Similar presentations:
Quantitative research in management: methodology. Introduction to IBM SPSS
1. Research methodology
RESEARCH METHODOLOGYOlga Konnikova
Ass.Prof. of Marketing Department, Saint-Petersburg State University of Economics
PhD in Economics
E-mail: [email protected]
2. Schedule
SCHEDULESeptember, 6, 2019 (Friday) 18.30-21.40 class № 2024
September, 13, 2019 (Friday) 18.30-21.40 class № 2024
October, 11, 2019 (Friday) 18.30-21.40 class № 2024
October, 29, 2019 (Friday) 18.30-21.40 class № 2024
November, 1, 2019 (Friday) 18.30-21.40 class № 2024
November, 8, 2019 (Friday) 18.30-21.40 class № 2024
Exam
2
3. Assessment requirements
ASSESSMENT REQUIREMENTSClass attendance/assignment
Hometasks
Final exam (written form)
Test (multiple choice questions)
Task (problem solution)
3
4. We will learn how to:
WE WILL LEARN HOW TO:1. Formulate research hypotheses
2. Select and conduct suitable types of statistical analysis to test hypotheses
3. Present the research results in the most understandable text and graphic form
4. Make predictions using multiple linear regression models and interpret their results
5. Conduct market segmentation and allocate clusters using the combination of characteristics
6. Predict the choice of the consumer according to the data we know about him
4
5. Agenda
AGENDA1.
Quantitative research in Management: methodology. Introduction to IBM SPSS.
2.
Data visualization. Descriptive statistics. Cross-tabulating (Contingency tables).
3.
Analysis of variance (dispersion analysis)
4.
Correlation and regression analysis
5.
Cluster analysis
6.
Summary
5
6. Topic 1.
TOPIC 1.QUANTITATIVE RESEARCH IN MANAGEMENT: METHODOLOGY. INTRODUCTION TO IBM SPSS.
7. Types of research
TYPES OF RESEARCHDesk-based (secondary) research (based on statistics, financial reports, …)
Empirical (primary) research (based on surveys, observation, experiments, …)
Qualitative (interviews, focus groups, expert surveys)
Quantitative (using different types of questionnaires (written, panel, telephone, PC, Internet, etc.))
7
8. Analyzing quantitative data
ANALYZING QUANTITATIVE DATADesk-based information:
- Statistics
- Financial statements
- CRM
- …
Empirical information:
- Questionnaires
- …
8
9. Hypothesis
HYPOTHESISHypothesis is the assumption of the connection of variables
In any hypothesis, a dependent and independent variable (-s) can be singled out
For each variable, you need to clearly understand in which scale it is measured (=how it cab be
measured?)
The number of hypotheses is unlimited
9
10. Hypothesis: dependent and independent variables
HYPOTHESIS: DEPENDENT AND INDEPENDENT VARIABLESIndependent variable (-s)
Hypothesis
Dependent variable (-s)
The number of training equipment and
training specialists in a gym has a positive
effect on the frequency of visits
When buying perfume, brand has a
stronger impact on customers than price
Women more often
unplanned shopping
than
men
do
10
11. Hypothesis: dependent and independent variables
HYPOTHESIS: DEPENDENT AND INDEPENDENT VARIABLESHypothesis
Independent variable (-s)
The number of training equipment
and training specialists in a gym Number of training equipment
has a positive effect on the Number of training specialists
frequency of visits
Dependent variable (-s)
Frequency of visits
Desire to buy
When buying perfume, the brand
Impact of brand
has a stronger impact on
Impact of price
customers than price
Women more often than men do
Gender
unplanned shopping
OR
Willingness to buy
OR
Purchase fact
Frequency of unplanned shopping
11
12. How to measure variables
HOW TO MEASURE VARIABLES3 main types of scales
1. Nominal scale
2. Ordinal scale
3. Quantitative scale
12
13. Nominal scale
NOMINAL SCALEObjects are classified by the presence and absence of a certain attribute
Categories of the attribute are not compared or measured in any way
Nominal scale is called binary if the number of categories is only two
E.g.: gender (male/female), native city (Moscow/London/New York), fact of purchase (yes or no)
13
14. Example of nominal / binary scale
EXAMPLE OF NOMINAL / BINARY SCALE1. What kind of soft drink do you prefer?
