Professional Documents
Culture Documents
CI Revision
CI Revision
Involves discussions with the decision makers, interviews with industry experts,
analysis of secondary data and some qualitative research (e.g. focused groups).
-> The research can be designed and conducted properly.
=> It should be noted that these steps are interdependent and iterative (phụ thuộc
lẫn nhau và lặp lại). Thus, at each step, the researcher should not only look back
at the previous steps but also look ahead to the following steps.
Secondary Data and Primary Data:
❖ Always start with secondary data
Secondary Research Primary Research
Benefits Save time, money and effort -> While it can take time and
Data from reputable sources, resources to collect good, rigorous
further adds to the researchers’ primary data. The data belongs to
credibility in identifying a the researcher or his/her
trustworthy source of organisation.
information.
Research Design:
1. exploratory -> descriptive -> causal
2. descriptive -> exploratory -> skip causal
3. causal -> descriptive to refine the findings
Exploratory Descriptive Research Causal Research
Research (Experimental Research)
● an experiment is
conducted in a
natural setting - a
field experiment.
● researchers’ office or
university classroom
- lab experiment.
Sampling Method:
● Defining the population and sampling units
● Identifying the sampling frame - a list of elements from which a sample may
be drawn; a list of units (persons, households, businesses, etc.) in the survey
population
● Estimating the sample frame
● Determining the sampling methodology
● Choosing a sampling technique
Probability:
simple random sampling, systematic sampling - has rules to pick; stratified -
homogeneous, meaningful
- Simple Random: every element has an equal chance of being selected in
the sample.
- Systematic: random starting point, then every n-th element -> order of
elements becomes important.
- Stratified:
- Elements are divided into meaningful groups (e.g. year of study) -
should be homogeneous.
- Random sampling within strata, all included.
- Over-represent some elements -> disproportional sampling
- Cluster:
- Elements are divided into heterogeneous groups (e.g. suburbs,
states, tute classes)
- are selected at random to be included, not all clusters are included.
Non-probability:
convenience, snowball (initial group of respondents is selected, who then recruit
more respondents) - introduce bias (all similar people)
- Convenience Sampling: choosing sample elements that are available/easily
accessed at the time (e.g. intercepts, students in my class)
- Judgemental Sampling: population elements are selected based on the
judgment of the researcher
- Quota Sampling: set limits (quotas) for each subgroup of people
- Snowball Sampling: an initial group of respondents is selected, who then
recruit more respondents -> introduces bias
EXAMPLE:
RS: association between smoking and gender
H0: no association
H1: Association
p > 0.05
-> reject H1, accept H0
-> There is no association between smoking and gender.
p < 0.05
-> reject H0, accept H0
-> There is an association between smoking and gender.
Sample Size:
● Larger sample = more accuracy
-> closer to census
● Factors to consider:
○ Level of heterogeneity (differences) in the population
○ Acceptable range of error
○ Confidence level
● Samples of 150 or more are generally sufficient
Sampling Error:
● Random sampling error: difference between sample and population due to
chance
● Systematic sampling errors: due to the researcher/researcher design
○ Sampling frame error
○ Sample/population definition
○ Self selection
Measures and Designing The Questionnaire:
Scaling:
Nominal Scale:
Numbers/alphabets/symbols serve only as labels for identifying and
classifying the objects: -> phân loại đối tượng
● No quantitative meaning
● Used for categorical data
● Characteristic: Description
-> What is your gender? (Male/Female)
-> What is your hair colour? (1-Brown; 2-Black; 3-Blonde; 4-Gray; 5-Other)
Ordinal Scale:
Numbers indicate the extent to which an object has more or fewer attributes
than others: -> định danh các giá trị được sắp xếp theo thứ bậc
● The order matters
● The magnitude differences do not matter
● Characteristic: Description & Order
-> ex: indicate the customers’ first, second, third and fourth choices of the hotels
-> Academic performance:
1. Weak, poor
2. Average
3. Fair
4. Excellent
Interval Scale:
Numbers indicate the extent to which an object has more or fewer attributes &
the distance between each number: có hai cực ở hai đầu dãy số thể hiện hai trạng
thái đối nghịch nhau -> cho biết được khoảng cách giữa các thứ bậc
● Order is known
● The differences can be compared
● “0” does not signify the absence of the attribute
● Characteristic: Description, Order & Distance
-> Measure attitudes, opinions, evaluations, beliefs, impressions, perceptions,
feelings, and intentions on a continuum.
-> takes the form of a series of continuous and regular digits 1-5, 1-7 or 1-10.
-> 1: completely disagree; 2: disagree; 3: normal; 4: agree; 5: Totally agree
-> 0 degrees Celsius does not mean the absence of heat
Ratio Scale:
Numbers indicate the extent to which an object has more or fewer attributes &
the distance between each number & absolute zero point: dùng cho đặc tính số
lượng - có đầy đủ đặc tính của thang đo khoảng cách -> cho phép lấy tỉ lệ so sánh
giữa hai giá trị của biến số.
● Order and difference are known
● “0” does signify the absence of the attribute
● Characteristic: Descriptive, Distance, Order & Origin
-> Number of children:_____
-> My monthly salary is:_____
EXAMPLE CASE:
1. Unilever is interested in utilizing secondary data before it starts
collecting primary data for a research study on Sydney-based
consumers. What possible sources of reliable information could the
company look at?
Internal sources:
● reports, previous research findings or old documents which may still be used
to understand a particular phenomenon.
External sources:
● public libraries, government departments, council offices, and various
organizations, as well as in newspaper or journal articles.
This respondent profile includes people of all ages and it is not specifically targeted
at teenagers. By including all ages you include family members who do the grocery
shopping 99% of the time and therefore you have more accurate responses and the
answers are raw.