IV' SURVEY RESEARCH STRATEGIES

1
IV. SURVEY RESEARCH STRATEGIES
2
A. Introduction

• 1. most social science knowledge is based
  on the
• process of asking questions and
  processing answers
• 2. the trick is asking the right
  questions and getting the
• right answers
• 3. survey research is the basic
  investigative tool used to
• conduct social science investigation
• 4. consideration must be given to the type
  of data that are
• best for the research conducted
• gt quantitative or qualitative

3
B. Types of Research Models

• 1. exploratory studies
• a. useful in gaining preliminary
  information and insights
• b. also called pilot studies
• c. initial general description/s of a
  particular
• phenomenon or research question
• gt most research contains at least
  some description of
• an exploratory nature
• d. fact-finding explorations
• e. usually univariate in nature
• f. can be or lead to one-shot or
  longitudinal in
• applications

4

• 2. explanatory research
• a. generally take the form of
  correlation research
• b. an examination of the relationship
  between
• variables leading to uni / bi /
  multivariate analysis
• c. process of identifying specific
  constructs /
• variables that will be measured and
  compared to
• another construct or variable
• d. this relationship is then examined
  as a strong or
• weak relationship
• gt i.e., what is the probability
  that A B ?

5

• 3. one-shot / fixed-point / snap-shot
  studies
• a. data are collected at one point in
  time only at a
• particular place and produces a
  particular array of
• data
• gt creates methodological problems
  with sampling,
• i.e., questions the validity of
  the information
• collected
• b. problem arises in trying to
  generalize these results
• across a wider population or in
  other situations
• c. generally this research is either
  descriptive /
• exploratory or explanatory

6

• 4. case studies
• a. an extensive theoretical,
  behavioral, operational,
• comprehensive examination of an
  individual,
• organization, incident, legal case,
  etc.
• b. an example of a particular situation
  to demonstrate
• a positive or negative procedure
• c. province of law schools, business
  schools, and most
• police training academies
• d. Gideon v Wainwright (372 US 335.
  1963)
• e. parental context Why cant you be
  more like ...?

7

• 5. true experiments
• a. the classic experiment
• b. a test / treatment is administered /
  given to a coterie
• of research subjects in a
  controlled environment
• designed to examine specific
  conditions or test
• conditions
• c. all operational definitions are
  established prior to the
• onset of research and their results
  are assessed
• d. design utilizes two unique reporting
  groups (samples)
• experimental and control
• gt Kansas City Preventive Patrol
  Study (Kelling, Pate,
• Dieckman, and Brown, 1974)

8

• 6. longitudinal studies
• a. investigations examining change
  occurring over time
• b. basic model
• T1 X
  T2
• (base-line (treatment
  (measure of
• measure administered)
  change, if any)
• taken)

9

• c. types
• 1) trend studies
• a) research seeking to
  determine attitudinal,
• behavioral, cognitive,
  opinion change, if any,
• over time
• b) repeated (same/similar)
  measures taken
• independently from
  similar units of analysis,
• over time
• c) e.g. What is the
  effect of gaining a college
• education?
  What changes occur based
• on acquiring more knowledge /
• experience?

10

• 2) panel / cohort studies
• a) examination of change,
  if any, in a specific
• sample over time
• b) repeated measures taken
  over the same
• sample at different
  points in time
• c) e.g.1 Wolfgangs
  (1972) Delinquency in a
• Birth
  Cohort study
• d) e.g.2 experiencing
  your 10-year high school
• reunion
• e) caution the Hawthorne
  Effect

11

• 7. to be discussed more thoroughly later
  (perhaps)
• a. field research / ethnography
• 1) generally a more qualitative
  than quantitative
• first-hand phenomenological
  account of the
• researchers experiences
• 2) e.g.s testimonial
  narratives, oral histories, or
• memoirs
• b. program evaluation research
• 1) determinations of the success,
  or failure of any
• policy initiative
• gt does community
  policing work?

