Answer the following question. Your answer should be at least three paragraphs in length (350-650 words)and should be presented in your own words. Any use of quoted material must be properly cited.Surveys are complicated to design, costly to administer and potentially suffer from response bias with respect to who answers the. Why use this data collection method? Draw on the case of the General Social Survey to support your answer.Connect the answer to the textbook. (Chapter 2)Reference Textbook: You May Ask Yourself An Introduction to Thinking like a Sociologist 5th Edition
sociology_week_2_12302018.ppt

sociology_week_2_12302018.ppt

Unformatted Attachment Preview

SOC 1100 Sociology
Week 2 Lecture
2
Sociological Investigation
• 2 Key Requirements
– Apply Sociological Perpective
– Be Curious & Ask Questions
• Sociology is a Science
– Based in Logic; Direct Systematic Observation
• Science is a Form of Truth
– Empirical Evidence; Verifiable
3
“Doing” Sociology
• 3 Ways
– Postivist Sociology
– Interpretive Sociology
– Critical Sociology
4
Operationalization
• Operationalization takes your beyond
conceptualization (defining) to developing
appropriate measures.
• If your (concepts and) variable are clearly
defined (conceptualized) and your testing
measured clearly defined (operationalization) you
will reduce/eliminate problems as the research
progresses – when analyzing data and reporting
results
5
Reliability & Validity
• Reliability and validity are key to a good research
study.
• Without both, the study results are in question.
• Reliability looks to whether or not you’re measuring
what you think you’re measuring what you thing
you’re measuring.
• Validity looks to the accuracy of the responses.
• Just because you have reliable questions or
measurement techniques does not guarantee valid
responses.
6
How Can We Improve Reliability?
– Test – Retest
• Test more than once
– Inter-Item reliability
• Multiple questions involving the same variables
– Alternate question forms
• Ask the same question in different ways to see if the
answer remains consistent
– Inter-Rater reliability
• Train all data collector’s the same way
• Have spot checks of data across collector’s as a
means to verify that questions are being coded
consistently
7
Variables
• Variables are (logical) groupings of attributes
• Independent Variable
– The independent variable is the ‘cause’ in a causal
relationship, it brings about change in one or more variables.
– The independent variable in one relationship may very well be
the dependent variable in another.
• Dependent Variable
– The dependent variable is the ‘effect’ in a causal relationship,
it is impacted by the causal variable(s).
– The dependent variable in one relationship may very well be
the independent variable in another.
• The relationship between the variables has to be testable and
logical
8
Validity & Causal Inference
• Validity
– The best available approximation to the truth or falsity
of propositions, including propositions about cause.
• (Cook, T.D. & Campbell, D.T. (1979). Quasiexperimentation: design & analysis issues for field
settings. Houghton Mifflin Company: Boston.)
– There are several types of validity to consider when
conduting research
• Statistical Conclusion Validity
• Internal Validity
• Construct Validity
• External Validity
9
Validity Explained
• Statistical Conclusion Validity
– The change in the dependent variable is statistically associated
with the effect of the independent variable
• Internal Validity
– Challenge causal statements through observation
• Construct Validity
– How well does the relationship represent the causal process
– Generalize what we observe and measure to real world things
• External Validity
– Is the relationship generalizable across similar populations
elsewhere
• Sample generalizability
• Cross population generalizability
10
Threats to Validity
• Statistical Conclusion Validity
– Problem with too small a sample size
– Poor measures of cause and/or effect
– Can result in a false positive
• Internal Validity
– Nonrandom or systematic error (you did something wrong)
– Effect is the result of a 3rd unidentified variable
• Construct Validity
– Improper variable definition impacts generalizability
• External Validity
– Inaccurate sample selection impacts generalizability
11
Cause & Effect Relationships
• Independent Variable
– Cause
– Brings About Change in Dependent Variable
• Dependent Variable
– Effect
– Changes in Response to Absence or Presence of
Independent Variable
12
Correlation
• Correlation Implies a Relationship Between
Variables
• Not A Causal Relationship
• 2 or More Variables Change at the Same Time
• Spurious Correlation
– Change is due to another variable
13
Causal Criteria
• Certain criteria must be met in order for an independent
variable to be accurately identified as causal
– The independent variable must precede the dependent
variable in time
• Think of the chicken/egg paradox – which came first?
The chicken or the egg… you can’t have an egg
without a chicken, but you can’t have a chicken
without the egg – identifying either as the causal
variable can be refuted.
– The observed correlation between the independent
variable and dependent variable cannot be explained
away because the variables are influenced by an third
unnamed/unidentified variable that is the true cause of
both (see spurious correlation)
14
2 Types of Cause
• There are two separate and distinct types of ‘cause’
– Necessary Cause
• The independent variable must exist for the dependent
variable to occur
– Ex: you need both sperm and an egg to produce a child
– Sufficient Cause
• The existence of the independent variable pretty much
guarantees the occurrence of the dependent variable
– Ex: cloud moisture and temperatures below 32 degrees
farenheit may in fact cause snow fall
• You are not likely to find both a necessary and sufficient
cause at the same time
15
Ethics
• Ethical Principles
– Achieving valid results
• Valid results necessary starting points
– Honesty and openness
• Concerns about validity require research to
be open and honest about their research
– Protect research participants
• ASA Code of Ethics 1997
16
Historical Background
– Milgram’s Obedience Studies
• Started in 1960 with Teacher/Student studies at Yale
– Resistance to or acceptance of authority
• Raised many ethical concerns about how subjects are treated during
research trials
– Development of Formal Ethics Procedures
• Result of widely publicized human rights abuses
– 1946 Nuremburg War Crime Trials
– 1930’s – 1970’2 Tuskegee Syphilis Study
• National Commission of Human Subjects of Biomedical & Behavioral
Research
– Respect for Persons
– Beneficence
– Justice
– Institutional Review Boards
• Members have different backgrounds
• Office of Protection from Research Risks in the National Institute of Health
Monitors IRBs
– American Sociological Association
• Adopted ethics guidelines
– More specific than federal regulations
17
ASA Code of Ethics





