Marissa Dean, PhD

1
Marissa Dean, PhD SOCIO 7002 Research Methods in
Criminal Justice
2
Lecture 3 General Issues In Research Design
3
Objectives for Lecture 3 Research Design
To continue our discussion on how to formulate a
research question To understand the methods
involved in selecting sample populations Discuss
the requirements of Progress Report 1
4
Research Design Establishing Causation
Our ultimate goal in research is to describe,
understand and predict the relationship between
key variables (e.g. what causes other things to
happen?) Social science research is highly
probabilistic (e.g. drug use is more likely to
happen in Group A - or - crime rates are less
likely to increase for Group B) A B ? C
5
Research Design and the Criteria for Causality
Claims must be valid, credible and
believable Research must also address why other
viable explanation are not adequate
6
How to Design a Research Project

• Research is about producing new knowledge (what
  is the contribution to theory? Will
  you..supportnegateexpand a theory?)
• You can only produce something new if you are a
  keen observer
• Knowledge builds in steps - and so does
  research!

7
General steps in Empirical Research The 5
Magic Steps Are Back !

• 1. Problem Formulation Review, select and
  specify area to be investigated
• 2. Research Design Theoretical framework and
  Methodological framework (i.e.Quantitative vs.
  Qualitative)
• 3. Data Collection Methods Surveys, interviews,
  observations etc
• 4. Presenting of Data Analyzing and summarizing
  your data clearly
• 5. Conclusions So what? - Why is your
  research compelling? Applicability?

8
How to Design a Research Project Were going to
talk about Step 1 Problem Formulation

• Problem Formulation Review, select and specify
  area to be investigated
• This means finding a research question and
  finding a sample of people to help answer your
  research question

9
How to Design a Research Project Problem
Formulation

• The Good Question Checklist
• Is the question right for me?
• Is the question right for the field?
• Is the question well articulated?
• Is the question doable?

10
How to Design a Research Project Problem
Formulation cont

• Where do you start to formulate your question?
  Recall from last week
• Observations What have you noticed that is
  different, unusual or interesting? Are there any
  new social trends that are changing the way
  people behave?
• Conduct a Literature Review What does
  current research say about a topic that interests
  you? Are there any research gaps?

11
How to Design a Research Project Problem
Formulation cont
Chicken or the Egg Dilemma Which steps comes
first? Define your research questions clearly so
that you can THEN (1) make observations (2)
identify the appropriate body of literature Be
keen observers and review the literature so that
you can THEN form appropriate research questions
12
How to Design a Research Project Problem
Formulation cont
Remember from last week Research can be done
Inductively (i.e. specific observations leading
to theory) or Deductively (i.e. general theory
and hypothesis leading to specific
observations) Research questions can also be
similarly constructed!
13
How to Design a Research Project Problem
Formulation cont
WARNING Finding the right research question is
one of the MOST difficult steps in the research
process !!
14
How to Design a Research Project Problem
Formulation cont
BUT DONT PANIC! Stay with it and you will find
the right question You will have your Ahh Ha
moment
15
How to Design a Research Project A Brief Word on
Confounding Variables
When constructing your research questionyou MUST
keep in mind if there are any confounding (rival)
explanations for your hypothesis!!
16
How to Design a Research Project A Brief Word on
Confounding Variables
Consider the following example
17
How to Design a Research Project A Brief Word on
Confounding Variables
Summary Make your question as narrow as
possible Examine all the possible alternative
explanations of your cause and effect
hypothesis Dont throw away good research
questions just because they are controversial
18
How to Design a Research Project Target Population
Once you have formulated your problem statement
(i.e. research question) what next ?? You must
also consider who will be your subjects (e.g.
participants) Must ask yourself Who is my
target population? Teens? Drug users?
Prisoners? The population for a study is that
group of people about whom you want to be able to
draw conclusions about
19
Identifying Your Target Population The Building
Blocks of Research

• Two essential components in research design
• Units of Analysis
• Time Dimension

20
The Building Blocks of Research Units of Analysis

• Units of Analysis Includes what or who is
  being studied
• Individuals
• Groups
• Organizations
• Social Artefacts

