Intro to Social Research Ch 4-6
Major Steps in Research Design
1. Getting Started 2. Conceptualization 3. Choose a research method 4. Operationalization 5. Sampling Techniques 6. Observe, process data, analyze 7. Apply
Sufficient Cause
A condition that if present, guarantees the effect in question
Criterion-relate validity
Degree to which a measure relates to some external criterion, also known as predictive reliability
Bogardus Social Distance Scale
Determines people's willingness to participate in social relations (of varying degrees) with others, this equals social distance
Dimensions
Distinctive, specifiable aspects of a concept ex.) Romantic love -> think, feel, behave
Causality: Correlation
Empirical Correlation, not the same as causation
False Causality Criteria: Exceptional Cases
Exceptions don't falsify causality
Precision and accuracy are synonyms
F
Research designs are descriptive, explanatory, or exploratory-never more than one.
F
Test-retest method
Making the same measurement more than once
A research proposal describes what you intend to accomplish and how.
T
The ecological fallacy refers to drawing conclusions about individuals based solely on the observation of groups.
T
The nomothetic model of explanation is probabilistic in its approach to causation.
T
The purpose of multivariate analysis in index construction is to discover the simultaneous interaction of the items to determine whether they are all appropriate for inclusion in the same index.
T
The result of combining several indicators of a variable is a composite measurement of the variable.
T
Necessary Cause
A condition that must be present for the effect to follow
Multivariate Relationship
A relationship between more than two variables
Bivariate Relationship
A relationship between two variables
Index
Accumulates scores assigned to individual indictors
Item Analysis
An assessment of whether each of the items included in a composite measure makes an independent contribution or merely duplicates the contribution of other items in the measure
Indicators
Any observation that we choose to consider as a reflection of a concept we wish to study (actual item) ex.) Romantic love -> How often does a couple show affection, have sex, give each other flowers, etc?
Units of Analysis: Social Artifacts
Any product of social beings or their behavior, ex.) books, excuses, scientific discoveries
Nominal Definition
Assigned to a term with consensus, it's what everyone agrees on
Scale
Assigns scores to patterns of responses
Measurement theory (the true score theory)
Based on the idea that empirical measure of a concept reflects three components, 1. True construct, or an ideal measure of it 2. Systematic errors or biases 3. Random error This is expressed as X = T+S+R where X is the empirical indicator/observation
Tensions between reliability and validity
Basically, there's a trade-off, it's really hard to get the same amount of reliability and validity in a study
Research-worker reliability
Checking the reliability of research workers (interviewers and coders), there are various methods to do this, clarity, specificity, training and practice prevent a great deal of unreliability
Typologies
Classifications in terms of something's attributes on 2 or more variables, uses mutually exclusive categories
Thurstone Scaling
Composite measure, develops questions, gathers judges, gets them to rank order everything, very time consuming and expensive
Units of Analysis: Groups
Considered a single entity, ex.) gangs, families
Internal consistency
Cronbach's alpha, mathematically equivalent to the average of all possible split-half estimates
Operational Definition
Defines how a concept is measured
Content validity
Degree to which a measure covers the range of meanings included within a concept
Construct validity
Degree to which a measure relates to other variables as expected within a system of theoretical relationships
A necessary condition represents a condition that, if it is present, will pretty much guarantee the effect in question.
F
A nominal measure can have only two categories
F
A set of questionnaire items that Guttman scaled on one data set will usually Guttman scale on another data set.
F
A split-half reliability test taps the idea of the general stability of the instrument over time.
F
Always include at least one item in an index on which all respondents give the same answer.
F
An index is constructed by assigning scores to patterns of attributes.
F
Dr. Jordan chronicles people's attitudes toward an upcoming property tax levy. Jordan is doing explanatory research.
F
If a measure is reliable, it must also be valid.
F
If three people (Smith, Jones, and Edwards) are rank-ordered with respect to the possession of an attribute on an ordinal variable, then we can conclude that the distance between Smith and Jones on that variable is equal to the distance between Jones and Edwards.
F
If two items in an index are perfectly correlated, both of them should be eliminated from the index.
F
If two variables are correlated with each other, there must be a causal relationship between them.
F
It is impossible to have several indicators of only one concept.