Cause no harm to subjects
Participation should be voluntary
Researchers must disclose identity
Maintain anonymity/confidentiality
Benefits outweigh risks
18
Research Methodology
• 4 Common Methodologies
– Experiment
– Survey
– Participant Observation
– Secondary Data Analysis
19
10 Steps in Sociological Investigation

What is Your Topic?

What Have Others Already Learned?

What, Exactly, Are Your Questions?

What Will You Need to Carry Out Research?

Are There Ethical Concerns?

What Method Will You Use?

How Will You Record Your Data?

What Do the Data Tell You?

What Are Your Conclusions?

How Can You Share What You’ve Learned?
20
Gender & Research
• Research is Affected by Gender in 5 Ways
– Androcentricity – approaching only from a male
perspective
– Overgeneralizing – studying only one gender to prove
a point
– Gender Blindness – ignoring gender as a variable
– Double Standards
– Interference because respondent reacts to gender of
researcher
21
Qualitative Research
• Observing
– Choosing a Role
– What are the possible roles?
• Complete Participation
• Mixed Participation/Observation
• Complete Observation
22
Compete Observation
• Complete Observation
– In complete observation, researchers try to see things as they
happen, without actively participating in these events.
• To study a political group
• To study a workplace
– Hospital
– School
– Potential Problem
• Reactive Effects
– People change the way they behave because they know
they are being observed and/or recorded
23
Mixed Observation/Participation

Mixed Participation/Observation
– In most settings at least one member of the group being observed is aware of the
presence of the researcher
– The researcher participates with the group in such a way as to develop a rapport
or some level of trust with the group members
• This is not easily accomplished or maintained
• This method has pretty clear ethical advantages
– The group is aware of the fact that they are being observed, therefore
they can choose what information/activities in which to include the
researcher
– The researcher is not obligated to participate in illegal, immoral, or
dangerous activities
24
Complete Participation
• Because the researcher has not identified him/herself the
choice to participate in illegal, immoral or dangerous activities
is greatly limited
• The researcher takes a typically covert role, participating in
all activities of the group
• Members of the group are not aware that they are being
observed
• Example:
– Laud Humphreys (1970) served as a “watch queen” so that he
could learn about men engaging in homosexual acts in a public
restroom
– Randall Alfred (1976) joined a group of Satanists to
investigate group members and their interaction
25
Data Sources

