exam 2 Flashcards
concepts
words, phrases or symbols in language that are used to represent these mental images in communication
ex. serious crime
dimensions
groups of indicators that reprint the aspects of the concept specifiable aspect of concept
ex . crime seriousness
- victim harm dimension
; indictors (physical injury, economic loss, psych consequences
conceptual def
working def. specifically assigned to term, provides focus to our observations
- gives up specific working definitions so that readers will understand the concept
operationalization
def. that spells out the measurement of the concept (not the meaning of the concept = conceptualization)
developing operational definitions
spells out precisely ho the concept will be measured
Be able to take the operational definitions and use them as specific measurements in a scientific study
- different measurement can produce different results
ex. time frame in which recidivism is measured might produce different results
-create the variables (categorization within variables)
ex. counties in California (variable)
(1) orange (2) Los Angeles (3) San Diego
applying conceptualization and operationalization example
recidivism
- You are going to do a proposal about the effect of a treatment program (X) on recidivism (Y) after release from prison (parole or direct release).
- How would you conceptualize recidivism?
ᾆ Recurrence of criminal behavior within a given period of time characterized by RE-ARRESTS, RECONVICTIONS, & REINCARCERATIONS
-How would you operationalize it? (example only
showing for reincarceration)
ᾆ Reincarceration for a technical violation or new offense within 1 year of release
ᾆ Reincarceration should be measured using “rap sheets” collected by the State’s DOC criminal history database
ᾆ Technical violation should be measured according to state guidelines
Understand operationalization and conceptualization together
- once conceptualization is completed, we have started to set up operationalization process (development of actual measures)
two qualities of a variable
exhaustive
mutually exclusive
understand exhaustive and exclusive measurement.
- Exhaustive: Classification for EVERY observation
into one attribute comprising the variable
(one must occur ) - Mutually exclusive: Every observation can only fit into ONE attribute. (can not occur at the same time)
but does not mean one has to occur
ordinal lv of measurement
attributes can be logically ranked but no scale across that order (nominal +rank order)
ex. income
$0-5000
$5001-10,000
nominal lv of measurement
attributes indicated by names w/o scale or hierarchy
ex. survey times that ask; Gender
male or female
by kind not quantity
reliability
does a particular measurement technique, applied repeated to the same thing, yield same result each time?
ex. IQ test: if person takes and retakes the exam will the same result occur (in so as all conditions are equal b/w settings)
validity
he measure accurately reflects the meaning of
the concept under consideration. Are you accurately
measuring what you say you are measuring?
operational definition really reflect the concept?
● The following is a poor operational definition of aggression:
● Concept: aggression among children
● Operational definition: Survey item: How many times have
you hit your sister/brother in the past year?
Criterion-related validity
Use a standard or benchmark to assess the
validity of your measure.
Substantiating PROXY measure for construct.
● You don’t have a measure of criminal behavior in a prison sample, so you use technical violations. You substantiate the use of technical violations by demonstrating that technical violations are highly related to criminal behavior by citing other’s work that shows this to be true.
Revisiting Measurement
The process of assigning numbers and/or labels to units of
analysis in order to represent conceptual
properties
Content validity
Does the measure cover the range of meanings of the
concept?
● One measure of fear of crime: How fearful are you to go outside late at night?
Narrow definition of fear of crime
- with content you have to ask more questions and less vague
Construct validity
The extent to which your items are tapping into the
underlying theory or model of behavior. (concept validity)
● One way: Multiple measures: compare measure with alternative measures of the same concept
● Does our theory of fear of crime support the variables that we are using to measure it? E.g.
Going out at night does not really reflect one’s overall fear of crime.
face validity
Does the operational definition of the concept make
sense?
● Does prison A have more violent prisoners than prison B? Compare the number of violent crimes and technical violations between prison samples.
