Beruflich Dokumente
Kultur Dokumente
Chapter 1
Business Research: a systematic + objective process of finding
(gathering) and analyzing data in order to help people or managers
make better decision.
-
(Review Chapter 1)
Example of Basic vs Applied Research
A. What are some determining factors that lead to diversified
market strategy?
B. What are customer perceptions and attitudes on Toyota Altis
compared with its competitors?
C. Could leadership style influence job commitment of employees?
D. How much loyalty of customers who shop at Central Department
Store?
Ways to find Information
1. Authority : from those who have power or authority
2. Experience : from those who have past experiences
3. Intuition : from our own sense, instinct
4. Research : from the past research
Skip Chapter 2
Chapter 3
Six stages / steps in Research Process (* exhibit 3.6 p.60)
1. Problem Discovery + Problem Definition (Chapter 5)
- Problem discovery: through Exploratory Research Situational
Analysis
1.1 secondary data analysis or literature review past studies
1.2 pilot study small sampling survey with customers
1.3 experience survey asking experts or authority
1.4 case study study or observe any similar case to our problem
- Problem definition (or Research Problem): a formal statement of
research problem + objective it should be in sentence, not a word.
There are 3 formats of writing the Research Problem:1.1 research question : who, what, when, where, how, why questions
1.2 hypothesis : statement of belief and relationship between factors
1.3 research objective : To study..., To measure.. , To identify etc.
2. Planning Research Design (Chapter 7-11)
2.1 survey (interview, questionnaire) Chapter 8-9 - all
2.2 experiment (field or laboratory experiment) Chapter 11 - some*
2.3 secondary data analysis Chapter 7 some*
2.4 observation Chapter 10 some*
** Design the Questionnaire ** Chapter 13-15
3. Sampling (Chapter 16-17)
3.1 target population - who
3.2 sample size how many of them (small size vs large size)
3.3 sampling method Probability sample vs Non-probability sample
4. Data Collection/Gathering (Field work) (Chapter 18 skip)
4.1 pre-testing trial of data collection (e.g. about 30-40 sets)
4.2 main study actual data collection (e.g. equal to sample size)
(Review Chapter 3)
Example of Forward Linkage or Backward Linkage
- Since we plan to use in-depth interview for our research design,
we then need data interpretation(qualitative analysis) rather than
statistics in data analysis.
(Research design Data analysis)
- If we plan to choose kindergarten kids as our sample, wed better
use observation technique rather than questionnaire survey.
(Sampling plan Research Design)
Research Questions
What are factors influencing service quality of MK restaurant?
What are perceptions of teenagers toward MK restaurant?
Which are the top 5 menus that teenagers like to order the most?
Research Objectives
To study factors that influence service quality of MK restaurant
To measure perceptions of teenagers toward MK restaurant
To identify the top 5 menus of MK restaurant from teenagers
Research Hypotheses
- Speed of service is the important factor that influences service
quality of MK restaurant.
- Level of satisfaction of MK restaurant is high.
- There is a positive relationship between the modern atmosphere of
MK restaurant and the satisfaction level of teenagers.
- There is a difference in satisfaction of MK restaurant between
teenagers and working people.
- There is a difference in satisfaction of MK restaurant between
customers in different income groups.
6
Skip Chapter 4
Chapter 5
Theory Building (theory explanation of facts or phenomena)
1. Observe reality
= objects, people, events
(or Do literature review)
2. Create concepts
= given names of objects, events
3. Create propositions = relationship bet. Concepts
4. Create a theory
= relationship bet. Propositions / combination
of many propositions (or networks of
propositions)
(from min. abstract to max. abstract)
<tangible> <intangible>
Abstract level: just concept, idea
Ladder of Abstraction
Proposition
(relationship bet. Concepts)
Concepts
(abstract words, things)
Empirical Level
Research
(testing or verifying theory
by doing a research to test
the conceptual model)
Hypothesis
(relationship bet. Variables)
(deriving tested hypothesis
from the conceptual model)
Variables
(designing the
measurement of concepts
in the model)
8
Scientific method
: technique used to analyze empirical reality or evidence
: to confirm prior theories or disprove them and get the new theories
(knowledge)
literature review look for concepts propositions
develop hypotheses research design (method) data collection
data analysis conclusion (a new knowledge or theory)
Concepts = words
: abstract names of things (at abstract level) e.g. learning aids,
furniture, happiness etc.
Variables = words
: = Concepts/things that can be observed / experimented and have
different numerical values.
: = Concepts at empirical level e.g. power point, chair, smiling face
etc.
