Conceptualization, Operationalization, and Measurement: Measurement Options

1
Conceptualization, Operationalization, and
MeasurementMeasurement Options

• October 25, 2004
• Sociology 487

2
Lecture Topics

• Asking Questions
• Single questions
• Question Sets
• Other Measurement Options
• Direct Observations
• Unobtrusive Measures

3
From Concept to Measurement

• Progression from sense of what a term means to
  measurement in a scientific study
• Conceptualization
• Nominal Definition
• Operational Definition
• Measurements in the Real World

4
Measurement Operations

• Measurement
• Process of linking abstract concepts to empirical
  indicants
• Operation
• Procedure for identifying or indicating the value
  of cases on a variable
• Single or Multiple Indicators

5
Asking Questions

• Different types of questions used in social
  research
• Single questions
• Closed-ended
• Open-ended
• Question sets
• Indexes
• scales

6
Closed-Ended Questions

• Offer respondents explicit response choices
• Advantages
• Easy to code
• Easy to analyze with statistics
• Reduce ambiguity
• Easy for respondents to answer
• Disadvantage
• Can obscure what people really think if response
  set is inadequate

7
Closed-Ended Questions continues

• Response choices must be
• Exhaustive
• Mutually exclusive
• Exceptions
• Partially closed Questions
• Questions that ask respondents to check all that
  apply

8
Open-Ended Questions

• Lack explicit response choices
• Advantages
• Useful when full range of responses cannot be
  anticipated
• Questions not used in previous surveys
• Questions asked of new groups
• Reduces confusion about meaning of responses
  involving complex questions
• Disadvantages
• Questions require careful review for clarity
  before use
• Can be used with only small number of people

9
Question Sets

• Index
• Composite measure based on responses to multiple
  questions measuring same concept
• Reponses to questions that comprise an index
  should be correlated
• Can assess correlation with reliability statistics

10
Advantages to Using an Index

• More comprehensive measure of concept than any
  one individual question
• Many pre-existing indexes already used in
  research
• Reliability proven
• Maximizes comparability

11
Examples

• College Performance
• Grade Point Average
• Depression
• CES-D

12
Cautions to Consider when Using an Index

• Individual questions may not measure same concept
• Combining responses to multiple questions can
  obscure important differences in meaning
• Questions may cluster in a multidimensional index

13
Calculating an Index

• Arithmatic average
• sum responses with all questions counting equally
  
• Weighted Scale
• Giving different weights to responses to
  different questions before summing or averaging
  responses

14
How to Design Clear Questions

• Avoid confusing phrasing
• Double negatives
• Double-barreled questions
• Minimize risk of bias
• Reduce likelihood of agreement bias
• Minimize fence-sitting and floating
• Exhaustive and mutually exclusive response
  choices

15
Disadvantages to Measurement with Questions

• Social desirability bias
• Poor approach for measuring behavior that is
• Socially stigmatized
• Illegal

16
Other Measurement Options

• Direct Observations
• Unobtrusive Measures
• Collecting data about individuals or groups
  without their knowledge

17
Direct Observations

• Advantages
• Good supplement to interview / survey data
• Optimal measurement for research in natural
  setting
• Avoids problems of recall and social
  desirability bias
• Disadvantages
• Can not observe everything
• Observations filtered by observe
• Presence of observer may affect behavior of
  respondents

18
Unobtrusive Measures

• Types of Unobtrusive Measures
• Physical trace evidence
• Archival Data
• Simple observation
• Contrived Observation

19
Combining Measurement Options

• Triangulation
• Use of two or more different measures of same
  variable
• Strengthens measurement
• Particularly when collected via different methods
  of data collection

20
Key Points

• Single question measures may be
• Closed-ended
• Partially closed
• Open-ended
• Question sets may be used to operationalize a
  concept
• Indexes and scales measure a concept by combining
  answers to several questions