Good face validity
different categories of validity
face
criterion related
construct
content
different type of tests reliability
test retest: doing it over and over agin
inter rater: have 2nd researcher get a different pov
how to produce valid indicators
Confidence increases with the use of multiple indicators
● The best reliable and valid measures usually use multiple
indicators (composite measure) <— (use of index and scales)
scale
Distinguishes indicators in terms of the
quantity/intensity of the concept each indicator is
suppose to reflect
● Then the items are ordered in some way by their
strength/intensity
● (Note: Just b/c something is called or referred to as a
“scale” might not necessarily be reflective of a true
scale)
▪ Eg. bogardus social distance (how willing you are to be associated with sec offenders (live in country, neighbor, next door, marry, etc)
- if person answers at the neighborhood one we are almost 100% they will say no to marry
index (composite measure)
Index – composite measure which is constructed
by combining different indicators of a specific
concept or dimension
● Most indexes are created by simply adding up the
scores of a set of indicators
● Provides an ordinal ranking of cases on a given
variable
composite measures
index
scale
UCR purpose and measurements of crime
FBI measurement of crime
summary based measure
reporting agencies are cities or countries
NCVS measurements of crime and its purpose
Criminal Victimization
– Frequency
– Characteristics
– Consequences
* Crime types include (Divided into personal and
property)
– Rape and sexual assault
– Robbery
– Assault (simple/aggravated)
– Pick-pocketing/purse snatching
– Theft
– Household burglary
– Motor vehicle theft
how do police based measures contain error
- Police know only a portion of the crimes that actually
happen.
– Examples of types of crimes that can go undetected: drug sales and
drug possession, traffic offenses, sex offenses, minor assaults, and
minor stolen property. - As well, police do not record every single crime they
encounter. - Does not try to count ALL crime reported to the
police, only collects index offenses…
* Reported to the police: Part I (8): Murder/non-
negligent manslaughter, forcible rape, robbery,
aggravated assault, burglary, larceny-theft, MV theft,
arson.
* Reported, arrested, and charged offender: Part II:
Drug sale and possession, fraud, prostitution, and
simple assault.
UCR strengths and weaknesses
- Good: History (trends since 1930), covers 95% of US
– Bad: Definitions, hierarchy rule, only official crime data
NIBR strengths and weaknesses
– Good: Improvement over UCR (incident-level reporting of
official crime and more crime types are collected)
– Bad: Not fully implemented (problems with implementation)
how does crime get measured through the use of victim surveys
– Screening instrument (promote recall, reduce
subjectiveness of whether or not an incident was a
crime, actively seek reporting when offender is
known)
* Detailed questions
* CUES respondents (examples of places, activities—typical
settings)
* No CJ jargon (i.e. Aggravated assault)
– Criminal incident report picks up detailed information
garnered from the screening instrument
victim surveys vs crimes know to police
UCR figures are counts of crime, NCVS
are statistics derived from a sample
- Definitional and categorization differences
– I.e. Rape
– I.e. Violent and property crime - UCR are summary-data, NCVS are
individual-level data
validity and reliability of victim report surveys
- Recall error—inaccurate memory (particular to
minor crime) and telescoping (when did crime
occur?) - Series victimization (several separate
victimizations that the respondent cannot
clearly describe separately to the interviewer) - Lack of honesty
* Rape, which also requires the victim to recognize
the crime as a “rape”
* Personal problem
alt reasons for measuring crime? specific research
policy purpose ?
Professor Smith gave an exam on Monday. On Wednesday Smith gave the same class the same exam. Professor Smith was clearly interested in assessing the exam’s
a. reliability
b. validity
c. face validity
d. conceptualization
e. precision
a. reliability *
b. validity
c. face validity
d. conceptualization
e. precision
Reliability involves
a. whether a particular technique applied repeatedly to the same object would yield the same results each time
b. ensuring accuracy
c. ensuring that your measure measures what you think it should measure
d. ensuring precision
e. all of the above *
e. all of the above *
Professor Koo decided to define socioeconomic status as a combination of income and education. Socioeconomic status
a. is a dimension.
b. is a composite measure.