- Types of variable (based on function) dependent variable,
independent variable, intervening variable, moderating variable
- Types of variable (based on measurement) categorical variable,
continuous variable
Attributes = characteristics of a concept
e.g. furniture: - type of furniture = indoor / outdoor
- raw material it is made of = wooden / metal
- size of furniture = big / small
concept - use of furniture = living room / kitchen
Proposition = statement (sentence)
- abstract level
- relationship bet. concepts
Hypothesis = statement (sentence)
- empirical level
- relationship bet. variables, a proposition that is empirically testable
an unproven proposition
9
10
(Review Chapter 5)
Ladder of Abstraction
- Terminator - Three; Movie; Entertainment
- Pantene; Shampoo; Consumer Product
- Activity; Tennis; Sport
11
(Extra Handout/Sheet)
Dependent Variable
(+)
Teaching Aids
Class Size
Class
Atmosphere
(-)
Student Learning
(+)
??
??
Research Topic ?? (of the above Conceptual Model)
a. Who could learn more effectively?
b. What are characteristics of learning?
c. What are consequences of student learning?
d. What are determining factors that influence student learning?
e. What are causes and effects of student learning?
12
Types of Variables
Independent Variable
(Predictor Var.)
Dependent Variable
(Criterion Var.)
(-)
Class Size
Student Learning
Class size:
A. How many students in one class? Fill the number in the blank
_____ students
(continuous variable can find mean value)
B. How big is your class?
___ Small ____ Medium ___ Large (categorical variable)
Intervening Variable (coming in between IV and DV)
Class Size
(-)
Level of Student
concentration
Student Learning
(-) (+)
Student Learning
Interestin
g topic
discussed
13
Chapter 6
Exploratory Research in order to clarify situation or help define
the right problem.
- Initial / preliminary research
- Need subsequent research
- Provide qualitative analysis / data
- To understand nature of problem, but not to solve the problem
-
14
15
Quantitative
Unstructured questions
Structured questions
Exploratory research
16
(Review Chapter 6)
Concept Testing is another example of Exploratory Research
- test the tendency of acceptance for any new product/service idea
- screen the most acceptable concept from many alternatives
- may apply the projective technique, e.g. word association
Pilot Study versus Experience Survey
A. People are selected for interview because they are articulate
(knowledgeable) rather than being representative.
B. People are selected for survey because they represent the target
population of the study.
C. _____ is similar to a survey method except that it is a small-scale
study and based on qualitative analysis.
Pilot Study vs Experience Survey vs Case Study
A. Su Su DVD rental shop intensively investigated the computerized
DVD renting process of Block Bluster. This investigation is
called __________ .
B. Discussing the economic crisis issue with the Minister of Finance
may be referred as the ___________ .
C. Discussing the impacts of crisis on SMEs with entrepreneurs or
investors is called the ______________ .
Moderator vs Depth Interviewer
A. He or she must be a high-skilled interviewer who can stimulate a
respondent to elaborate on the topic discussed.
B. His or her job is to be a good listener who can also develop a
rapport with everyone in order to stimulate spontaneous answers.
17
Chapter 7
(Overview)
*only p.159-169*
Secondary data
: data gathered before the current research
not for the purpose of the current needs of researcher.
Primary data
: data gathered at the current research and for the purpose of the
current need of researcher.
Data conversion
: changing original form of data to format suitable for current
research.
Advantages of secondary data
- less expensive
- quickly obtained
- get some info. that may not be obtained by primary data.
Disadvantages
- not designed to meet current researchers need
- outdated
- different in definition of terms
- different units of measurement
- no proof for accuracy
Sources of secondary data
1. Internal
= usually organization has many internal records kept for future use,
but researcher may ignore these records (e.g. Annual Report, Memo,
Marketing Plans, etc.)
2. External
- public library
- government sources (e.g. BOI, Bank of Thailand, etc.)
- commercial sources (e.g. companies that sell the data)
- media sources (e.g. newspaper house, TV stations, etc.)
- trade associations (e.g. Association of Garment Trading, etc.)
18
Secondary Data
gathered and analyzed
before the current research
Vs
Primary Data
gathered and analyzed
at the current research
- data vs information
- data conversion = data transformation
- Advantages and Disadvantages of Secondary data
- Sources of Secondary data ---- internal
external
- Reference or Bibliography (base on APA style see the next page)
Objectives of secondary data research
1. fact finding simple
2. model building more complex (conceptual model)
3. data mining the most complex
using computer to categorize people/things
19
Chapter 8
Survey (an obtrusive research method)
: method of primary data collection based on communication between
researcher and respondents (based on asking questions method)
Verbal
Written questionnaire
- popular method, quick, accurate, inexpensive
- mostly for descriptive research measuring awareness, attitudes,
behaviors, opinions.
