PSY 3392 Study Guide - Midterm Guide: Uncertainty Principle, Naturalistic Observation, Participant Observation
27 Mar 2019
School
Department
Course
Professor
CHAPTER 4: OBSERVATIONAL DESIGNS
• Descriptive methods can establish correlations, but DO NOT demonstrate causation
o True experiments use random assignment of participants to IV levels to establish
causation
• Observational Methods –
o Direct observation
▪ Without intervention (most commonly known):
• Naturalistic Observation –
o Direct observation of a behavior in a natural setting without
attempt of observer to intervene
o Goal: describe behavior as it naturally occurs
o Observes is a passive recorder
o Can often establish external validity – the extent to which a
behavior generalizes to different populations, settings, and
conditions
• Experimenter/observer effect –
o Aka Heisenberg Uncertainty Principle: every action has an
(uncertain effect)
o States that measurements of a system cannot be made without
effecting the system
o Our observations change the outcome and behavior of those
involved in the show
▪ With intervention
• Participant observation (least amount of intervention)
o Observe and participate
o Open/undisguised – individuals who are being observed, know
the observer is present for purpose of collection information
about their behavior
▪ Often used by anthropologists
o Disguised – those who are being observed, do not know they are
being observed
▪ Avoids major problem when observing behavior is
reactivity – occurs when people react to the fact they are
being observed by changing their normal behavior
▪ Allows them to gain access to situation that is not usually
open to scientific observation
o Self-observation – report on what you are doing
o Example of design: disguised participant observation (with
intervention)
▪ Classic study of psychiatric diagnosis and hospitalization
of the mentally ill (Rosenhan 1973), employed disguised
participant observers who sought admission into mental
hospitals
▪ Each complained of the same general symptom: hearing
voices
▪ Most of the pseudopatients were diagnosed with
schizophrenia
▪ Immediately after admission, participant observers
stopped complaining of any symptoms and waited to see
how long it took for a “sane” person to be released from
the hospital
▪ Hospitalized from 7-52 days and when discharged, their
schizophrenia was said to be “in remission”
o Potential problems –
▪ Loss of objectivity – observers lose their scientific
objectivity if they identify too closely with the people and
situation they are observing
• Kirkham, 1975 “Doc cop” – criminologist goes
through police academy training, as an
undisguised participant observer and became
uniform patrol office assigned to a high-crime
area
o Changed his personality and views, he lost
objectivity
▪ Privacy violations – there is an absence of informed
consent by participants when covert observers are present
(pro-ana websites)
▪ Observers may affect phenomena (even disguised)
• Structured observation – set up a situation to observe a specific event
and record it more easily (e.g., development of lying)
o Evans and Lee (2013) used a simple structured observation to
investigate lying in very young children
▪ Tested 2 and 3 children to see if they would peek in a bag
when told not to and then ask to see if they lie about
doing it
▪ 80% of 2 and 3 year old’s peek, but most 2 year old’s will
tell the truth
▪ As they grown they tend to lie more
o Different from naturalistic observation because researchers
intervene to exert some control over the events they are
observing
• Field experiment (if done well, you can state causality) – when
experimenters manipulate 1 or more independent variables in a field
setting
o Bias with direct observation
▪ Observer/experimenter effect: Heisenberg uncertainty principle
• Experimenter influences the observed
▪ Observer bias: systematic errors in observation often resulting from the
observer’s expectancies regarding the outcome of the study (i.e. “expectancy
effects”)
▪ Reactivity: when people react to the fact that they are being observed
