Week 14: Exam Preparation: Slide 1
Week 14: Exam Preparation: Slide 1
Slide 1
Research Design Question Example
Slide 2
Exercise
Based on the scenarios given below please suggest how a researcher can
use a suitable research design (give reasons for your choice)
1. Type of research
2. Extent of researcher interference
3. Study Setting
4. Time horizon
5. Unit of analysis
Slide 3
Example
• Tan Mei Mei, a matron in a general hospital, found that the nurses in Pulau Angsa General Hospital often
complain about the extreme stress that they are facing at work due to the lack of emotional support. She
records all the complaints received according to the ward to date in the hospital’s book of complaints. In
order to test how far each complaint is true, Tan Mei Mei decides to conduct a research concerning the
relationship between the perceived emotional support and the stress faced by the nurses.
• The Pulau Angsa General Hospital is divided into 3 divisions, the accident ward, the children’s ward and
the maternity ward. According to the report, the nurses at the accident ward are the most stressed. Tan Mei
Mei then arranged a strategy whereby two counselors will be stationed at the accident ward, 1 counselor at
the maternity ward while no counselor will be placed at the children’s ward.
• She will begin the strategy today. After a month she will check the hospital’s book of complaints to
ascertain whether the complaints have increased or decreased.
Slide 4
Research Model Question Example
Slide 5
Answer these questions
Slide 6
Exercise
• A marketing manager would like to predict the factors that influence purchase
decision. He conducts a literature search and also interview with several clients.
After going through this process, he forwards this suggestion.
• When the product is of high quality, the confidence towards the product will
increase and this will lead to higher purchase intention. He also found that when
the price is low, the intention to purchase will also increase. If the brand of the
product is more known, then the purchase intention will also increase. However,
he also found that if the client has high brand loyalty for a different brand than
the one marketed by the company, then the brand knowledge does not play an
important role in the intention to purchase.
Slide 7
Research Question
a. What are the factors influencing purchasing decision?
Slide 8
Research Model
Purchase Decision
Slide 9
Research Model
Quality
Brand Knowledge
Slide 10
Research Model
Quality
Confidence
Brand Knowledge
Slide 11
Research Model
Quality
Confidence
Brand Knowledge
Brand Loyalty
Hypotheses
H1 Quality has a positive effect on Confidence
H2 Confidence has a positive effect on Purchase Decision.
H3 Price has a negative effect on Purchase Decision.
H4 Brand Knowledge has a positive effect on Purchase Decision.
H5 The relationship between Quality and Purchase Decision is mediated by Brand
Confidence.
H6 The positive relationship between Brand Knowledge and Purchase Decision will be
stronger when Brand Loyalty is high.
Slide 13
Multiple Regression Question Example
Slide 14
Exercise 1
The MBA was started in USM since 1995 and the dean would like to
analyze the factors that influence the performance of the candidates
(MBAGPA). One hundred students who have graduated the last three years
were selected and their GPA during their first degree (UNDERGPA), GMAT
score (GMAT) and their number of years working experience (WORK)
before they joined were recorded. The data was analyzed using the SPSS
software and the output is presented below:
Slide 15
Data Set
Slide 16
Output
b
Model Summary
ANOVA b
Sum of
Model Squares df Mean Square F Sig.
1 Regression 52.063 3 17.354 30.516 .000 a
Residual 54.594 96 .569
Total 106.656 99
a. Predictors: (Constant), WORK, GMAT, UNDERGPA
Coefficientsa
Standardi
zed
Unstandardized Coefficien
Coefficients ts Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
1 (Constant) .530 1.323 .401 .689
Slide 18
Collinearity Diagnostics
Collinearity Diagnostics a
Variance Proportions
Condition
Model Dimension Eigenvalue Index (Constant) UNDERGPA GMAT WORK
1 1 3.738 1.000 .00 .00 .00 .02
2 .252 3.854 .00 .00 .00 .98
3 8.437E-03 21.048 .02 .20 .75 .00
4 2.063E-03 42.569 .98 .79 .25 .00
a. Dependent Variable: MBAGPA
Slide 19
Histogram
Histogram
12
10
4
Frequency
0 N = 100.00
Slide 20
Normality of the error term
Normal P-P Plot of Regression Standardized Residual
Dependent Variable: MBAGPA
1.00
.75
.50
Expected Cum Prob
.25
0.00
0.00 .25 .50 .75 1.00
Slide 22
Based on the output answer the questions below:
a) Develop a suitable regression model.
Coefficientsa
Standardi
zed
Unstandardized Coefficien
Coefficients ts Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
1 (Constant) .530 1.323 .401 .689
UNDERGPA 8.236E-02 .105 .057 .782 .436 .988 1.013
Slide 23
Based on the output answer the questions below:
b) What can say about the model developed in part (a).
The R2 is 0.488 indicating that 48.8% of the variance in MBAGPA can be explained by the 3
independent variables and there are another 51.2% of the variance that cannot be explained.
b
Model Summary
Slide 24
Based on the output answer the questions below:
c) Can we conclude at the 1% level that the model is a valid model to predict MBAGPA?
H0: All he 3 independent variables cannot explain the variance in MBAGPA
H1: All he 3 independent variables can explain the variance in MBAGPA
The F-value is 30.516, p< 0.01 thus we can reject H0 and conclude that the model can be used to
predict MBAGPA.
ANOVA b
Sum of
Model Squares df Mean Square F Sig.
1 Regression 52.063 3 17.354 30.516 .000 a
Residual 54.594 96 .569
Total 106.656 99
a.