Types of survey (based on Time)
1. Cross-sectional study more than one group at one point in time
2. Longitudinal study one group for a series of time
- Panel study same group (same individuals) for a series of time
- Cohort study more than one group for a series of time
Survey error or Total error (Exhibit 8.1, p.188 )**
not precise = 1. Random sampling error (chance or probability error)
: chance variation (nobodys mistake, it is mistake by
chance)
: results from our sample results from the real
population
not accurate = 2. Systematic error
: mistake in conducting research (human error)
: it is divided into Respondent error and
Administrative error
20
21
(Review Chapter 8)
Acquiescence Bias vs Extremity Bias
SA = strongly agree
DA = disagree
A = agree
N = neutral
SD = strongly disagree
Statement of attitude
1. Internet service at ABAC is of
good quality.
2. Computer lab at ABAC Bangna is
convenient.
3. Staff at computer lab is very
helpful.
4. Operating hours of the lab is
reasonable.
5. Fees at the lab is reasonable.
SA
x
A
x
N
x
DA
x
SD
x
Or
1.
2.
3.
4.
5.
4
4
4
4
4
3
3
3
3
3
2
2
2
2
2
1 bad quality
1 inconvenient
1 unhelpful
1 unreasonable
1 unreasonable
22
F. acquiescence bias
G. interviewer error
H. auspices bias
I. extremity bias
J. self-selection bias
1.If a researcher noticed that some security guards with higher authority tended to
refuse to answers the interviewers questions about ghost rumor on campus.
2. Researchers forgot to ask some personal data from some respondents.
3. When a job applicant is asked by HR manager (during job interview) about how
long he or she intends to work for this company, the possible bias may come from
______ .
4. If Oishi conducts it own survey about preferences of Japanese food of teenagers,
the findings may suffer ______ .
5. The question, Do you come from a good family?, might suffer ____ .
6. Respondents disagreed very strongly with all ideas proposed in the
questionnaire items.
7. Researchers wanted to study about MSN in internet, so they interviewed
senior citizens and farmers in rural towns.
8. Housewife respondents tend to agree with all proposed qualities of the new
product survey.
9. The researcher filled in the income level for some respondents who did not
write the answers on the questionnaire.
10. Bias from a hotel guest who was so upset with the room service and after that
he wrote a complaint on the questionnaire placed near the check-out counter and
rated the service quality of the hotel very negatively.
23
(Chapter 9)
Face-to-face
Verbal
(Interview)
Mall intercept
interview
Survey
Telephone
Written
(Questionnaire)
Sent by hand
Sent by mail, email, fax
Not sent
(self-administered questionnaire)
Survey
Personal
(door-to-door)
interview
Central
Location
telephone
interview
Computer
Assisted Tele.
Interview
Computerized
Voice
Activated
Tele. Interview
Cross-sectional
Longitudinal
Panel Study
Cohort Study
24
Chapter 9
Survey: is divided into 3 types (based on method)
1. Personal interview door-to-door
(face-to-face) mall intercept interview
(or high traffic area interviews)
2. Telephone interview Central Location (has interviewer and paper)
CATI - Computer Assisted Telephone
Interview (has interviewer and computer)
Computerized Voice Activated Telephone
Interview (use answering machine system,
no human interviewer)
3. Mail questionnaire
Personal Interview (about 1 - 1 hr.)
Advantages
1. Immediate feedback
2. Probing complex questions Can you tell me more about
3. Longer period
4. Can collect more complete info.
5. Can show product, visual aids
6. High participation
Disadvantages
1. no anonymity between researcher and respondents
2. influence of the demographics of interviewer (interviewer bias)
3. interviewer cheating
4. cost
5. callback problem
25
26
Mall
Intercept
Telephone
Interview
Questionnaire
- immediate feedback
- short time, short question
- cover wide area with low cost
- sensitive / personal issue
- anonymity of respondents
- low response rate problem
- easy to call back
- easy to probe
- get complete information
- highest cost
- no interviewer bias
- can show pictures/products
- most structured questions
- respondents convenience
- high versatility (= flexibility)
- appropriate for retired +
housewives
- appropriate for busy
executives
- appropriate for less-educated
people
- appropriate for high-educated
27
people
(Review Chapter 9)
Response Rate =
100 x (No. of questionnaires returned and completed without bias)
(No. of questionnaires sent to eligible persons without problems)
Example
You mailed 1,000 questionnaires, 200 were returned to you as wrong address.
After 4 weeks, 600 were returned and completed, but 50% of those returned
contained acquiescence bias and extremity. What is the response rate of this
study?
Answer = 37.5%
Different Types of Survey
A. mail survey
E. door-to-door interview
B. computer-assisted telephone interview F. mall intercept interview
C. self-administered questionnaire
G. telephone interview
D. computerized voice activated telephone interview
1. Honda would like to identify the characteristics of potential customers who
showed interest on their new car models during the MOTOR SHOW.
2. Department of Public Health used its own budget to study about the usage of the
30-Baht healthcare policy from Thai citizens all over the country.
