Nonreactive Measures

1
(No Transcript)
2
Nonreactive Measures
Webb, E. J., Campbell, D.T., Schwartz, R.D.,
Sechrest, L. Grove, J. B. (1981). Nonreactive
measures in the social sciences. Boston
Houghton Mifflin Company.
3

• Bridgman (1927) in the Logic of Modern Physics
  had asserted that every concept is defined by the
  operations used to measure it.
• Although the operational definition is not the
  only dimension of a definition of a construct, it
  is a requirement of empirical studies.
• Mono-operationism assumed that one operational
  definition existed for the construct.
• The existence of multiple operations for one
  construct was the downfall of the operationalist
  philosophy.

4

• We have recognized that a single psychological
  construct has more than one operation. Anxiety,
  for example, could be measured through
  physiological, self-report, and behavioral
  measures.
• To the degree that these measures agree
  (converge) we feel more confident about our
  definition of our construct we have achieved a
  degree of construct validity.
• We assume that each measure of our construct has
  a certain degree of agreement and a certain
  degree of independence.
• For each operation, the score a participant gives
  us, or X, is described as
• X Truth Error

5

• The philosophy of multiple operationism
  suggests that a construct is best studied with
  multiple operations that converge on the concept.
• The approach assumes that the error part of the
  measures is different. Two measures cannot be
  affected by the same confounding variable.
• For example, suppose you study sexism in young
  men by giving them a survey and then interviewing
  them. Both surveys and interviews may lead to a
  reactive problem the respondents know that
  sexism is generally disapproved so they answer
  both measures in a socially acceptable way. This
  bias infects both measures, therefore multiple
  operationism is violated.

6

• Examples of reactive measures
• Guinea Pig Effect The participants know that
  they are being studied.
• Changes in The Research Instrument Record
  keepers change their methods because they know
  you wish to inspect the records.
• Experimenter Expectancies The experimenter
  inadvertently (or deliberately) influences the
  outcome.

7
Webb, et al. suggested a solution to the problem
in 1966 in a book originally titled Unobtrusive
Measures. While taking multiple measures, why not
employ at least one that avoids human reactivity?
8

• Physical Trace Measures
• Erosion
• Sherlock Holmes comment on Dr. Watsons new
  office.
• Floor tile wear.
• Worn parts of books.
• Accretion
• Cigarettes on the floor.
• Studies of human trash.
• Accretion of laundry.

9

• Archives
• Use the CIRP for baseline student attitudes.
• Check student records for GPA, etc.
• Do background checks for security.
• Compare voting records in a district to archives
  of ethnicity or religious affiliation.
• Check hospital and police records to determine if
  a full moon affects behavior.

10

• Simple observation
• Direct
• Sit on the Forum porch and observe people.
• Go to a grocery store and watch the customers.
• Indirect
• Go to the library and observe the use of
  territorial markers.
• Use the lost letter technique.

11

• Contrived Observation
• Direct
• Watch the participant via a video camera.
• Indirect
• Record images or conversation for later coding.

12

• Examples of unobtrusive measures
• Henle and Hubble (1938) study of egocentric
  statements in college students.
• Sullivan, Queen, and Patrick (1958) study of
  attitudes of enlisted men in the Air Force.
• Humphreys (1970) study of gay men in public
  restrooms.
• Middlemist, Knowles, and Matter (1976) study of
  mens behavior in a restroom.

13

• Three queries
• Does a measure taken during deceit constitute an
  unobtrusive measure?
• What is the role of ethical guidelines in judging
  unobtrusive measures?
• Does psychology depend on the naïve subject?