Prof. Samart Powpaka (Associate Professor, Marketing Department; and Director of MSc Program in Marketing, CUHK Business School of The Chinese University of Hong Kong (CUHK)) was our guest speaker.
In the beginning, Prof. Powpaka introduced different types of research. He said the goal of academic research was to discovery theories; and a theory was an explanation of a phenomenon, as well as, it was facts.
Then he explained the research methodology that separated into “Concept” to “Construct” to “Valuable”. Where Concepts had constitutive definitions but was not able to be measured directly. Constructs were concepts that could be measured if operational definition added. The relationships among Concepts, Constructs and Variables were discussed. For instance, 1 concept → 1 construct → 1 scale → 1 or more observed variable. Different Scales were explained as follows:
i) Nominal scale – ID
ii) Ordinal scale – Ranking
iii) Interval scale - Degree of difference between items (without zero)
iv) Ratio scale – with zero value
After that Prof. Powpaka drew another diagram. The upper part (Perceived Quality (PQ) → Attitude (A)) was concept (in Empirical World) and the lower part (Xn → Yn) was Variables/Scales (in Abstract World). Since Concept was not be measured, we could only measure its variables and verified our concept relationship from Abstract World to Empirical World.
Three research design methods were introduced and they were Sample Surveys, Experiments and Field Studies. For A → B, Sample Surveys used when A and B both measurable; Experiments used when A was measurable and B was manipulable (with / with something); and Field Study when data collected by observation in the field. Each research designs had different strengths and weaknesses as the diagram showed.
- Sample Surveys are high in generalizability but low in precision/control and realism.
- Experiments are high in precision/control but low in generalizability and realism.
- Field Studies are high in realism but low in precision/control and generalizability.
There are two types of sampling that were probability and non-probability sampling. Prof. Powpaka drew a diagram to explain the following probability samplings:
- Simple random sampling (1)
- Stratified sampling (2)
- Cluster sampling (in red color)
- Systematic sampling (No particular pattern)
- Area sampling (e.g. US City → Block → Room → Person)
Prof. Powpaka plotted a table to explain different statistic techniques employed in specific conditions. Where DV – Dependence Variable, IV – Independence Variable, m – matrix, C – Character variable (e.g. Gender)
For Positioning Research, Prof. Powpaka showed the steps from “Need Recognition (Problem)” to “Search” to “Alternative Evaluation” and then obtained the Perception Map(s). Where A to F in diagram was Shopping Centre and 60 degree was ID line. It could estimate the ideal location.
After that Prof. Powpaka demonstrated the data analysis using SPSS using the model that X1-X11 were parameters and Y1 was Brand and Y2 Purchasing Intention. The first step used Factor Analysis to find how many factors (Data reduction) and then compared their Means (FS) and using factor scores for multiple regression with Y1. Finally, he used simple regression to find the relationship between Y1 and Y2.
At the end of the workshop, I asked Prof. Powpaka how to calculate Cronbach’s Alpha. He demonstrated and said it should be larger than 0.7. He also told me that Cronbach’s Alpha used as reliability test for sample survey under “Structural Equation Modeling”. (P.S. usually used to test model with large scales)
Reference:
HKSTP - http://www.hkstp.org
In the beginning, Prof. Powpaka introduced different types of research. He said the goal of academic research was to discovery theories; and a theory was an explanation of a phenomenon, as well as, it was facts.
Then he explained the research methodology that separated into “Concept” to “Construct” to “Valuable”. Where Concepts had constitutive definitions but was not able to be measured directly. Constructs were concepts that could be measured if operational definition added. The relationships among Concepts, Constructs and Variables were discussed. For instance, 1 concept → 1 construct → 1 scale → 1 or more observed variable. Different Scales were explained as follows:
i) Nominal scale – ID
ii) Ordinal scale – Ranking
iii) Interval scale - Degree of difference between items (without zero)
iv) Ratio scale – with zero value
After that Prof. Powpaka drew another diagram. The upper part (Perceived Quality (PQ) → Attitude (A)) was concept (in Empirical World) and the lower part (Xn → Yn) was Variables/Scales (in Abstract World). Since Concept was not be measured, we could only measure its variables and verified our concept relationship from Abstract World to Empirical World.
Three research design methods were introduced and they were Sample Surveys, Experiments and Field Studies. For A → B, Sample Surveys used when A and B both measurable; Experiments used when A was measurable and B was manipulable (with / with something); and Field Study when data collected by observation in the field. Each research designs had different strengths and weaknesses as the diagram showed.
- Sample Surveys are high in generalizability but low in precision/control and realism.
- Experiments are high in precision/control but low in generalizability and realism.
- Field Studies are high in realism but low in precision/control and generalizability.
There are two types of sampling that were probability and non-probability sampling. Prof. Powpaka drew a diagram to explain the following probability samplings:
- Simple random sampling (1)
- Stratified sampling (2)
- Cluster sampling (in red color)
- Systematic sampling (No particular pattern)
- Area sampling (e.g. US City → Block → Room → Person)
Prof. Powpaka plotted a table to explain different statistic techniques employed in specific conditions. Where DV – Dependence Variable, IV – Independence Variable, m – matrix, C – Character variable (e.g. Gender)
For Positioning Research, Prof. Powpaka showed the steps from “Need Recognition (Problem)” to “Search” to “Alternative Evaluation” and then obtained the Perception Map(s). Where A to F in diagram was Shopping Centre and 60 degree was ID line. It could estimate the ideal location.
After that Prof. Powpaka demonstrated the data analysis using SPSS using the model that X1-X11 were parameters and Y1 was Brand and Y2 Purchasing Intention. The first step used Factor Analysis to find how many factors (Data reduction) and then compared their Means (FS) and using factor scores for multiple regression with Y1. Finally, he used simple regression to find the relationship between Y1 and Y2.
At the end of the workshop, I asked Prof. Powpaka how to calculate Cronbach’s Alpha. He demonstrated and said it should be larger than 0.7. He also told me that Cronbach’s Alpha used as reliability test for sample survey under “Structural Equation Modeling”. (P.S. usually used to test model with large scales)
Reference:
HKSTP - http://www.hkstp.org