3. A-Day magazine wanted to identify the demographic characteristics of its
subscribers and measure their perceptions about the magazine.
4. The survey technique that allows researchers to do a lot of probing in order to
get as much as information as possible.
5. The survey that conducts the telephone interview without human interviewers.
6. The survey that allows data collection and data processing (keyboarding the
answers and coding) to occur simultaneously.
7. Right after the new campaign of drug problems was on air, the researchers used
the very short and simple survey questions to measure the awareness of target
respondents toward this campaign.
28
Chapter 10
(Overview)
(only p.236-244)*
Observation: scientific observation observer
subjects
Being the scientific observation if it .
has research objectives
is planned systematically
is record systematically
can be checked for validity + reliability of results
Observation is unobtrusive method =
without communication
without asking any questions
29
3.
-
eye-tracking monitor
pupilometer
psychogalvanometer
voice pitch analyzer
30
A. content analysis
B. role playing
1. BTS would like to estimate the number of student passengers using its service
during the rush hours.
2. To study the interpersonal relationship between students and teacher in a
classroom of western culture. A student was asked to behave as an American
teacher and the expected results were observed.
3. We can know that the current government aims to solve the poverty problems as
its top priority by investigating the government policies announced in the
Parliament Meeting.
4. A car company wants to determine the most popular radio stations listened by its
customers by checking at the setting programs from one to six radioprogrammed buttons of all cars parked at the check-up service.
31
Chapter 11
(Overview)
(only p.255-265)
Experiment
: the study of causal relationship between variables in the condition
controlled by research (experimenter and test units)
units of analysis = test units in the experiment
1. Select independent variable.
2. Manipulate or make some changes on independent variable.
= experiment treatment (variation of independent variable)
3. Observe and measure an effect on dependent variable.
- experimental group: receive experiment treatment (= real treatment)
- control group: not receive experiment treatment
(sometimes, receiving placebo = fake treatment)
field experiment: in natural setting
laboratory experiment: in artificial + controlled setting
1.
2.
3.
4.
Constancy of condition
: a way to control errors by keeping conditions / situations the same in
experimental groups (only treatments that are not the same)
: if there is no constancy of condition, there will be constant error
32
Dependent variable
Lotion
Skin
Music
Work performance
Extraneous variable
Irrelevant variable
- sunlight, soap, food, body weight
- working hours, type of music, salary, religion of security guards
Experimental Treatment : how to manipulate the independent
variables
- Lotion using Nevia Lotion twice a day for 3 months
- Music listening to country music during rush hours
Testing Unit (= dependent variable of ??)
Women measure skin of women
Security guards measure work performance of security guards
Example
Sending employees to training at Phuket Island seemed to result in
higher work morale. The least we can conclude about out-of-thehouse training and morale is that these two variables have
____.
(causal relationship or concomitant variation)
34
35
36
Chapter 13
Measurement: a process of assigning numbers to objects or concepts
- Concept cannot be directly measured, so researcher has to give
Operational Definition to the concept to be measured in the
research.
Operational Definition: giving empirical meaning to a concept by
specifying the activity or how to measure that concept
( Conceptual Definition abstract meaning)
Rule of measurement: rule or guideline of how to assign the number
Scale:
low level 1.
2.
3.
high level 4.
( single measure)
Index Measure (Composite Measure): asking more than one question
to measure one concept which is complex, e.g. social class, attitude, etc.
37
38
Important Notes
When a scale is reliable, it does not need to be valid. But when the
scale is valid, it should be also reliable. Reliability is a necessary
condition of Validity.
39
- Nominal
- Ordinal
- Interval
- Ratio
percentage, mode
percentage, mode, median, ranking
percentage, mode, median, arithmetic
mean (x), standard deviation (S.D.)
percentage, mode, median, x, S.D.,
index number, geometric mean
40
41
Review Chapter 13
Four Levels of Scale
(Nominal, Ordinal, Interval, Ratio Scale)
Do you like Pizza?
a) __ Yes
__ No
b) __ like so much
__ like
__ dislike
Rank 1 to 3 for the following pizza brands that you like. 1 means
like the most and 3 means like the least
__ Pizza Hut
__ The Pizza Company
__ Narai Pizzeria
42
__ 2. No
b) __ 1. dislike so much
__ 3. like
c) like so much
__ 2. dislike
__ 4. like so much
4 3 2 1
dislike so much
43
To measure gender
__ Female
Chapter 14
Attitude: enduring disposition toward something
1. affective component feeling / emotion
2. cognitive component thinking / thought / belief / awareness
3. behavioral component buying intention / behavior intention
expectation
Hypothetical construct: any variable that cannot be directly
observed, but must be measured by the indirect way (such as by
survey) e.g. attitude, perception.