F
Item analysis allows you to examine the extent to which the composite index is related to other items in the questionnaire that are not part of the index.
F
Longitudinal studies may be either trend or cohort studies, but not panel studies.
F
Numbers assigned to ranks on an ordinal scale can legitimately be added, multiplied, subtracted, and divided.
F
Operationalization is the process of clarifying what is meant by the concepts being used in a study.
F
Panel attrition is comparable to experimental mortality.
F
Precise measurement is more important than accurate measurement.
F
Researchers ignore previous work on a topic so that their research can be original.
F
The independent (effect) variable must occur later in time than the dependent (cause) variable.
F
Operationalization
Range of variation, the specific way we measure something, includes degrees of precision and dimensions
Semantic Differential
Rating something in terms of two opposite adjectives, uses very, somewhat, neither, etc; hard to use on dumb people
Causality: Non-spurious Relationship
Relationship between variables must be real, there can't be a third variable connecting them
Exploratory Research
Research as means of exploring
Descriptive Research
Research that describes, answers who what when and where
Explanatory Research
Research that explains, answers why
False Causality Criteria: Majority of Cases
Social science is supposed to explain variation not averages, if there's no variation there's no point
Split-half method
Splitting the data in half, the two sets should classify the data the same way.
A major shortcoming of exploratory studies is that they seldom provide satisfactory answers to research questions.
T
A particular variable can usually be measured in several distinct ways using different sources of information and various observation techniques.
T
Being at least 18 years of age is a necessary cause for voting in the United States.
T
Changing definitions almost inevitably results in different descriptive conclusions.
T
Conceptualization and operationalization are processes that continue throughout the research process.
T
Descriptive research answers the question "What's so?" and explanatory research answers the question "Why?"
T
Explanatory research depends on the notion of cause and effect.
T
For an index or scale to be considered unidimensional, its component items should be indicators of only one dimension.
T
Predictive validity is another term for criterion-related validity
T
Thurstone scales use judges in their construction.
T
Typologies are nominal composite measures.
T
Validity refers to the link between the operational and conceptual definitions.
T
When one's research purposes aren't clear, it is advisable to choose the highest level of measurement possible.
T
External Validation
Testing the validity of a measure by examining its relationship to other presumed indicators of the same variable
Levels of Measurement: Interval (continuous)
The amount of distance between two variables can be specified and equal ex.) test scores
Range of Variation
The extent one is willing to combine attributes in fairly loose categories
Trend Study
Type of longitudinal study, a given characteristic is monitored over a period of time
Cohort Study
Type of longitudinal study, a specific subpopulation is studied over time, although data may be collected from different members in each set of observations, also known as a generational study
Panel Study
Type of longitudinal study, data is collected from the same set of people over time
Likert Scaling
Used to determine relative intensity of different items, typically measures attitudes (ordinal) SD<->D<->N<->A<->SA
Guttman Scaling
Used to summarize several discrete observations to represent a more general variable
Using established measures
Using measures that have proved their reliability in previous research
Levels of Measurement: Ordinal
Variables whose attributes can be logically rank-ordered ex.) religiosity, how religious are you? strongly agree, agree, neutral, disagree, strongly disagree
Units of Analysis: Social Interactions
What goes on between humans, ex.) kissing, dancing, arguments
False Causality Criteria: Complete Causation
You can never have 100% certainty, causation is probabilistic
Units of Analysis: Individuals
Individual human beings, most common
Units of Analysis: Organizations
Individual social organizations, ex.) corporations, colleges
Causality: Theoretically Meaningful
It has to make sense
Causality: Time Order
Kind of intuitive, cause always precedes effect
Levels of Measurement: Ratio (continuous)
MEANINGFUL ZERO UNIT, variables are based on a true zero value ex.) income, because zero income is meaningful
Ecological Fallacy
Making an observation about a large group and and trying to apply it to a small number of people, or aggregates
Conceptualization
Making something fuzzy into a solid, concrete definition
Levels of Measurement: Nominal
Measurement that has mutually exclusive, exhaustive categories ex.) gender
Longitudinal Research
Observations made over an extended period of time
Cross Sectional Research
Observations of a sample made at one point in time
Reductionism
Observing smaller units of analysis and applying observations to large groups of those individuals
Face validity
Quality of an indication that makes it seem a reasonable measure of some variable