12

• 8. secondary analysis
• a. overview
• 1) analyzing available information
  collected by
• someone else
• gt a different unit of analysis
  
• 2) can take the form of document
  or content analysis
• gt i.e., examining various
  media for underlying
• themes or
  significance
• 3) e.g.1 the Unabomer Manifesto
• 4) e.g.2 the John Gotti tapes

13

• b. advantages
• 1) cheaper and faster to conduct
  than original /
• primary analysis
• 2) can examine existing
  information and tailor-make
• your analysis / examination
• c. disadvantages
• 1) validity concerns
• gt how can information
  collected by someone
• else, for possibly
  other purposes be relevant
• for your study?
• 2) question of accuracy and
  precision

14

• d. a sampling of sources of secondary
  data available
• for criminal justice research
• 1) National Opinion Research
  Center
• a) national sample
• b) questions social concerns
  of education, public
• policy, environment,
  crime, and health
• 2) Juvenile Court Daily Statistics
• a) compiled by Office of
  Juvenile Justice and
• Delinquency Prevention
• b) national aggregate
  compilation of juvenile and
• family court cases

15

• 3) National Jail Census
• a) compiled by National Institute of
  Justice
• b) aggregate head count,
  charges, and
• socio-demographic
  information of the 100
• largest lock-ups in the
  country
• 4) Sourcebook of Criminal Justice
  Statistics
• a) compiled by Bureau of
  Justice Statistics
• b) a comprehensive
  compendium of a range of
• criminal justice
  official data generated by BJS
• and agency funded
  research

16

• 5) National Crime Victimization
  Survey
• a) generated by the U.S.
  Census Bureau
• b) the crime rate
• Number of reported
  offenses
• CR
  X
  100,000
• Total U.S. Population
• c) benefits
• 1 amelioration of the
  Dark Figure
• 2 because of research
  techniques, survey
• goes to more
  representative households

17

• 
  d) limitations
• 1 telescoping
• 2 suspect validity
• 3 trivial v serious
  crime reports
• 6) Uniform Crime Reports
• a) compiled by the Federal
  Bureau of
• Investigation
• b) reported major felonies
• gt Murder / Non-negligent
  Manslaughter,
• Forcible Rape,
  Aggravated Assault,
• Burglary,
  Larceny-Theft (GE 250),
• Motor Vehicle Theft,
  and Arson

18

• c) benefits
• 1 makes sense of a
  large array of local data
• 2 standardized
  reporting of official data
• 3 good indicator of
  big city police
• bookkeeping
  procedures
• d) limitations
• 1 initial bias /
  discretion
• 2 rural v urban bias
• 3 under-reporting of
  Part II and white-collar
• crime
• 4 reporting hierarchy
  problem

19

• 7) National Incident-Based
  Reporting System
• a) began in 1991 by
  Department of Justice
• upgrade of UCR
• b) reports individual
  crimes, not aggregated
• summary of crime types
• c) no offense hierarchy
• d) taxes resources of small
  departments and
• clearinghouse locale

20

• 9. self-report questionnaires
• a. overview
• 1) traditional paper-pencil ask
  question / get
• answer
• 2) most popular form of social
  science data
• gathering strategy
• b. advantages
• 1) allows great freedom in
  seeking and acquiring
• information to support
  theoretical constructs
• 2) strong validity
• 3) good for large samples with
  lots of variables

21

• 5) ameliorates the Dark Figure
• c. limitations
• 1) no standardization
• 2) much research comes from
  juvenile populations
• 3) fear of social desirability /
  Hawthorne Effect
• 4) sampling bias
• d. Types
• 1) survey face-to-face interviews
• a) predetermined questions
  are constructed and
• asked in a specific
  order

22

• 
  c) advantages
• 1
  high response rate
• 2 guards against confusing
  and ambiguous
• questions
• 3 interviewer can
  interpret respondents
• physical cues
• d) limitations
• 1 hermeneutics
• 
  gt how good a listener
  are you?
• 2 poor reliability