Four techniques to measure attitude
1. Ranking: list a rank order of preference
2. Rating: estimate magnitude and value or scores
3. Sorting: classify objects according to some criteria
4. Choice: choose from alternatives
Attitude Rating Scales
1. Simple attitude scale: 2 choices to choose
2. Category scale: more than 2 choices to choose
3. Likert scale: rate degree of agreement or disagreement
4. Semantic differential scale using bipolar adjectives cheap __ __ __ __ __ expensive
- fast 5 4 3 2 1 slow
5. Numerical scale
- 3 2 1 clean +1 +2 +3
6. Stapel scale
- clean
dirty
7. Graphic rating scale
- Happy face scale
- Ladder scale
for study of kids or small children (low literacy = cant read)
= using picture as scale
45
shopping time
duration of shopping
amount of money paid / shopping
attitudes toward store
response latency in buying perfume
reasons for choosing a perfume
marital status of shoppers
46
Ranking technique
Rating technique
Sorting technique
Choice technique
__ Mitsubishi
__ Honda
__ Nissan
2. Rank 1 to 4 to the car that you like 1 means like the least and 4
means like the most.
__ Toyota
__ Mitsubishi
__ Honda
__ Nissan
Like
3.2 Honda:
Like
3.4 Nissan:
Like
47
Filter q.
Simple dichotomy
Determinant choice
Frequency determinant
Semantic differential
Graphic rating scale
- Checklist q.
- Simple attitude scale
- Category scale
- Likert scale
- Numerical scale
- Stapel scale
A.
B.
Agree
C.
-2
Agree
D.
-1
+1
+2
Disagree
SA A
__ __
__ __
N
__
__
D
__
__
SD
__
__
G.
___ agree
___ neutral
___ strongly disagree
H.
I.
48
J.
____ coffee
___ tea
____ soft drink ___ fruit juice
(you can choose more than one beverage)
K.
Coffee
L.
M.
____ at home
N.
____ like
____ dislike
O.
____ like
____ so-so
P.
____ shopping
Q.
R.
____ never
____ rarely
____ sometimes
____ often
____ always
Tea
____ dislike
___ listening to radio
49
Chapter 15
Questionnaire: - instrument to collect data in survey research
- should have grammar and good common sense
Two qualities to judge good questionnaire
1. Relevance
2. Accuracy
Two types of questions
1. Open-ended question
2. Fixed-alternative question
2.1
2.2
2.3
2.4
2.5
2.6
2.7
Simple-dichotomy question
Determinant-choice question
Frequency-determination question
Attitude rating scales
Checklist question
Filter question
Ranking question
50
Review Chapter 15
Questionnaire
Each question below represents What type of question? (from
Ch.15)
A. What is your opinion about ABAC Bang-Na?
___ No
52
N
__
__
D
__
__
SD
__
__
I.
- Food: delicious ___: ___: ___: ___: ___: ___: ___: not delicious
- Service: friendly___: ___: ___: ___: ___: ___: ___: unfriendly
- Area:
dirty ___: ___: ___: ___: ___: ___: ___: clean
J.
(Please circle one number)
- Food: delicious
7
- Service: unfriendly 1
6
2
5
3
4
4
3
5
2
6
1
7
not delicious
friendly
K.
- Food:
-3
-2
-1
delicious
+1
+2
+3
L.
+3
+2
+1
friendly
-1
-2
-3
- Service:
M.
(Please mark on the line)
- Food: delicious -------------------------------- Service: friendly --------------------------------
not delicious
unfriendly
53
Not delicious
P. Have you ever tried McDonalds?
_____ Yes (Go to Part 3)
____ No (Go to Part 4)
Q. What type of technique is used below? (choice, rank, rate, or sort?)
Please classify the following characteristics or factors under Japanese car
or European car
1.
2.
3.
4.
More safety
More reasonable price
Better after-sales service
Save energy
Japanese Car
?
?
European Car
?
?
54
___ No
___ No
55
Research Problem
- To measure job satisfaction of factory workers.
- To study if marital status can influence job satisfaction of
factory workers.
- To study if gender can influence job satisfaction.
Questionnaire
____ male
____ female
____ married
OR
56
____ No
____ Siemens
____ other, ________
____ Hutch
Never
8. How much do you pay for the monthly mobile phone bill?
_____________
9. What is/are problems you find from the current mobile phone you are
using?
____ not have many functions
____ not look modern
____ not enough service center
____ not appropriate size
____ too big size
____ no advanced technology
____ other, ________________
57
10. What is your attitude toward the following characteristics of the mobile
phone you are using now?