Three Sources of Data
– Asking Questions
• Survey data
• Focus groups
• Individual interviews
– Leaves possibility of additional questions
– Making Observations
• Direct observation
– Ex: lighting levels, security measures, counting, drug tests
• No interaction with the subjects involved
• More subjective than self reports or survey data
– Examining Written Reports
• School reports
• Medical records
• Police reports
• This is all second hand data
– Be aware of how the data were originally collected and analyzed.
26
Sampling

Probability sampling
– Impacted by population heterogeneity
• To be truly representative, you must closely match the population break down
– Advantages
• More representative
• Less bias
• Better estimate of accuracy of representativeness

Representativeness
– Sample has basically the same characteristics as the total population
• Representativemess only needs to apply to the relevant variables

EPSEM Samples
– Equal probability of selection
• Everyone has an equal chance of being selected as a member of the study sample

Rare to have a perfectly representative sample
27

Threats to Internal Validity with Experimental Designs
– History
• Exposure to outside effects may impact subjects
– Maturity
• Growth and change of the subjects
– Testing
• Alters behavior of subjects
– Instrumentation
• The result of changes made to the research instrument after onset
of study
– Statistical regression
• Widely skewed variables tend to regress naturally toward the
mean
– Selection biases
• Can be reduced through randomization
28
• Threats to Internal Validity with Experimental Designs (cont)
– Experimental mortality
• Subjects drop out the study
– Causal time order
• Ambiguity about time order of intervention
– Diffusion or imitation of treatments
• Control and experimental groups share information
– Compensatory treatment
• Control group feels deprived
– Compensatory rivalry
• Control group works harder to make up for the missing
intervention
– Demoralization
• Control group gives up as a result of the missing
intervention
29
• Survey Research
– Standard questions asked of a sample (or samples)
drawn from a larger population
– Very old
• Examples of old surveys
– Biblical times (earliest being recorded in the book
of numbers)
– Marx 1880
» Surveyed workers
30
Survey Types
• Survey Types
– Self-administered questionnaires
– In-person interviews
– Telephone surveys
– Specialized interviewing
– Focus Groups
31
• Survey Benefits
– In Person Interviews Provide
• More completed interviews
• Higher response rates
• Better handling of complex questionnaires
– Mail Surveys Provide
• More privacy
• Better for use with sensitive issues
– Consider
• Research needs
• Resources available to you
32
• Research Purpose
– Exploration
• Develops understanding
– Description
• Counts or documents
– Explanation
• Just as the name describes
– Application
• Applies research findings to public policy
• Research Projects are not limited to one purpose
33
Traditional Model of Science
• The traditional model of science is made up of
three elements
– Theory
• Defined above
– Operationalization
• Develop questions to be asked and method of
study
– Observation
• Data collection
34
Wheel of Science
Theories


