Please enable JavaScript.
Coggle requires JavaScript to display documents.
Dissertation Lecture 3 Research methods and data analysis - Coggle Diagram
Dissertation Lecture 3
Research methods and data analysis
Descriptive
often choose cross-sectional
sample survey
make a simple prediction
Exploratory
methods
Expert/experience surveys
pilot surveys
case studies
secondary data: qualitative
qualitative researech
Research designs
Deduction: theting theory
Induction: building theory
Mixed methods: Combing research approaches
process
Theory
use the theory to make a prediction
prediction
design an experiment to test the prediction
experiment
perform the experiment
observation
create or modify the theory
Experimental design
Nature of experimentation
Start from the theory and find
independent variable
dependent variable
2 types experiments
Laboratory
effects of seeing advertisement A vs. B on attitude toward the brand
consider running such an experiment in a lab vs. at home
the independent variable is manipulated and measures of the dependent variable are taken in an artificial setting contrived to control extraneous variables
Lab experiments maximize internal validity, external validity is a question
you can control the noise in the lab, home is noisy, your attitude may be get affected
Field (Now during quarantine cannot do)
effects of a point of purchase display on sales
Consider potential confounds
Experiments conducted in a natural setting.
Lab experiments maximize external validity, internal validity is a question
Validity of experiments
Internal Validity
Lab is better
External Validity
Field is better
Evidence required to infer causality
Experiment example
Two treatment group
an control group
Two treatment group
Factorial(statistical) designs
simultaneously more than one IV, GENERATING A COMBINATION OF ANY POSSIBLE LEVEL OF ONE iv TO THE OTHER LEVELS OF ALL THE REMAINING IVs
Types of design
3
Problem: confounding (extraneous) variables
operational confounds
a confounding variable can be classified as properational if the IV does not manipulate
Procedural confounds
variables relate to a situational condition rather than to the relationships between independent and dependent variables
you should experiment at the same time same day
Person confound
you should randomizaiton of the sample
T-TEST
Independent t-test
Dependent t-test
One-way ANOVA
when we want to compare means we can use a t-test
Cross sectional VS. Longitudinal research
study design depends greatly on the nature of the research question
Cross-sectional
involves one set of measurements
Longitudinal research
involves a series of measurements over time
Conceptualization and Operationalization Table
不要自己想問題，要從literature裡面挖出來
做一個表格
Measurement scale
量表 (From literature)
the measurement items can come from different paper
有一個變量是process moitoring
you should provide the statements or question that you should used in your questionnaire
Types of measure scales
名目
等距
等比
序
Questionnaire design
Should contain meaures of all the variables that are in your hypostheses
valid and reliavle
simple and easy to understand
demographic
proof read
aoid jargons
pilot test the questionnaire in a sample from the same population
Sampling methods 1
(Methodology section) 2-3000 words
Sampling frame: complete list of all the cases in the population from which your sample will be drawn
Probability sampling
simple random
systemativ
stratified random
cluster
Sampling methods 2
Quota
a stratified but non-random sample
snowball
convienece
Quantifative data analysis
Descriptive
Computing me
Inferential
Reliability
One fundamental concept in measurement
The extent to which th measures (or broadly, data collection techniques) will yield consistent findings
most widely used measure of reliability is Cronbach's alpha
Ranges from 0 (NO reliabity ) to 1 (complete reliability)
Scales with alpha values above 0.7 are considered reliable
Tutorial
Correlation
Linear and multiple regression
(For testing your hyposis)
Qualititive research
Qualitative methodology
Theory
theory validation empiracal research
Reality
theory formation exploratory search
methodology
Type of research question
what? why?
type of questions
open questions
what type of data will you need to coleect to answer th Research questions
explore, explain and understand phenomena- in-depth-interview
When to use qualitative research
Build theory and propositions
develop hypotheses for furhter testing and for quantitative questionnaire develoment
Understand thefeelings, values, and perceptions that underlie and influence behaviour
Perception of marketigng
Conducting fieldwork
create a discussion guide to use during the interviews
identify and contact respondents to arrange interviews
Have a consent form ready
Record the interviews
Observe behaviour during the interviews and took notes about it
Agree with you supervisor on how to analyse the data
Transcribe the interview yourself
Develop hypotheses or propositions
Micro expressions
Left
truth
Right
Lying
Data analysis
Content anlysis
Narrative analysis
Discourse analyisis
Framework analysis
Grounded theory