10.1 Price:
very reasonable 5
4
3
2
1 very unreasonable
10.2 Design:
modern
5
4
3
2
1 out-of-date
10.3 Technology: high advanced 5
4
3
2
1 low advanced
10.4 Promotion:
satisfied 5
4
3
2
1 dissatisfied
11. Where did you buy the current mobile phone?
____ from the authorized dealer
____ from the unauthorized dealer
____ from your friend/relative
____ from the service center
12. The service center should be_____________ .
widely available
5
4
3
1 opened 24 hours
13. How often do you go to the service center of your current mobile phone?
____ less than 2 times a month
___ 3-4 times a month
____ 5-6 times a month
___ more than 6 times a month
14. What is the most appropriate price of the good mobile phone in your
opinion? (in Baht)
____ less than 4,000
____ 4,000 8,000
____ 8,001 10,000
____ more than 10,001
15. Are you satisfied with your current mobile phone?
____ Yes
____ No
16. In case you have a chance to buy a new mobile phone, you will buy
____________ brand.
17. What is the first word came to your mind when you think about the most
wanted mobile phone?
_____________________________
18. What is the chance of you buying a new mobile phone?
Never -2
-1
Soon
+1
+2
19. What is the chance of you switching to the other brand of mobile phone?
Switch sure
May or may not
Dont switch sure
58
Chapter 16
Sampling: method of choosing some respondents from a total group of
respondents to survey and draw a conclusion to that total population.
- sample: some items that researcher chooses to study sampling
units (each unit to be sampled)
- population: a total + complete group of items sharing common
characteristics. (population vs universe) population elements
- census: a national survey or a study of everybody in a population
- we can do sampling by using sampling frame or working
population to draw a sampling unit from
Sampling frame: a list of names / elements in a population such as
telephone directory (listed by surnames of house owners), class
attendance sheet, membership list, reverse directory (listed by city or
address) systematic list vs non-systematic list
Why do we do sampling? (Why dont we do the census?)
1. save energy, time, labor, money: less job more saving
2. get accurate and reliable results: less job lower human errors
3. avoid destruction of all elements or the whole things such as blood
test by sampling 5 cc bloodnot taking all blood, or Q.C of finished
products in the production department
Seven steps of Sampling Process
1. define target population
2. select the relevant sampling frame
3. decide sampling method between
probability sampling randomly
non-probability sampling non-randomly
4. plan detailed procedures for choosing sampling units
59
pq
n
* n = sample must be drawn by probability sampling only!!
SP
1.
2.
Z 2 pq
2
E
3. Adjusted n =
( N n)
x
( N 1)
(N = population size)
61
E. sampling frame
F. probability sampling
G. non-probability sampling
1. Sample ________.
2. Sampling _________ .
3. Researchers usually collect data from the _______ to get results, then
they will generalize or infer those results to explain the __________ .
4. Researchers use the ______ as a source to select sampling units for
their sample. If every unit has an equal, known chance to be selected,
the method is called _______ .
5. In the more homogeneous sample, the method of _______ is better.
6. In the more heterogeneous sample, the method of ______ may be
possible.
Random sampling error versus Systematic error
A. Increasing sample size would reduce _______ error while
increasing the accuracy of data collection would reduce _______.
B. To get the precise data, ______ error should be lowered and to
get the accurate data, ______ error should be lowered.
62
Sampling Method
A. Quota sampling
B. Stratified sampling
C. Judgment sampling
D. Snowball sampling
E. Convenience sampling
F. Simple random sampling
G. Cluster sampling
H. Systematic sampling
1. There are only six pediatric medical doctors in Thailand. They are
specialized in the heart surgery. The best way to sample these experts is
by ________.
2. In a marketing survey, if all customers are divided into two groups:
urban and rural groups. A sampling unit is randomly drawn from each
group to study.
3. In a test marketing of coffee consumption of Thai people in the
northern part of Thailand, we randomly selected 3 northern provinces
(Chiangmai, Chiengrai, Lampang) and then we randomly selected
people in these three provinces.
4. Researcher decides to interview 100 customers who are below the age
of 30 and 50 customers who are above 30 years old.
5. BBC news reporter interviewed the first person they met on the street
in front of his TV station.
6. In a manufacturing firm of instant noodle, an inspector does a
quality control check on the product once a month. He usually goes to
the inventory room and randomly selects 50 packs of instant noodle for
the quality check. His technique of selection is called _______ .
7. During an investigation of corruption issue within hospitals in rural
areas, the committee decided to inspect only those hospitals they
suspected of being corrupted. The committees selection is based on
________ .
63
64
Chapter 17
Sample Size
Three factors to determine sample size (n)
1. variance value or heterogeneity of population: variance, sample
size
2. magnitude of accepted error (E): error, sample size
3. confidence level: confidence level, sample size
One principle about sample size
- When E is double, n will be equal to of the original n ***
Ex.
In the first study, the E was 4, the sample size used was 2,000. What
should be the sample size of the second study if the E becomes 8?
= E from 4 to be 8 double, so use the above principle to get n
= n of the second study is 1/4 (2,000) or equals to 500.