Empirical Generalization

Hypotheses

Observations
35
Deductive Model
• When using a deductive model:
– Start at the top of the wheel with the development of a
theory or theories
– Then develop you hypothesis or hypotheses
– Make observations
– Draw empirical generalizations based upon the
observations
– If empirical generalizations do not fit the original
theory, start over
36
Inductive Model
• When using an inductive model:
– Start at the bottom of the wheel with observations or
by using previously collected data
– Make empirical generalizations based upon the
observations or data
– Develop a theory or theories based upon the
generalizations
– Develop your hypothesis or hypotheses
– Make observations to test the new theory and to
replicate the original data/observations used to draw
empirical generalizations to begin with
37
Not a Cop Out
• Despite the need for causes, the deterministic model and social
scientists:
– Don’t believe that everything is pre-determined
– Recognize that patterns are usually not simple, but rather
complex
– Recognize that not everyone is controlled by the same factors
• What works for one may not work for another
• Social Science follows (generally) the probabalistic – causal model
– Variable A may and probably does cause Variable B, but we
cannot say for sure
• However, we have to consider the deterministic model in social
science research
38
SOC 1100 Sociology
Week 2 Lecture
2
Sociological Investigation
• 2 Key Requirements
– Apply Sociological Perpective
– Be Curious & Ask Questions
• Sociology is a Science
– Based in Logic; Direct Systematic Observation
• Science is a Form of Truth
– Empirical Evidence; Verifiable
3
“Doing” Sociology
• 3 Ways
– Postivist Sociology
– Interpretive Sociology
– Critical Sociology
4
Operationalization
• Operationalization takes your beyond
conceptualization (defining) to developing
appropriate measures.
• If your (concepts and) variable are clearly
defined (conceptualized) and your testing
measured clearly defined (operationalization) you
will reduce/eliminate problems as the research
progresses – when analyzing data and reporting
results
5
Reliability & Validity
• Reliability and validity are key to a good research
study.
• Without both, the study results are in question.
• Reliability looks to whether or not you’re measuring
what you think you’re measuring what you thing
you’re measuring.
• Validity looks to the accuracy of the responses.
• Just because you have reliable questions or
measurement techniques does not guarantee valid
responses.
6
How Can We Improve Reliability?
– Test – Retest
• Test more than once
– Inter-Item reliability
• Multiple questions involving the same variables
– Alternate question forms
• Ask the same question in different ways to see if the
answer remains consistent
– Inter-Rater reliability
• Train all data collector’s the same way
• Have spot checks of data across collector’s as a
means to verify that questions are being coded
consistently
7
Variables
• Variables are (logical) groupings of attributes
• Independent Variable
– The independent variable is the ‘cause’ in a causal
relationship, it brings about change in one or more variables.
– The independent variable in one relationship may very well be
the dependent variable in another.
• Dependent Variable
– The dependent variable is the ‘effect’ in a causal relationship,
it is impacted by the causal variable(s).
– The dependent variable in one relationship may very well be
the independent variable in another.
• The relationship between the variables has to be testable and
logical
8
Validity & Causal Inference
• Validity
– The best available approximation to the truth or falsity
of propositions, including propositions about cause.
• (Cook, T.D. & Campbell, D.T. (1979). Quasiexperimentation: design & analysis issues for field
settings. Houghton Mifflin Company: Boston.)
– There are several types of validity to consider when
conduting research
• Statistical Conclusion Validity
• Internal Validity
• Construct Validity
• External Validity
9
Validity Explained
• Statistical Conclusion Validity
– The change in the dependent variable is statistically associated
with the effect of the independent variable
• Internal Validity
– Challenge causal statements through observation
• Construct Validity
– How well does the relationship represent the causal process
– Generalize what we observe and measure to real world things
• External Validity
– Is the relationship generalizable across similar populations
elsewhere
• Sample generalizability
• Cross population generalizability
10
Threats to Validity
• Statistical Conclusion Validity
– Problem with too small a sample size
– Poor measures of cause and/or effect
– Can result in a false positive
• Internal Validity
– Nonrandom or systematic error (you did something wrong)
– Effect is the result of a 3rd unidentified variable
• Construct Validity
– Improper variable definition impacts generalizability
• External Validity
– Inaccurate sample selection impacts generalizability
11
Cause & Effect Relationships
• Independent Variable
– Cause
– Brings About Change in Dependent Variable
• Dependent Variable
– Effect
– Changes in Response to Absence or Presence of
Independent Variable
12
Correlation
• Correlation Implies a Relationship Between
Variables
• Not A Causal Relationship
• 2 or More Variables Change at the Same Time
• Spurious Correlation
– Change is due to another variable
13
Causal Criteria
• Certain criteria must be met in order for an independent
variable to be accurately identified as causal
– The independent variable must precede the dependent
variable in time
• Think of the chicken/egg paradox – which came first?
The chicken or the egg… you can’t have an egg
without a chicken, but you can’t have a chicken
without the egg – identifying either as the causal
variable can be refuted.
– The observed correlation between the independent
variable and dependent variable cannot be explained
away because the variables are influenced by an third
unnamed/unidentified variable that is the true cause of
both (see spurious correlation)
14
2 Types of Cause
• There are two separate and distinct types of ‘cause’
– Necessary Cause
• The independent variable must exist for the dependent
variable to occur
– Ex: you need both sperm and an egg to produce a child
– Sufficient Cause
• The existence of the independent variable pretty much
guarantees the occurrence of the dependent variable
– Ex: cloud moisture and temperatures below 32 degrees
farenheit may in fact cause snow fall
• You are not likely to find both a necessary and sufficient
cause at the same time
1 …
Purchase answer to see full
attachment