21
The Building Blocks of Research Units of Analysis

• I. Units of Analysis Includes what or who is
  being studied
• Individuals as a unit of analysis
• Researchers make observations describing certain
  characteristics of offenders or crime victims
  (i.e. age, gender or race)
• These observations provide a descriptive picture
  of the population that comprises those individuals

22
The Building Blocks of Research Units of Analysis
2. Groups as units of analysis (Very similar
to individual units of analysis) Example If
we study members of a gang to learn about teens
who join gangs, the individual (gang member)
becomes the unit of analysis. However, if we want
to learn how big gangs differ from small ones
then the gang (group) becomes the unit of analysis
23
The Building Blocks of Research Units of Analysis
3. Organizations as units of analysis
Organizations as units of analysis differ from
groups in that organizations are more formally
structured Example If we study how one police
department may differ from another, the unit of
analysis becomes the police department
(organization)
24
The Building Blocks of Research Units of Analysis
4. Social Artefacts as units of analysis
Social artefacts are products of social beings
and their behaviour (i.e. mass media stories
about crime, TV news reports, etc) Example
Analyzing crime-related news stories (over the
last 18 months) as a way of establishing crime
trends for research
25
Problems to Consider Units Analysis and the
Ecological Fallacy
Ecological Fallacy refers to making conclusions
about individuals based solely on the
observations of groups Example A study shows
that community A has a crime rate that is three
times higher than community B. Making the
assumption of a person (randomly selected) from
community A would have experienced crime would
be to commit ecological fallacy
26
Problems to Consider Units Analysis and
Reductionism
Reductionism is an overly strict limitation on
the kinds of concepts and variables to be
considered as causes in explaining the broad
range of human behaviour Example Explaining
crime rates solely in terms of economic reasons
(while ignoring other critical social factors)
27
The Building Blocks of Research The Time
Dimension
Time also plays a role when designing research.
Time related options for research
Cross-sectional studies Longitudinal
studies Retrospective studies
28
The Time Dimension Cross Sectional Studies
Cross-sectional Studies Data are collected at
one point in time on several variables such as
gender, income and education (NOTE Data is
collected within a limited time frame) Example
Researching the differences between males and
females from different socio-economic groups in
terms of how they approach their school work in
Hong Kong
29
The Time Dimension Longitudinal Studies
Longitudinal Studies Study is designed to
permit observations over an extended period of
time Example Analysing newspaper stories
about crime over a period of five years
30
The Time Dimension Retrospective Studies
Retrospective Studies Research that requires an
analysis of a past event or issue Example A
study of crime behaviours might select a group of
prison inmates and analyze their history of
delinquency or crime
31
How to Design a Research Project Problem
Formulation Study Participants
Must find a way to access the people (population)
or situations (social artefacts) that will help
answer your research question 2 ways
Probability Sampling and Non-Probability
Sampling
32
How to Design a Research Project Problem
Formulation Study Participants
Key terms Population Usually noted as N.
Represents the total number of people which
research findings are to be generalized. Sampling
frame a list of an accessible population - or
units of analysis (i.e. organizations, people,
households) - that you will derive your sample.
Example list of registered students in HKU may
be a sampling frame for HKU student body. Sample
Usually noted as n. Represents a sample derived
from the larger population (N). Example Of the
entire student body in HKU, randomly select 100
students (50 male and 50 female).
33
How to Design a Research Project Problem
Formulation Study Participants

• Key terms
• Generalizability the extent by which your
  research results (derived from your sample
  population) can be applied (generalized) to the
  larger population (N).
• How well your results can be generalized depends
  on
• Representativeness of sample population
• Size of sample population
• Randomness of sample population

34
How to Design a Research Project Sampling
Probability vs. Non-Probability

• Probability Sampling
• You cant study everyone in a population (i.e.
  all teens in Hong Kong). Must break populations
  down into manageable units. Probability sampling
  uses some form of RANDOM selection
• 4 kinds of random (Probability) sampling
• Simple Random Sampling
• Stratified Random Sampling
• Systematic Sampling
• Cluster Sampling

