Research Design

1
Research Design Analysis 2 Class 14

• Small n-designs (continued)
• Review of and more on External and Ecological
  Validity
• Using animals in research

2
2-Way ANOVA Output From Lab
Note F-ratios are the Mean square terms for the
effect being tested divided by the mean square
error.
3
Characteristics of single-subject baseline (from
table 10-1 of text)

• Individuals are observed under each of several
  phases with multiple records under each phase
• Extensive baseline observations are made
• Each subject is observed under all phases with
  each treatment repeated at least twice
  (intrasubject replication)
• A stability criterion is often used
• Multiple subjects may be used (intersubject
  replication)

4
Functions of the Baseline

• 1) establishes level of the DV prior to
  intervention
• 2) allows assessment of variability in DV
• May involve a stability criterion to minimize
  error variance so that any effect of the
  intervention will be more apparent

5
Assessing Generalizabiliy

• Intersubject Replication
• Present data from each subject individually

6
Fig 10-1
7
Fig 10-3
8
Assessing Reliability

• With single-subject designs, reliability is
  assessed through intrasubject replication
• (e.g., ABAB design)
• in contrast with
• Group designs in which inferential statistics are
  used to assess reliability after a fairly complex
  chain of logical inferences
• Note that in both cases, the key is degree of
  overlap of distributions

9
Dealing with Error Variance

• Philosophical difference between single-subject
  approach and group approach in dealing with error
  variance.
• With group approach, error variance is controlled
  somewhat, either experimentally (through
  randomization) OR statistically - if effects are
  significant (can reject the null) little further
  effort is made to determine the sources of error
  variance.
• With Small-n designs, control as much as possible
  experimentally

10
Time-Series (Small-N) Designs

• Dont involve random assignment
• control and manipulate variables sequentially
  (not simultaneously as in analytic experiments)
• anything else that occurred in this time is
  therefore a potential confound.
• Remember with experimentation we look after
  these things with simultaneous manipulations
  (between groups designs) or by counterbalancing
  (within groups designs)

11
Types of Replication

• Intrasubject - to assess reliability (given time
  confound)
• Intersubject - to assess generalizability
• also
• Distinction between systematic replication
  (introduces some extension or variation on the
  original research - often to assess
  generalizability of the phenomenon) and direct
  replication

12
Variation in baseline

• Variability can be due to
• 1) chance variation
• 2) carryover effects

13
Problematic Baselines

• Drifting baselines
• Unrecoverable baselines (due to carryover
  effects)
• Unequal baselines between subjects
• Inappropriate baselines -- floor and ceiling
  effects

14
Drifting Baseline
15
Small-n Design Terminology

• Not standardized but
• Single-factor
• e.g., AB, ABA, ABAB (reversal design),
  multilevel (ABACA..)
• Multi-factor designs (gt 1 IV)
• Multiple baseline design (gt1 DV)

16
Multivalent Single-Factor Small-n Designs
A-baseline B-Placebo cookie C- artificially
coloured cookie
17
AB (ABA) Type Non-Experimental Studies

• Example Drunk driving in Michigan
• legal age 21 ? 18 in 1972, then 18 ? 21 in 1978
• alcohol related traffic accidents 15 ? 22 in
  72, decreased again in 78
• Does a lower drinking age cause alcohol related
  traffic deaths?
• Confounds?
• Wider alcohol availability in 72
• oil crisis in 78 -- 55 mph limit imposed in
  1973/74

18
Other Examples of AB Studies

• TV effects on children Tannis MacBeth-Williams
• David Phillips Research
• Motor vehicle accidents after publicized suicides
• Airplane accidents after publicized
  murder-suicides (non-equivalent control groups)
• Homicides after heavyweight prize fights

19
Berkowitz 70
20
PhillipsCar suicides?
21
PhillipsNoncommercial Plane suicides?
A all plane crash fatalities B multifatality
crashes C Single fatality crashes
22
PhillipsConverging evidence?
23
Phillipsprizefights
Highest peak in homicides on day 3
24
Multiple baselines
25
Liberman Smith 1972
26
Non-Equivalent Control Groups

• Different than AB type non-experimental studies
  in that a comparison individual is also studied
• e.g., Mining Safety study Lucky Star mine
  (Overhead)

27
Non-Equivalent Control Groups
28
External and Ecological ValidityUsing Animals
in Research
29
Stress and Cancer
93 lab rats randomly assigned from a defined
population
Yoked control
Based on Visintainer, Volpicelli, Seligman, 1982
30
Results
31
Factors Influencing External Validity

• Issue of generalizability beyond the specific
  experimental conditions
• Population sampled
• Operational definitions
• Parameter values
• Demand characteristics
• Ecological validity?