c. is comprised of multiple dimensions
d. is comprised of a single indicator
e. a & d
f. b & c
g. a & b
h. a & c
i. c & d
f. b & c
Conceptions are
a. empirical measurements
b. variables
c. definitions
d. hypotheses
e. mental images
e. mental images
Troiden and Jendrek used the SAS (sexual attitudes) questions to examine the link between sexual attitudes and sexual experience. They reasoned that people expressing liberal sexual attitudes would have a wider range of sexual experiences than those expressing conservative attitudes. They were concerned with validating the SAS items using
a. split‐half validity
b. face validity
c. criterion‐related validity
d. content validity
e. construct validity
e. construct validity
Hudson et al. developed a series of questions to examine sexual attitudes (SAS). The SAS scores of religious fundamentalists, a group known to be conservative regarding sexual expression, were compared with the scores of social work graduate students, a group known to be liberal regarding sexual expression. The researchers were examining the instrument’s
a. content validity
b. face validity
c. criterion related validity
d. criterion related reliability
e. construct validity
c. criterion related validity
The result of combining several indicators of a variable is a composite measurement of the variable. (T/F?)
true
A particular variable can usually be measured in several distinct ways using different sources of information and various observation techniques. (T/F?
true
Conceptualization is the development of specific research procedures that will result in empirical observations representing those concepts in the real world. (T/F)?
false
A nominal measure can have only two categories. (T/F?)
false
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. (T/F?)
false
Changing definitions almost inevitably results in different descriptive conclusions. (T/F?)
true
Precision and accuracy are synonyms. (T/F)?
false
If a measure is reliable, it must also be valid. (T/F?)
false
Precise measurement is more important than accurate measurement. (T/F?)
false
It is impossible to have several indicators of only one concept. (T/F?)
false
What is the lowest level of measurement in which there is an exact difference between attribute values?
a. nominal
b. interval
c. ratio
d. ordinal
e. all of the above
interval
The variable educational level was measured as last year in school completed (i.e., none, 1st, 2nd, 3rd, grades, etc.). It is, therefore, measured at the level.
a. nominal
b. interval
c. ratio
d. ordinal
e. not enough information to decide
ratio
Professor Tilton measured the variable “feelings toward drafting women” with the categories strongly agree, agree, indifferent, disagree, and strongly disagree. Professor Tilton was using the level of measurement.
a. nominal
b. interval
c. ratio
d. ordinal
e. not enough information to decide
ordinal
Classifying someone as employed or not employed treats employment as
a. a ratio variable
b. an interval variable
c. an ordinal variable
d. a nominal variable
e. a dependent variable
nominal variable
Measuring how people feel about proposed income tax hikes when you really want to know how well informed they are on the proposal is a problem of operationalization concerning
a. the use of single or multiple indicators
b. how observations are going to be made
c. the specific dimensions of the variable to be studied
d. the relevant range of variations
e. the degree of precision needed between extremes
c. the specific dimensions of the variable to be studied
Which of the following is a nominal variable?
a. education
b. age
c. employment status
d. occupational prestige
e. not enough information to know
e. not enough information to know
Professor Myth asked respondents whether or not they had ever been divorced. One year later Professor Myth asked respondents the same question. Myth found that with repeated applications of the measure different responses were obtained for the same respondent. This means that the measuring instrument was
a. unreliable
b. invalid
c. unreliable or the value on the variable had changed
d. lacking face validity
e. inaccurate
c. unreliable or the value on the variable had changed
ordinal lv of measurement
attributes can be logically ranked but no scale across that order (nominal +rank order)
ex. income
$0-5000
$5001-10,000
interval
attributes collect w/logical rank order and scale across that order is equal: meaningful distance b/w attributes
no absolute zero
ex. IQ and temp
ratio lv of measurement
attributes collected w/logical rank order and scale across that order equal and true zero
ex age and income
what does high lv of measurement give you ?
- give you a better accurate measurement
- generally more numerous differ nation b/t attributes
what is the highest lv of measurement ?
ratio
when deciding measurement design
what do you take in for account ?
range of variation: you want to capture how your group varies (ex.based on area)
precision of measurement
Revisiting Measurement
Revisiting Measurement: The process of
assigning numbers and/or labels to units of
analysis in order to represent conceptual
properties