35
How to Design a Research Project Probability
Sampling
Simple Random Sampling Everyone from the larger
population (N) has an equal chance of
participating in the smaller sample (n) (think of
it as winning the lottery) Most widely used
RANDOM sampling process
36
How to Design a Research Project Probability
Sampling
Stratified Random Sampling (sometimes referred to
as quota random sampling or proportional
sampling) Population is first divided into two
or more mutually exclusive segments based on some
categories of interest (gender, race, income,
etc) These sub-populations are non-overlapping
(i.e. if you add them all up, they will equal the
larger population) Once each sub-population has
been identified, a sample is then randomly drawn
using equal number of people from each group
37
How to Design a Research Project Probability
Sampling
Systematic Random Sampling Sample is taken
from larger population systematically yet
randomly N 100 n 10 Then divide N/n
10 Select a random number between 1-10 (say
3) Start with number 3 and take every 10th
unit
38
How to Design a Research Project Probability
Sampling

• Systematic Random Sampling
• Example
• Unit Variables
• High income, Female, 34
• Low income, Female, 45
• High income, Male, 29
• Low income, Male, 34
• Low income, Male, 45
• High income, Male, 23
• Mid income, Female, 39
• Mid income, Male, 31
• Mid income, Male, 32
• Low income, Female, 44
• Mid income, Male, 46
• Low income, Male, 38
• Mid Income, Female 41

39
How to Design a Research Project Probability
Sampling
Cluster Sampling Unlike simple random
sampling and stratified sampling where subjects
are selected from the larger population (N), in
cluster sampling subjects are selected in groups
or clusters. This approach allows researchers to
overcome the constraints of costs and time
associated with big and dispersed populations
40
How to Design a Research Project Probability
Sampling
Cluster Sampling Example A researcher wants
to know how to improve drug awareness in primary
schools. Ten schools are randomly selected (i.e.
clusters) and interviewed
41
How to Design a Research Project Probability
Sampling
IMPORTANT NOTICE There is a difference between
Random sampling and Random assignment
42
How to Design a Research Project Non-Probability
Sampling

• What happens when you can not do random
  sampling? Say you want to research homelessness.
  How do you obtain a population of all the
  homeless people living in Hong Kong?
• 4 kinds of non-probability sampling
• Availability sampling
• Quota sampling
• Purposive sampling
• Snowball sampling

43
How to Design a Research Project Non-Probability
Sampling
Availability sampling is a method of choosing
subjects who are available or easy to find
(referred to as accidental, haphazard or
convenience sampling) Example Asking your
office mates to participate in your study
44
How to Design a Research Project Non-Probability
Sampling
Quota sampling is designed to overcome the flaws
of availability sampling (NOTE Do not confuse
with quota random sampling) Rather than just
taking anyone, you set quotas to ensure that the
sample you get represents certain characteristics
in proportion to their prevalence in the larger
population Example When asking your office
mates to participate in your study, you include
48 males and 52 females to reflect the gender
distribution of the larger population
45
How to Design a Research Project Non-Probability
Sampling
Purposive sampling is a method which elements are
chosen based on the overall goal of the study. As
a rule, it does not represent a sample that is
representative of a larger population (like you
find in non-probability sampling) Example You
interview gang members who frequent a particular
club
46
How to Design a Research Project Non-Probability
Sampling
Snowball sampling is a method in which
researchers identify one member of some
population of interest and asks him/her to
introduce them to others that might want to
participate in a study Example You interview a
gang member and ask them to introduce you to
other gang members they might know
47
How to Design a Research Project Things to
consider
You must also consider practical issues like
(1) Scheduling concerns (i.e. if you are
interested in studying university student
behaviours, then summer months might not be a
good time to plan interviews) (2) Costs of
running your study (i.e. time, money, staff etc)
48
How to Design a Research Project Things to
consider
When you identify the best sampling procedure
for yourself, you must also figure out how you
are going to analyze your results Quantitative? Q
ualitative? Variables of interest?
49
How to Design a Research Project Can now move on
to STEP 2 of the 5 MAGIC STEPS

• Problem Formulation Review, select and specify
  area to be investigated
• Research Design Quantitative vs. Qualitative
• Data Collection Methods Surveys, interviews,
  observations etc
• Presenting of Data Analyzing and summarizing
  your data clearly (Is it reader-friendly?)
• Conclusions So what? - Why is your research

50
Tune in Next Lecture
Questions ?