Docstoc

MEASUREMENT

Document Sample
MEASUREMENT Powered By Docstoc
					VI.


MEASUREMENT
A. Introduction
 1. definition: the process of assigning a counting scheme
                to units of analysis in order to analyze
                theoretical concepts
 2. actually moving from abstract theory to concrete, real
    world examples
B. The Process
 1. conceptualization
   a. T H I N K I N G
   b. creating a mental image
   c. what do you see when someone is *doing* your
      problem? your concept?
 2. identification / specification of indicators (naming your
    variable/s)
    a. what is it you are actually trying to measure?
    b. variable: anything that can be measured
       1) these are the images, concepts, and questions you believe
          provide the solution / explanation to your problem
2) the process of identifying the relevant variables is
   called specification of indicators

3) the answers, elements, characteristics that make
   up any particular variable are its attributes

4) Dependent Variable: the concept / phenomenon
   being explained or solved by your theory
   a) the specific focus of your research
   b) sometimes called the “outcome” variable
5) Independent Variable/s: the concept / phenomenon
     expected to have an effect on the concept /
     phenomenon under examination
        a) those concepts considered that explain / solve
           your research problem
        b) sometimes called “predictors” or “indicators”
     6) these are ostensibly all your creation

3. observations
  a. direct observables
     > that which can be easily seen or identified
  b. indirect observables
     1) that which requires methodologically sophisticated
        finesse to discern
      2) e.g., stress, frustration, anger
   c. constructs
      > an abstraction requiring identification
4. operationalization
   a. the creation of articulated definitions describing, explicitly, that
      which you will be measuring
   b. i.e., the creation of operational definitions
5. levels of measurement
   a. who, what, where are your units of analysis
   b. qualities of measures
      1) exhaustive
      2) mutually exclusive
  c. types of measures
      1) nominal: named / titled
      2) ordinal: rank-ordered
      3) interval: distance between attributes relevant
      4) ratio: precise, exact attributes with absolute zero response
C. The Model

   > the thinking process



   CONCEPT               Nominal     Operational
                        Categories   Definitions




             REAL WORLD MEASUREMENTS
D. Errors in Measurement
    (mistakes made by unit of analysis)

 1. the “wrong” answer
   > eye-witness testimony
 2. sources
   a. real error
      > observation, reporting hermeneutics issue
   b. bias
      1) social desirability
      2) Hawthorne Effect
3. types of measurement error
   a. systematic: predictable
      > errors that may persist over time, place
   b. random: unpredictable
      > errors that occur inconsistently, through chance
4. controls for error
   a. check unit/s of analysis
   b. watch for experimenter bias
   c. check level of measurement
   d. caution: precision v accuracy
5. criteria for assessing quality of measures
   a. reliability
      1) dependability
      2) freedom from random error leads to more
         reliable information
      3) improving reliability
          a) pretest
          b) triangulate
          c) eliminate distractions
          d) eschew obfuscation
b. validity
   1) the assurance that the measure you take
      actually reflects that which you intend it to
      measure
   2) systematic errors lead to problems of validity
   3) unlike reliability problems, validity problems
      cannot be directly corrected
   4) available checks
      a) consult with the available literature
          > make sure your definitions are similar to
            others
      b) use your own best judgment
5) types of validity

   a) subjective / face
   b) content
   c) criterion – related
   d) construct
E. Summary
 1. measurement is your attempt to collect the
    information (data) you need to prove your point
 2. failure of reliability = random error
 3. failure of validity = systematic error
 4. any concept has no definition other than the one you
    provide

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:9
posted:3/23/2012
language:English
pages:13