Predictors: (Constant), WORK, GMAT, UNDERGPA
b. Dependent Variable: MBAGPA
Slide 25
Based on the output answer the questions below:
d) Predict MBAGPA if given the below:
UNDERGPA = 8
GMAT = 630
WORK = 5
MBAGPA = 8.52
Slide 26
Based on the output answer the questions below:
e) Test if each variable has a significant impact on MBAGPA at the 5% level.
H0 b1 = 0 b2 = 0 b3 = 0
H1 b1 ≠ 0 b2 ≠ 0 b3 ≠ 0
t = 0.782 t = 8.505 t = 4.225
p > 0.05 p < 0.05 p < 0.05
Coefficientsa
Standardi
zed
Unstandardized Coefficien
Coefficients ts Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
1 (Constant) .530 1.323 .401 .689
Slide 28
Assumption
Histogram
Dependent Variable: MBAGPA
14
12
10
Frequency
4
Std. Dev = .98
2 Mean = 0.00
0 N = 100.00
Collinearity Diagnostics a
Variance Proportions
Condition
Model Dimension Eigenvalue Index (Constant) UNDERGPA GMAT WORK
1 1 3.738 1.000 .00 .00 .00 .02
2 .252 3.854 .00 .00 .00 .98
3 8.437E-03 21.048 .02 .20 .75 .00
4 2.063E-03 42.569 .98 .79 .25 .00
a. Dependent Variable: MBAGPA
Based on the output answer the questions below:
g) What can you suggest from the results of this analysis?
The variable GMAT ( = 0.622) and WORK ( = 0.310) were the 2 most important predictors of
MBAGPA. Thus, in future recruitments of potential MBA students the Dean should focus on
recruiting students with high/good GMAT scores and with more years of work experience as work
experience also plays a role in the achievement of the students.
Slide 30
Exercise 2
Ahmad a manager at a university feels that lecturers ought to have high intentions to
share information among them so that the competitiveness of the university can be
enhanced. Lately there have been many complaints about the poor quality of
students from the public institutions of higher learning. Ahmad has collected data
from 96 lecturers and the variables that has been emphasized are Gender (gender of
the respondent 1 = Male, 0 = Female), Reciprocal (Reciprocal relationship between
lecturers), Self (self efficacy), Climate (Climate of the organization) and Intention
(Intention to share information). The data was analysed using SPSS and the output is
given below:
Slide 31
Data Set
Sum of
Model Squares df Mean Square F Sig.
1 Regression 24.131 4 6.033 37.006 .000a
Residual 14.835 91 .163
Total 38.966 95
a. Predictors: (Constant), climate, self, gender, reciprocal
b. Dependent Variable: intention
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
1 (Constant) .379 .339 1.117 .267
gender .243 .096 .165 2.533 .013 .986 1.014
reciprocal .690 .092 .607 7.528 .000 .644 1.553
self .231 .079 .229 2.927 .004 .685 1.460
climate .022 .060 .025 .370 .712 .882 1.134
a. Dependent Variable: intention
Slide 34
Collinearity
Collinearity Diagnosticsa
Slide 35
Based on the output answer the questions below:
Histogram
40
30
Frequency
20
10
Mean = 1.17E-15
0 Std. Dev. = 0.979
-3 -2 -1 0 1 2 3 N = 96
Scatterplot
1
Residual
-1
-2
-3
-4
-3 -2 -1 0 1 2 3
Regression Standardized Predicted Value
Slide 38
Partial Regression Plot
1
intention
-1
-2
Slide 39
Partial Regression Plot
1.0
0.5
intention
0.0
-0.5
-1.0
-1.5
Slide 40
Partial Regression Plot
1.0
0.5
intention
0.0
-0.5
-1.0
-2 -1 0 1 2
climate
Slide 41
Residuals
Residuals Statisticsa
Std.
Minimum Maximum Mean Deviation N
Predicted Value 2.5117 5.0165 3.8188 .50400 96
Residual -1.15633 1.14693 .00000 .39517 96
Std. Predicted Value -2.593 2.377 .000 1.000 96
Std. Residual -2.864 2.841 .000 .979 96
a. Dependent Variable: intention
Slide 42
Based on the output answer the questions below:
a. Can we develop a regression equation?
H0: All he 4 independent variables cannot explain the variance in Intention
H1: All he 4 independent variables can explain the variance in Intention
The F-value is 37.006, p< 0.01 thus we can reject H0 and conclude that we can develop a
regression equation.
ANOVA b
Sum of
Model Squares df Mean Square F Sig.
1 Regression 24.131 4 6.033 37.006 .000a
Residual 14.835 91 .163
Total 38.966 95
a. Predictors: (Constant), climate, self, gender, reciprocal
b. Dependent Variable: intention
Slide 43
Based on the output answer the questions below:
Slide 45
Based on the output answer the questions below:
Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
1 (Constant) .379 .339 1.117 .267
gender .243 .096 .165 2.533 .013 .986 1.014
reciprocal .690 .092 .607 7.528 .000 .644 1.553
self .231 .079 .229 2.927 .004 .685 1.460
climate .022 .060 .025 .370 .712 .882 1.134
a. Dependent Variable: intention
Collinearity Diag nosticsa
Slide 46
Based on the output answer the questions below:
e. Have all the assumptions of the regression fulfilled? If not, explain how they should be tested?
Assumptions:
1. Normality Histogram is normal
2. Normality of the error term P-P plot is not given so cannot test
3. Constant Variance Plot shows no pattern
4. Linearity Linear plots all shows linear relationship
5. Multicollinearity No issue as VIF ≤ 5, Conditional Index ≤ 30
6. Autocorrelations D-W = 1.633 is acceptable as it is within 1.5 – 2.5
7. Outliers There are no outliers as the std. residuals are ± 3
Slide 47
Click to edit Master title style
Questions Please
Slide 48