URBP 204A QUANTITATIVE METHODS I Survey Research II

1
URBP 204A QUANTITATIVE METHODS ISurvey Research
II

• Gregory Newmark
• San Jose State University
• (This lecture is based on Chapters 4 6 of Earl
  Babbies
• The Practice of Social Research, 10th Edition.
• All uncredited cartoons are from CAUSEweb.org by
  J.B. Landers.)

2
Research Purposes

• Exploration
• Description
• Explanation

3
Research Purposes

• Exploration to get to know a new topic
• Examples
• What is transit oriented development?
• What is equity planning?
• Purposes
• To satisfy researchers curiosity
• To test the feasibility of more extensive study
• To develop methods to be employed in future study
• Benefits and Drawbacks
• Breaks new ground and yields new insights
• Seldom provides satisfactory answers

4
Research Purposes

• Description to describe observations
• Examples
• What are the characteristics of TOD in
  California?
• How does this equity planning program work?
• Purposes
• To describe situations and events
• To chronicle activities
• Benefits and Drawbacks
• Answers questions of what, where, when, and how
• Does not answer why

5
Research Purposes

• Explanation to explain phenomena
• Examples
• Why are TODs not more successful in California?
• Why does this equity planning program work so
  well?
• Purposes
• To explain things
• To understand why something occurred
• Benefits and Drawbacks
• Causality

6
Explanation

• How do you know if your explanation is credible?

7
Nomothetic Causality

• Real Criteria
• Correlation
• A relationship exists between variables
• Housing values are negatively related to commute
  length
• Time Order
• Cause precedes effect
• Gender impacts political opinions
• Non-Spurious (Non Coincidental)
• Effect can not be explained by a third variable
• Height is not explained by Zodiac sign

8
Nomothetic Causality

• Why is correlation not enough for causality here?

9
Nomothetic Causality

• Does this cartoon pass the causality criteria?

10
Nomothetic Causality

• False Criteria
• Complete Causation
• Nomothetic explanations are usually incomplete
• Three common factors for driving a Prius are . .
  .
• Exceptional Cases
• Nomothetic explanations are not voided by
  exceptions
• I am an environmentalist, but I dont drive a
  Prius.
• Majority of Cases
• Nomothetic explanations still valid even if
  effect only occurs in the minority of cases
• Most environmentalists dont drive a Prius.

11
Nomothetic Causality

• Necessary Causes
• Represents a condition that must be present for
  the effect to follow
• You need to be female to get pregnant
• You need to go to college to get a college
  degree
• Sufficient Causes
• Represents a condition that guarantees effect
  will follow
• Skipping the final exam will result in automatic
  failure
• Participating in our survey will get you a
  t-shirt

12
Nomothetic Causality

• Unit of Analysis
• What or whom are being studied
• Individuals, groups, organizations, social
  artifacts, etc.
• Typically, but not always, the unit of
  observation
• For example, a study of couples might observe
  both partners separately
• Data from individual units can be aggregated
• 32 percent of our sample is college educated
• This neighborhood predominantly does not vote

13
Nomothetic Causality

• Faulty Reasoning regarding Units of Analysis
• Ecological Fallacy
• Drawing conclusions about individual based on
  observation of groups
• Younger neighborhoods support recycling,
  therefore younger people are more likely to
  support recycling
• Reductionism
• Explaining complex phenomena with narrow concepts
• Oversimplification
• What is the single cause of the American
  Revolution?
• Economists see everything in economic terms.

14
Nomothetic Causality

• Ecological Fallacy

15
Nomothetic Causality

• Reductionism

16
Time Dimension of Research

• Cross-Sectional Studies
• Observations are all made at one point in time
• E.g. our FWBT survey
• Longitudinal Studies
• Observations are extended over time
• Trend Studies track a characteristic
• Cohort Studies track a subpopulation
• Panel Studies track the same people (the panel)
• Problem of panel attrition (dropping out of the
  study)

17
The Research Proposal

• Problem/ objective of research
• Relevance of research
• Literature review
• Research question
• Hypothesis
• Variables of interest
• Research method/s
• Types of data to be collected - methods, sources
• Plan for data analysis
• Outline of research report - main chapters,
  sub-chapters
• References/ Bibliography
• Schedule
• Budget

18
Composite Measures

• Some concepts are not easily represented by a
  single variable
• Particularly attitudes and orientations
• E.g. religiosity, alienation, prejudice, etc
• Composite measures can help
• Can combine information from several variables
• Can expand the range of variation
• Can be useful for data analysis

19
Indexes vs. Scales

• Both are ordinal measures of variables
• Both are composite measures of variables
• Indexes are the simple accumulation of scores
• We award a point for each task completed and
  tally those scores to create an index value
• Dow Jones Industrial Index
• Scales assign scores to patterns of responses
• Consider differences in intensity of the response
• Have some sort of logical or empirical order
• We assess the health risk to humans of a mix of
  pollutants and then award scores on a scale from
  least to most harmful.

20
Indexes vs. Scales

• Determine an index measure which would be
  contradicted by a scale measure.
• For example, if you were to index conservative
  politicians by votes on conservative bills, you
  might incorrectly consider a politician who voted
  for several moderately conservative bills as more
  conservative than a politician who did not vote
  for those bills because they were not
  conservative enough.

21
Index Construction

• Item selection
• What should we include in our index?
• Face validity, Unidimensionality, Specificity,
  Variance
• Examination of empirical relationships
• Are variables related to each other? Generally
  useful.
• No relation probably should not include one
• Too close a relation probably should not
  include one
• Index Validation
• Internal Validation (Item analysis)
• Each item should have independent contribution
• Should not be perfectly correlated with other
  another item
• External Validation
• Index should correlate with other presumed
  indicators of variable
• People who score strongly on the index should
  score strongly on other related measures.

22
Scale Construction

• Bogardus Social Distance Scale
• Determines the willingness of people to
  participate in social relations of varying
  degrees of closeness with other kinds of people
• (Least extreme)
• 1. Are you willing to permit immigrants to live
  in your country?
• 2. Are you willing to permit immigrants to live
  in your community?
• 3. Are you willing to permit immigrants to live
  in your neighborhood?
• 4. Are you willing to permit immigrants to live
  next door to you?
• 5. Would you permit your child to marry an
  immigrant?
• (Most extreme)
• E.g., agreement with item 3 implies agreement
  with items 1 and 2.
• Guttman Scale
• More general case than that above
• Binary (Yes/No) responses to a set of questions
  can be ranked, so that some one answering yes to
  a more difficult question will also answer yes to
  an easier one.
• A coefficient of reproducibility is used as few
  Guttman Scales are perfect

23
Scale Construction

• Thurstone Scale
• Measures intensity structure among indicators by
  use of judges to assign scores
• Likert Scale
• Uses standardized response categories in survey
  questionnaires to determine the relative
  intensity of different items
• Semantic Differential
• Two opposing possibilities are given and the
  respondent selects their gradation between those
  poles

24
Scale Construction

• Likert Scale

25
Scale Construction

• Semantic Differential Scale

26
Typology

• Classification of observations with respect to
  two or more attributes.

Typology of Housing High Quality Low Quality
Pre-War Construction
Post War Construction