23

• 2) interactive
  interviews
• 
  a) focused conversation
• 
  b) must remember what areas
  (questions) are
• important to the
  research endeavor
• c) listen and respond
• gt
  focus must be on the answer the
• respondent, is
  providing
• 3) mailed surveys
• a) best strategy for a
  broad, well-funded project
• gt large number of
  variables (questions), large
• number of
  respondents

24

• b) funding is critical due to
  reproduction and
• postal charges and
  considerations
• c) strong validity / weak
  reliability
• d) process
• 1 cover letter,
  questionnaire, instructions,
• self-addressed
  stamped envelope
• 2 addresses
• 3 consider follow-ups
• e) concern over response
  rate
• gt likely best rate to
  expect is between
• 40-60

25

• 4) the self-report
• a) by far the most popular
  data gathering
• strategy
• b) questionnaires do NOT have to
  be a
• collection of
  questions
• c) use various scaling
  techniques
• d) open / closed-ended
  questions
• 1 include both
• 2 OE items allow
  the respondent a chance
• to interact with
  the questionnaire
• 
  gt
  CE items are forced choices

26

• 3 OE
  questions may generate
• unforeseen,
  unanticipated, interesting,
• irrelevant
  responses
• 
  
  gt CE items may not provide enough
• choices
• 4 a caution on CE
  items, remember to
• insure that the
  attributes (responses)
• are mutually
  exclusive and exhaustive
• gt allow the
  respondent opportunity to
• answer the
  question honestly and
• 
  appropriately

27

• e)
  which type of answer will generate the most /
• best information
• 
  f) lack of response may be as
  informative as an
• abundance of garbled
  responses
• g) issues
• 1 clarity
• 2 avoid double-barreled
  questions
• 3 beware of respondent
  competency
• 4 be parsimonious
• 5 avoid biased
  questions

28

• 
  h) questionnaire construction
• 1 cosmetically
  attractive
• 
  2 order of items
• 3 clear
  instructions
• i) remember this is
  your test instrument
• 1 what is it you are
  looking for?
• 
  2 what are you trying to
  prove?
• 3 you are tying to
  ameliorate the Dark
• Figure
• 4 SRs are good at
  maintaining confidentiality

29

• 5 SRs good on
  validity, less so on reliability
• gt secondary sources are
  vice versa
• 6 remember, this is your
  attempt at gaining
• The Truth

30
C. Experimental / Internal Validity

• 1. overview
• a. addresses the issue of relevance
• gt is your study valid?
• b. def the relevance between the
  dependent variable and the
• operational definition
  created to explain it
• gt obtaining the results you
  actually intended / hoped to obtain
• 2. problems that may affect respondents
  responses
• a. history / current events
• 1) some life issue, unanticipated
  event, may directly affect
• performance or response
• 2) e.g., physical illness,
  epiphany, affect of substance abuse

31

• 2) increased awareness /
  sophistication in
• completing the examination /
  boredom
• c. statistical regression
• 1) a mathematical convention
• 2) in a normal distribution, all
  responses will
• eventually form a bell-shaped
  curve
• 3) responses will regress toward
  the mean
• 4) scores at the extremes of the
  distribution are
• rare
• d. sample selection bias
• gt do those elements that / who
  participate actually

32

• e. subject mortality
• 1) respondent drop-outs cause
  statistical
• quandaries
• gt most social science
  statistical models assume
• 100 completion rates from
  samples
• 2) a form of sampling bias

33
D. External (In)Validity

• 1. addresses the real concern of whether or not
  the
• results gained in any one experiment can be
  used in
• any other research setting
• 2. can the results of your experiment be
  generalized
• beyond the parameters of your research
  setting?
• gt will the results be useful beyond the
  parameters of
• your research setting
• 3. do results come from a Hawthorne Effect?
• 4. some programs and policy initiatives are
  effective
• specifically because of the charismatic
  personality in
• charge

34
E. Summary

• 1. research designed is based SOLELY on the
  type of
• information desired
• gt remember what your research question
  is
• 2. does your research actually reflect that
  which you had
• intended
• 3. there is no guarantee that the perfect
  model or project
• can be designed
• gt problems inevitably will always
  occur