Dr.Pepper
Pepsi
Sprite
2. Would you continue to buy your favorite cosmetics brand if its price rose by 10%?
yes
no
14
15. Ordinal scale
ORDINAL SCALECategories have a logical order
We can compare the intensity of studied feature in the object, so we can dispose the categories on the
basis of "more - less", but without indicating how much more or less
There are different types of ordinal scales: scale of importance, Likert-type scale, interval scale, …
E.g. level of education (bachelor/master/PhD)
15
16. Example of ordinal scale
EXAMPLE OF ORDINAL SCALE1. Rank these beverages according to the degree of your preference
Dr.Pepper
Pepsi
Sprite
16
17. OrdinaL interval scale
ORDINAL INTERVAL SCALEThe categories in this type of scale are not only logically ordered, but also separated by certain intervals
Example:
2. Evaluate each drink on a 5-point scale, where 1 - do not like it, 5 - extremely like
Dr.Pepper
12345
Pepsi
12345
Sprite
12345
17
18. OrdinaL scale of importance
ORDINAL SCALE OF IMPORTANCEIntervals determine the degree of importance of a characteristic (-s)
Example:
3. Indicate how important each of these factors is for you when purchasing soft drinks
Factor
Almost
unimportant
Partly
unimportant
Sometimes
important,
sometimes not
A bit
important
Very
important
Critically
important
Price
Bottle design
Taste
Amount of sugar
18
19. OrdinaL Likert-type scale
ORDINAL LIKERT-TYPE SCALERespondents are given some statements and they are asked to what extent they agree or disagree with them
Example:
4. Below there are several statements regarding soft drinks. Please indicate the extent of your agreement or disagreement
with each of them.
Statement
Definitely
disagree
Generally
disagree
Partly
agree
Generally
agree
Definitely
agree
Strongly
agree
I believe that the taste of Pepsi is more
intense than that of Cola
A bottle of Sprite is more attracted to the
touch than a bottle of Pepsi
I believe that the consumption of soft
drinks is harmful to the stomach
19
20. Quantitative scale
QUANTITATIVE SCALEA measurement that uses absolute zero and, therefore, allows to make comparison of absolute values
of categories
Quantitative scale can be easily transformed into interval one
E.g. age, income, …
20
21. Example of quantitative scale
EXAMPLE OF QUANTITATIVE SCALEIndicate how many times a month you buy each of the drinks listed:
Dr.Pepper -
Pepsi -
Sprite -
21
22. Hypothesis: measuring variables
Scale forIndependent
measuring
Hypothesis
HYPOTHESIS: MEASURING
variableVARIABLES
(-s)
independent
variable (-s)
The number of training
equipment and training
specialists in a gym has a
positive effect on the
frequency of visits
Number of training
equipment
Number of training
specialists
Dependent
variable (-s)
Scale for
measuring
dependent
variable (-s)
Frequency of visits
When buying perfume, the
brand has a stronger Impact of brand
impact on customers than Impact of price
price
Desire to buy
OR
Willingness to buy
OR
Purchase fact
Women more often than
men
do
unplanned Gender
shopping
Frequency
of
unplanned shopping
22
23. Hypothesis: measuring variables
Scale formeasuring
Independent
Hypothesis MEASURING VARIABLES
independent
HYPOTHESIS:
variable (-s)
variable (-s)
The number of training
equipment and training
specialists in a gym has
a positive effect on the
frequency of visits
Number of
equipment
Number of
specialists
training Quantitative
Frequency of visits
Quantitative
Desire to buy
OR
Willingness to buy
OR
Purchase fact
Nominal (binary)
OR
Ordinal interval
training Quantitative
When buying perfume,
the
brand
has
a Impact of brand
stronger impact on Impact of price
customers than price
Women more often
than men do unplanned Gender
shopping
Scale for
measuring
Dependent variable
dependent variable
(-s)
(-s)
Ordinal
importance
Ordinal
importance
Nominal (binary)
Frequency
of
unplanned shopping
*Only binary
Quantitative
OR
Ordinal interval
23
24. Why is it so important (variables, scales, …)
WHY IS IT SO IMPORTANT (VARIABLES, SCALES, …)INDEPENDENT VARIABLE
TYPE OF VARIABLE
Quantitative scale
DEPENDENT
VARIABLE
Nominal / Ordinary
scale
Quantitative scale
Nominal / Ordinary
scale
Correlation and
regression analysis
Analysis of variance
(dispersion analysis)
Discriminant analysis
Cross-tabulating
(Contingency tables)
24
25. Hypothesis: main and alternative
