Measuring Social Life

1
Measuring Social Life

• Ch. 5, pp. 112-137

2
Measuring Social Life

• Connecting the specifics you observe in the
  empirical world to an abstract idea you cannot
  see directly
• Inferring from this sample or measure to an
  entire population or to abstract ideas
• making generalizations

3
WHY MEASURE?

• Measurement transforms our ideas and general
  observations into specific and concrete data
• Measuring helps communicate thoughts and
  observations more effectively

4
MAKING ASPECTS OF THE SOCIAL WORLD VISIBLE

• Measurement extends the range of our senses
• Scientific measurement produces a more accurate
  measure than ordinary experience, and it varies
  less with the specific observer
• Measurement makes visible ideas that are
  otherwise unseen

5
MEASURING with NUMBERS or WORDS

• In all research, data is collected systematically
• Depending on whether data are quantitative or
  qualitative, the process differs in 4 ways
• Timing
• Direction
• Data form
• Linkages

6
Two Parts of the Measurement Process

• All measurement builds on two processes
• conceptualization
• operationalization

7
Conceptualization

• conceptualization "refining an idea by giving it
  a very clear, explicit definition" (117)
• conceptual definition "defining a variable or
  concept in theoretical terms with assumptions and
  references to other concepts" (118)

8
Operationalization

• operationalization "the process of linking a
  conceptual definition with a specific set of
  measures" (117)
• operational definition "defining a concept as
  specific operations or actions that you carry out
  to measure it" (117)

9
Quantitative Conceptualization
Operationalization

• Measuring quantitative data flows in a 3-part
  sequence
• conceptualization think through the idea and
  create a conceptual definition
• operationalization link the conceptual
  definition to specific measurement procedures
• measurement apply the operational definition to
  collect the data

10
The measurement process connects three levels of
reality, from abstract to concrete

• conceptual, operational, and empirical
• conceptual hypothesis stating a hypothesis with
  the variables as abstract concepts
• empirical hypothesis the hypothesis stated in
  terms of specific measures of variables

11
Racially biased policing determinants of citizen
perception

• Whether a person is a member of the dominant or
  nondominant racial group
• A persons belief that the police are or are not
  racially biased
• Number and type of experiences with the local
  police
• Amount of exposure to media reports about police
  actions of corruption or brutality

12
Fig. 5.1 Conceptualization Operationalization
Abstract Construct to Concrete Measure
Independent Variable
Dependent Variable
Hypothetical
Abstract Construct
Abstract Construct
Causal Relationship
theoretical level
Conceptualization
Conceptualization
Conceptual Definition
Conceptual Definition
operational level
Operationalization
Operationalization
Tested Empirical Hypothesis
empirical level
Indicator or Measure
Indicator or Measure
13
Qualitative Conceptualization and
Operationalization

• In qualitative research, you use basic working
  ideas during the data collection process,
  rethinking old ideas and developing new ideas
  based on observations
• Qualitative measurement is integrated with other
  parts of a study

14
Naturalization of white culture?

• naturalization means that a culturea set of
  values, outlooks, assumptionsis so fully taken
  for granted that it becomes invisible
• white culture is a culture associated with the
  white racial group

15
HOW TO CREATE GOOD MEASURES Reliability
Validity

• reliability a feature of measuresthe method of
  measuring is dependable and consistent
• validity a feature of measuresthe concept of
  interest closely matches the method used to
  measure it
• you are actually measuring what you say you are
  measuring

16
Measurement validity is the fit between
conceptual operational definitions

• Three types of measurement validity
• face validity
• content validity
• criterion validity

17
Putting Reliability and Validity Together

• Reliability is a necessary but not sufficient
  condition for validity
• ? You can have a reliable measure that is invalid

18
Levels of measurement

• levels of measurement the degree a measure is
  refined or precise
• the way in which you conceptualize variable
  limits the levels of measurement you can use

19
Continuous discrete variables

• continuous variable a variable that can be
  measured with numbers that can be subdivided into
  smaller increments
• has an infinite of values that flow along a
  continuum
• discrete variable a variable measured with a
  limited number of fixed categories
• has a fixed set of separate values or categories,
  instead of smooth continuum, discrete variables
  have 2 or more distinct categories

20
Levels of Measurement

• Nominal measures only indicate a difference among
  categories
• Ordinal measures indicate a difference among
  categories, and the categories can be order or
  ranked
• Interval measures do everything above, plus
  specify the distance between categories
• Ratio measures do everything all the other levels
  do, plus they have true zero

21
Specialized Measures Scales and Indexes

• scale a measure that captures a concepts
  intensity, direction, or level at the ordinal
  level measurement
• index a composite measure that combines several
  indicators into a single score

22
Mutually exclusive and exhaustive attributes

• mutually exclusive each unit fits into one, and
  only one, category of a variable
• exhaustive all units fit into some category of a
  variable

23
Unidimensionality

• unidimensionality all items of an index or scale
  measure the same concept or have a common
  dimension

24
ADDING MEASURES TO GET A SCORE INDEX CONSTRUCTION

• To create an index, you combine two or more items
  into a single numerical score
• examples
• FBI crime index
• consumer price index (CPI)
• index of leading economic indicators
• consumer confidence index (CCI)
• (The Conference Board Consumer Confidence Index
  )

25
(No Transcript)
26
Two Complications in Index Construction

• Count items equally or weigh them?
• Unless you have a very good reason, it is usually
  best to weight them equally
• In a weighted index, you value or weigh items
  differently, depending on your conceptualization,
  assumptions, conceptual definition, or
  specialized statistical techniques
• Missing data
• - If data for one of your items (in a 4-item
  index) is missing for some of your cases (e.g.,
  in a societal development index, literacy data is
  missing for 3 of 50 countries, you must decide
  whether to drop the cases (3 countries) or
  substitute weaker measures (using only 3 items in
  your index)

27
CAPTURING INTENSITY SCALE CONSTRUCTION

• Most scales help us measure the intensity,
  hardness or extremity of a persons
  feelings/opinion at the ordinal level
• The simplest scale is a visual rating
• e.g., a feeling thermometer" is used to see how
  people feel about various groups in society,
  political candidates, public issues, etc.

28
Likert scale

• The Likert scale offers a statement or questions,
  and respondents indicate their response with a
  set of answer choices, such as strongly agree,
  agree, disagree, or strongly disagree, or
• - approve/disapprove of X
• - support/oppose X
• - believe X is always/never true
• - do X frequently/rarely