32
Factors Influencing External Validity

• Population sampled
• be careful generalizing beyond population studied
• look for converging evidence that there is
  nothing importantly unique
• External validity not crucial for basic research
• Sample - to - population generalizations
• Sometimes unusual populations are sought out
• HIV resistant individuals, spotted hyenas

33
Spotted Hyena
34
Factors Influencing External Validity

• Operational definitions
• construct validity
• Unavoidable shock-- unnatural but are effects
  unique?
• Look for converging evidence

35
Operational Definitions of Stress
Phenomena associated with stress
Operational definition of stress
Overcrowding
Gastric ulcers
Inescapable shock
Depression
Sensory deprivation
High blood pressure
Sleep deprivation
36
Factors Influencing External Validity

• Parameter values
• Values selected for each variable
• both independent and controlled

37
Factors Influencing External Validity

• Demand characteristics
• subtle cues in a research procedure that
  influence the participants
• serious problem in social sciences
• characteristics of volunteers
• can influence both internal and external validity
• Students holding poisonous snakes, throwing
  acid in anothers face

38
Factors Influencing External Validity

• Ecological validity
• How generalizable are the experimental results to
  the specific set of conditions -- those of the
  natural context in which the phenomena usually
  occurs

39
Research Examples Alcohol consumption

• Alcohol consumption extremely stable in adults
  (rgt.85)
• Substantial individual variability
• Four patterns of drinking
• Abusive binge drinking
• Heavy steady drinking
• Moderate drinking
• Avoidance of drinking

40
Research Examples Alcohol consumption

• Abusive drinkers also prefer stronger drinks
• Can be distinguished pharmacologically
• lower 5-HIAA levels (serotonin metabolite)

41
Abusive drinking ...
42
Research Examples Alcohol consumption

• Females gt Males to be moderate drinkers
• Females prefer sweetened mix
• Subjects

43
Fruit flies alcoholism

• Dr. Vanessa Auld Dept. of Zoology at the
  University of British Columbia studies alcoholism
  in fruit flies Drosophola spp. (with alcohol
  vapor)
• Researchers have been studying fruit flies to see
  if they can understand alcoholism in humans.
• Increased activity and courting, then poor
  coordination etc. then pass out.
• Quirk quarks Jan 15th approx 1900 in Real Audio

44
Fruit flies alcoholism

• Inebriometer DV time to fall Norm approx
  20min, mutant, 12 min (sensitive to alcohol)
• Isolated cheap-date gene
• had already been studied in learning and memory
  (amnesiac gene)
• First evidence for a defect in the brain
  associated with alcohol susceptibility
• Looking at the mechanism of alcohols long term
  effects on the brain
• Quirks quarks Jan 15th _at_ approximately 1900 in
  Real Audio

45
Using animals in psychological research

• Much of what we know in psychology comes from
  animal research
• Models of human psychopathologies
• Basic research on how the brain works,
  perception, learning, memory, motivation
• Which animal to use, depends on
• research question, previous literature,
  facilities, availability/cost

46
Using Animals in Psychological Research

• Why?
• Ethical reasons e.g., effects of brain lesions,
  extreme levels of some IVs that can be studied in
  humans
• Greater control and internal validity
• Convenient
• Ethical considerations and guidelines (CCAC)
• Cost-benefit ratio
• Generality (external validity question)

47
Thalidomide
48
Prenatal stress feminized behaviour

• Environmentally stressed females give birth to
  male rats/mice whose behaviour is feminized and
  demasculinized
• Prenatally stressed males display less sexual
  behaviour
• If castrated and given female hormones, more
  likely to show lordosis
• Show more paternal behaviour
• Less masculine anatomy and SDN-POA

49
Using animals in psychological research

• Animal rights movement
• Alternatives to using animals
• in vitro studies
• computer simulations
• there are no viable alternatives to using
  animals in behavioral research

50
(No Transcript)
51
Scientific Explanations

• Science is a self correcting process
• single research findings are seldom conclusive
• problem especially in applied areas where answers
  are needed quickly