Formula to calculate a sample size (n)
ZS
E
1.
2.
Z 2 pq
2
E
3. Adjusted n =
( N n)
x
( N 1)
(N = population size)
65
1.96 120
10
553.19
Z 2 pq
E2
0.035 2
753
3. Based on the above problem, if the entire working people are 10,000
people, what should be the more appropriate sample size?
5% (N) = .05 (10,000) = 500
n from the above problem is 753, and it is more than 500 (5%N)
So, we need to find the Adjusted n to get the more appropriate sample size
Adjusted
N n
N 1
x n
10,000 753
x 753
10,000 1
= 724
66
67
Data Processing
Coding
Data entry
(keyboarding)
Error checking
and verification
Editing
Data analysis
Descriptive
analysis
Inferential
analysis Univariate
analysis
Inferential
analysis Bivariate
analysis
Interpretation
Inferential
analysis Multivariate
analysis
of Result (Findings)
Conclusions
Recommendation
Report
68
Chapter 19
Data Processing:
Editing
Coding
Editing:
Data Entry
Data Analysis
no
- no overlapping case.
Data entry: a keyboard operator will type all codes into a computer for
SPSS or any statistical software to do data analysis
Re-coding: change some codes to better fit the final results adding or
collapsing (Data Transformation)
Error checking: data cleaning stage = the last step of coding before
researchers go to do data entry (or after data entry)
Data storage (in computer)
- Field a collection of characters that represent a single type of data
(such as a field of gender data, a field of age data, income field)
- Record a collection of related fields (such as a record of Personal
data part, a record of User part questions, a record of Non user
part questions)
- File a collection of related records (such as a file of Questionnaire
Set 1, a file of Questionnaire Set 2)
70
Review Chapter 19
Editing
Ex. 1
Q1. Do you have a PC at your house?
___ Yes
___ No
___ Compaq
___ Acer
___ Other, specify ___________
Ex. 2
Q1. Do you have a PC at your house?
___ Yes
___ No
___ IBM
___ Compaq
___ Acer
___ Other, specify ___________
Ex. 3
Q1. What is your monthly income?
I dont know
Q2. What is your family monthly income?
--- or
I have no idea
Q3. What is your savage income?
Dont know!
71
Test Tabulation
Sample size = 30
.
.
15. Why dont you like Rungsit Campus?
__
__
__
__
__
__
__
__
Recode 5, 4
3
2, 1
3
2
1
=
=
=
72
__ 2. 16 20
__ 3. 21 24
__ 4. 25 30
__ 5. more than 30
Recode 1,2, 3
4, 5
1
2
=
=
73
I enjoy cooking
__
__
__
__
__
6.2
__
__
__
__
__
6.3
I am a good cook
__
__
__
__
__
__
__
(5)*
__
__
__
(X6.1 + X6.2 +
X6.3 + X6.4 +
X6.5 + X6.6) 6
Their answers are: 26, 20, 38, 55, 19, 277, 42, 35, 22, 50
x 584
Find average age = n 10 58.4
334 = 33.4
10
- You need to look through all codes or answers on the coding sheet
(or at the Data View page of SPSS) AGAIN.
** To prevent Garbage-in Garbage-out
(wrong input) (wrong output)
74
Chapter 20
Data Analysis
Level of Data
Ratio
Interval
Ordinal
Nominal
Descriptive
Statistics
Central tendency
- Mean **
- Mode
- Median
Dispersion
- Range
- Variance, S.D
- Median
- Mode **
- Percentage
- Counting
Frequency
Inferential
Statistics
Standard error
of mean
Hypothesis Testing:
Univariate
Bivariate
Multivariate
Hypothesis Testing
Difference
Correlation
- T-test
-Pearsons r
- Z-test
or Correlation
- ANOVA coefficient
- Regression
Standard error
of proportion
-Chi-square
test
76
55
45
100
Total
Yes
No
Total
Male
10
40
50
77
n = 100
M
(50)
F
(50)
= quota sampling
Conclusion: Female like shopping more than Male. correct??
B. Do you like shopping? by gender
Male
Female
Total
Yes
20
40
60
No
10
30
40
Total
30
70
100
n = 100
M
(30)
F
(70)
random sampling
Conclusion: Female like shopping more than Male. correct??
We need to compare Percentages, not Frequency
Total
60
40
100
60%
40%
100%
Conclusion: Male like shopping more than Female. (67% > 57%)
78
Elaboration Analysis
Do you like shopping at Central? By gender
Male
Female
Yes
60
80
No
40
20
Conclusion: Female like shopping at Central more than Male (80% 60%).
