 .537
21 .413 .526
22 .404 .515
23 .396 .505
24 .338 .495
25 .381 .485
Degree of freedom (df) 5% 1%
26 .374 .478
27 .367 .463
28 .361 .463
29 .355 .456
30 .349 .449
35 .325 .418
40 .304 .393
45 .288 .372
50 .273 .354
60 .250 .325
70 .232 .302
80 .217 .283
90 .205 .267
b. Reliability Test
Reliability is an index that indicates the extent to which a
measure can be reliable or unreliable (Ancok, 1989).

1. Coefficient Alpha (Cronbach)

ANALYZING QUANTITATIVE DATA
 Correlation Models

• Answers ‘How strong is the linear

relationship between two variables?’
• Coefficient of correlation
– Sample correlation coefficient denoted r
– Values range from –1 to +1
– Measures degree of association
– Does not indicate cause–effect relationship
 Coefficient of Correlation
SS xy
r
SS xx SS yy

 x
2

SS xx   x 2


where n
 y
2

SS yy   y 2


n

SS xy   xy 
  x   y 
n
 Table of Correlational Statistics
Bivariate Statistics Purpose
- Product-moment correlation
- Rank-difference correlation Use to describe the strength of
- Kendalls tau relationship between two
- Biserial correlation variables
- Point – biserial correlation
- Phil coefficient
- Contigency coefficient
- Correlation ratio

Multivariate Statistics Purpose

-Multiple linear regression -Used to describe the strength of
discriminant function relationship between several
independent variables and on
dependent variable
 Coefficient of Correlation
Values
Perfect Perfect
Negative No Linear Positive
Correlation Correlation Correlation

–1.0 –.5 0 +.5 +1.0

Increasing degree of Increasing degree of

negative correlation positive correlation
 Linear Regression Model

Relationship between variables is a linear

function
Population Population Random
y-intercept Slope Error

y  0  1x  
Dependent Independent
(Response) (Explanatory)
Variable Variable
• Satisfaction = 38.480 -1.187 Motivation

Pre-University Bridging Program 8-26

August 2011
 First–Order Multiple
Regression Model
Relationship between 1 dependent and 2 or
more independent variables is a linear
function
Population Population Random
Y-intercept slopes error
y   0   1 x1   2 x 2     k x k  
Dependent Independent
(response) (explanatory)
variable variables
 Step in Regression Analysis
1. Analysis of Data (Demographic Analysis)
2. Descriptive Analysis
3. Model Evaluation (Classical Assumption Test)
a. Normality Test (Histogram/normal distribution/diagonal line)
b. Multicollinearity Test (Variance Inflation Factor/VIF = 0.1< VIF< 10)
c. Heterocedasticity Test (Scatter plot...randomly)
d. Autocorrelation (Durbin Watson / DW = -2 < DW < 2)
4. Interpretation Results
a. Goodness of Fit Test ( F-Test , t-Test)
b. Coefficient of correlation (r)
c. Coefficient of determination (r square)
• Normality Test : determine whether normal data distribution
or not, see the probability plot around the diagonal line
• Multicollinearity Test : test whether a regression model found
a correlation between the independent variables
• Heterocedasticity Test : whether there is inequality of
variance of the residual from one observation to other
observations.

Note : homocedasticity is good

Figure : Normality

Closest to a
straight line…
 Figure : Heterocedasticity

Copyright © 2006 Pearson

Addison-Wesley. All rights 13-45
reserved.
• F – Test : is used to know all independent variable (X) toward
dependent variable (Y)
• t – Test : is used to know each independent variable (X)
toward dependent variable (Y)
• Coefficient of determination : value that indicates how much
the contribution variable (X) to change the variable (Y)
 Examples :

1. Analysis of the Effectiveness of Advertising Marketing

Strategies One of Printing and Publishing PT. Rambang Using
EPIC Model Method (Yudi Farola Bram)

2. Analysis of Factors Affecting Location Selection Of Success

Service Business (Case Study on the Micro-Small Business
Services Around Campus UNDIP Pleburan (Azizah Pratiwi)

3. Analysis of Capital Structure on Plastics and Packaging Sector

Company in the Indonesia Stock Exchange (Indra Widjaja)
• Y = Bo + B1X1 + B2X2 + Ꜫ
• Performance = 0.065 + 0.648 Stress + 0.249
Satisfaction

Pre-University Bridging Program 8-26

August 2011
 Group Assignment (3 – 4 Persons)
• Topic
• Previous Research
• Framework
• Operational Definition
• Data & Analysis (Validity, Reliability, Respondent Profiles, Descriptive Statistics, Regression)
- 3 – 5 independents, 1 dependent variable
- Respondent = 20 – 30 (validity & reliability)
- Data: 50 – 100 respondent
- Interpretation
- Conclusions & Recommendation
- Total Pages = 15 - 30 pages (exclude references), space 1,5, times new romans
- Deadline to Submit: Week 15th (Before Final Test), max plagiarism 20%
- References
- Appendix (Plagiarism Result, Draft of Questioner, Raw Data)
-

Pre-University Bridging Program 8-26

August 2011
Performance = 0.065 + 0.648 Stress + 0.249
Satisfaction
Performance = 25.234- 0.662 Stress – 1.070
Satisfaction + 0.069 Moderating

Pre-University Bridging Program 8-26

August 2011
• Customer satisfaction = 2.264 + 0.193Tangible +
0.245Emphaty + 0.085Reliable +
0.211Responsiveness + 0.361Assurance
(Model 1)
• Customer satisfaction = 2.264 + 0.193Tangible +
0.245Emphaty + 0.211Responsiveness +
0.361Assurance
(Model 2)

Pre-University Bridging Program 8-26

August 2011
Moderating

Stress (X1) Performance (Y)

Satisfaction
(X2)/Moderating
Y = a + b 1 X 1 + b 2X 2 + b 3X 1X 2
Y = dependent variable (performance)
X1 = independent variable (stress)
X2 = moderating variable (satisfaction)
X1.X2 = interaction between independent and
moderating variable.
Intervening

0.723
Stress Performance

-0.536
0.555

Satisfaction

-0.536 x 0.555 = - 0.297