HYPOTHESIS: MAIN AND ALTERNATIVEMain hypothesis (H0) states that there is no connection between variables
Alternative hypothesis (H1) states that there is any type of connection between variables
According to the methodology of quantitative research, we must first test the main hypothesis, and only if it
is refuted, check the alternative hypothesis
25
26. Hypothesis: main and alternative
HYPOTHESIS: MAIN AND ALTERNATIVEMain hypothesis
Alternative hypothesis
*gym
*perfume
*unplanned shopping
26
27. Hypothesis: main and alternative
HYPOTHESIS: MAIN AND ALTERNATIVEMain hypothesis
Alternative hypothesis
Gyms of the same network with different number of In the gyms of the network, where there are more
specialists and equipment are visited by the same specialists and equipment, the average number of
frequency
visits is more
When buying perfume, the price and the brand have Consumers more often buy perfume under the
the same influence on consumers
influence of the brand and not the price
Men and women equally often do unplanned shopping Women more often than men do unplanned shopping
27
28. Introduction to ibm spss
INTRODUCTION TO IBM SPSSSPSS (Statistical Package for Social Sciences OR Superior Performing
Software Systems) — a system (software package) of statistical information
processing that provides the user with a wide range of data transformation
and analysis capabilities, as well as visual representation of obtained results
28
29. How to change language in IBM SPSS
HOW TO CHANGE LANGUAGE IN IBM SPSS29
30. How to change language in IBM SPSS
HOW TO CHANGE LANGUAGE IN IBM SPSS30
31. The way of data organizing
THE WAY OF DATA ORGANIZINGTwo main windows (views) – window with the data (“Data View”) and window with the information about
variables (“Variables View”)
“Data view” window – rows contain observations, columns contain variables
Observation can be a respondent, product, brand, enterprise, …
Variable can be a question in the survey or some data we know about the observation
“Variables View” window – rows contain variables from the “Data view” and columns contain its description as
name, type, width and so on
31
32. Variables view
VARIABLES VIEWName – the working title of the variable (e.g. child) - no spaces, maximum 10 symbols
Type – type of the data in variable (numeric, string, data etc., e.g. numeric)
Label – full name of the variable (e.g. Number of children) – any number of spaces and symbols
Values – possible codes of the variable (e.g. 1 – high, 2 – medium, 3 – small).
All variables in nominal and ordinal scale should be coded.
Missing – missing values code in order not to take into account (optional, e.g. “98” – code for missing values)
Measure – type of the variable (scale, ordinal, nominal)
32
33. SPSS Statistics menu tools
SPSS STATISTICS MENU TOOLSFile – import, export functions, to save, to open, to create a new project etc.
Data – all types of work with data presentation, such as sort, split, select etc.
Transform – all kinds of data transformations, such as computing variables, recoding variables, missing
data replacing
Analyze – everything related to data analysis, such as descriptives, methods and models
Graph – graphical visualization of the data
33
34. Basic commands: data sorting
BASIC COMMANDS: DATA SORTINGData – Sort Cases
Example №1:
sort respondents by age
sort respondents by family status and education
34
35. Basic commands: data selection
BASIC COMMANDS: DATA SELECTIONData – Select Cases
Example №1:
select only men for future analysis
select only respondents with income more than 64000 rub.
35
36. Basic commands: Creating a new variable
BASIC COMMANDS: CREATING A NEW VARIABLETransform – Compute variable
Example №1:
Create a new variable ”Income in euro” by transforming Income variable according to the exchange rate
36
37. Basic commands: Recoding one variable into another
BASIC COMMANDS: RECODING ONE VARIABLE INTO ANOTHERTransform – Recode into Different Variables
Example №1:
Divide respondents into ”High level income” (>50000 rub.) and ”Low level income” (<50000 rub.)
Divide respondents into ”Younger” (< 35 y.o.) and ”Older” (>35 y.o.)
37
38. Task №1
TASK №1For 6 hypotheses in the file Task №1, determine dependent and independent variables, in which scale
each variable is measured, formulate main and alternative hypothesis
38