Do Female like shopping at Central more than Male always? or
In all circumstances, or no matter what, that Female will like shopping
at Central more than Male? We want to test this doubt by doing the
Elaboration Analysis adding one more variable marital status
Case 1: Do you like shopping at Central? By gender and marital status
Male
Female
Married
Single
Married
Single
Yes
30
30
40
40
No
30
10
20
0
Conclusion : Married female respondents like shopping at Central more
than married male (40% 30%). And single female also like shopping
at Central more than single male (40% 30%).
Therefore, based on Elaboration Analysis between attitude about
shopping and gender by adding marital status, it is concluded that
female still like shopping at Central more than male. Adding the
marital status does not change the original relationship.
So the marital status is called _________ , and the relationship between
attitude about shopping and gender is called ____________ . (in Case 1)
Case 2 : Do you like shopping at Central? By gender and marital status
Male
Female
Married
Single
Married
Single
Yes
10
50
40
40
No
30
10
20
0
Conclusion : ??
Therefore, based on Elaboration Analysis ?? (what is found??)
79
Rank 1
(4x1) +
(1x1) +
(7x1) +
Rank 2
(4x2) +
(3x2) +
(2x2) +
Rank 3
(2x3)
(6x3)
(1x3)
Total
= 18
= 25
= 14
Result
2nd
3rd
1st
Rating
Rank 1
(4x1) +
(1x1) +
(7x1) +
Rank 2
(4x2) +
(3x2) +
(2x2) +
Rank 3
(2x3)
(6x3)
(1x3)
Total
= 18
= 25
= 14
Result
?
?
?
80
Chapter 21
Statistical Analysis
1. Descriptive Analysis: counting, percentage, central tendency
(mean, mode, median), and dispersion (range, variance, S.D)
Ch.17, 20
2. Inferential Analysis:
(2.1) Estimating population parameters (standard error of mean
and standard error of proportion) Ch.17
(2.2) Testing Hypothesis
Univaritate Analysis: investigate significance of hypothesis of one
variable Ch.21
Bivariate Analysis: investigate significance of hypothesis about two
variables (test of relationship or test of difference) Ch.22-23
Multivariate Analysis: investigate significance of hypothesis about
more than two variables (Advanced statistical analysis) Ch.24
Three factors to determine which method of statistical analysis
-What is proportion of those who go to EGV?
-What are three most important factors in ..?
-What is average satisfaction level of ?
81
Parametric Statistics: statistics used for interval and ratio scale and it
assumes that population has a normal distribution
Non-parametric Statistics: statistics-used for nominal and ordinal scale
and there is no assumption about normal distribution of the population
82
Hypothesis Testing
1. Null hypothesis (HO): assumption about status quo, no change
e.g.
= 19 or 19, or 19 (must have equal sign)
= There is no difference; There is no relationship
2. Alternative hypothesis (Ha): opposite of null hypothesis
e.g.
19 or < 19, or > 19
= There is a difference, There is a relationship
Type I error ( Alpha): reject null hypothesis when it is true
Type II error ( Beta): accept null hypothesis when it is false
type I error will lead to type II error
type II error will lead to type I error
- We cant completely eliminate any type of error, so researchers keep
one type of error constant, normally setting Type I error at 0.05 and
calling it as significance level of 0.05 there are less than 5 times out
of 100 times that we make erroneous decisions (reject the null
hypothesis when it is true/correct).
Univariate Statistical Method
1. T-test: used when n is small (n 30), S.D. is not given
Z-test: used when n is large (n > 30), S.D. may or may not be given
To compare mean or average with hypothetical value in the null
hypothesis
2. Chi-square test for goodness of fit ( 2)
To compare distribution of a variable between two or more groups of
respondents/ or to test a difference in frequency distribution (frequency table)
3. Z-test of a proportion
83
84
(critical value)
X
S/
X
S/
2 value
Z value
S
X
or
O E
E
p -
O2
E
or
pq
n
Note: - Chapter 21-24 Read Hand-out about:Statistics for Data Analysis in Business Research Chapter
21-24
***Please read Text Book***
- Chapter 24 Only overview of what is the Multivariate
Analysis
85
Age
q.5
choice scale
(2 groups, e.g. gender)
q.16
Gender
- One-way ANOVA
q.17
Age
- Simple regression
87
__ dislike
__ over counter
__ by phone
__ via internet
Efficiency of security
Scale
data
+
+
Cleanliness of restroom
+
Attitude about
Liking to see
movies at EGV
Comfort of seat
88
Efficiency of security
Efficiency of security
89
Test of Difference Between more than two groups: e.g. age groups
- Respondents at age of more than 30 rate the efficiency of security
of EGV differently than other age groups. (There are total of five age
groups) (q.17 q.9.1)
There is a difference in attitude about efficiency of security of EGV
between people with different age groups.
Age
Efficiency of security
90
Chapter 25
Written Report Format
1. Title page
2. Letter of Transmitted
- researcher client
Objectives
(skip)
91
The End
93