Research Handouts

Rabia Farooq

Complex Designs
• Researchers often investigate the effects of two or more independent
variables simultaneously; this type of experiment is referred to as a
complex design.
• The simplest complex design has two independent variables (IVs) and
one dependent variable (DV). The simplest independent variable has two
levels (conditions
• Complex designs are also called factorial designs.
• An experiment conducted by Baaren and his colleagues (2004)
conducted an experiment using 2 x 2 design to see whether mimicry
(coping) helps in promoting pro social (helping) behaviors

2 x 2 design
• The factorial design with 2 independent variables both with 2 levels
is called a 2 x 2 design (read “2 by 2"). It has four conditions.
• So what is meant by
• 2x2 x2
• 2x2 x2 x2
Research hypothesis
• Mimicry increases helpfulness.
• 1st Experiment: Single IV was manipulated using RGD.
• IV… 2 levels….mimicry or non mimicry.
• DV… whether they were helping.
• One group was randomly assigned to mimicry or another to non
mimicry group and than DV was assessed.
• All participants in the mimicry group whereas 33% of non mimicry
group helped.
• 2nd Experiment to test whether helped the participant other than the
experimenter.
• Same IV as in 1st experiment were used.
• New experimenter was not a manipulated variable after completing
the 1st experiment the new experimenter entered in the room.
• Mimicked one 84% copy.
• None mimicked one 48 % copy.
•
Main Effect
• The overall effect of an independent variable In a complex design is called a
main effect.
• The effect one IV has on other DV or second IV have on another DV.
• The main effect is the effect of the independent variable on the
dependent variable as if only that variable was manipulated in the experiment

Interaction Effect
• The combined effect of independent variables in a complex design is called
an interaction effect.
• An interaction effect occurs when the effect of an independent
variable differs depending on the level of the second independent variable.
• The effect one IV with interaction to another IV have on DV

Subject Assignment and factorial

designs
• The independent variables can be manipulated using an independent
groups design or a repeated measures design, or both.
• Independent group
• Repeated measure
• Mixed factorial design
Advantages
• Allows to be manipulation of more than one IV in one experiment.
• Allow to see interaction effect.
• Control
• Conserve time and energy.
• Test predictions.
• Resolve contradictions
Dis advantages
• Involved more complex statistical calculations.
• Take more time.
• Impractical in many situations

Single Case Research Design (2nd Notes)

Single Case
Research Design

Single Subject
Case Experimental or
study small n research
metho design
d
Single Case Research Design
 It uses only one case or one group to investigate specific phenomena.
 It has been found useful in experiments of perception, memory, and
classical conditioning.
 Used by Freud to develop Psychoanalytic theory.
 Study individual organism.
 Case Study Method

 It is a descriptive account of behavior, Past history, and

And other relevant factors concerning an individual.
Case study is an intensive description and analysis of single individual.
It has been used by child psychologists, animal behaviorists and in the field of
criminology, neurology and sociology.
Researcher uses the case study method obtain their data from several
sources.
 Naturalistic Observation
 Interviews
 Psychological Tests
 Archival Records

Characteristics
 It frequently describes the application and results of a particular
treatment.
 E,g., a clinical report of a case study may describe an individual
symptoms, methods used to understand and treat symptoms and evidence
for treatment effectiveness… does provide a potentially rich source of
information about the individual.

They lack Control:

 They often lack high degree of control.
 Treatment variables are rarely controlled systematically Instead
several treatments are applied simultaneously psychologists have little
control over extraneous variables (home work and home environment )
influence the client.
 Without control it is difficult for a researcher to make causal inference
about variables that influence the client behavior.

Content of case study

 Content of case study is extremely varied… Published case study can

be of few pages to of a book.
 Krisch (1978) attempt to implement self management training (SMT)
with a woman experiencing low self confidence and social inhibition. Case
study describes 9 therapy sessions and 5 months follow up in just 4 pages
of one issue of journal Psychotherapy: Theory Research and Practice.
 Personality Psychologist Gordon Allport described relationship of a
mother with her son as revealed in 301 letters that mother wrote to her
friends. First reported in 2 consecutive issues of journal of Abnormal &
Social Psychology. This study was published in 1965 as a 223 page book
under the title “Letters from Jenny”.

Aspects of Control
 It is descriptive and exploratory in nature.
 In some cases is the only option…Some information is better than
none.
 Many aspects make it unique.
 Affective in isolating relationship.
 Source of hypothesis and ideas about normal and abnormal behaviors

Advantages of Case study

 Rich source of information about individual and insight into possible
causes of people’s behavior. Insights when translated I into research
hypothesis can be tested using more controlled research methods.
 It is starting point for a researcher who is entering in area of a study
about which relatively little is known.
 It provide an opportunity to try new treatments or therapeutic
techniques and application of existing techniques. Use of self management
technique.
 Useful to study rare cases. ( Feral Children …case of Aveyron).
 Challenge theories. (Case of Genie & Theory of Lenneberg, 1967)
 Support theory ( Not conclusive evidence but evidence in support).
 Idiographic approach … Detailed description of individual…
 Study Unusual condition. (Case of Sybil)
 No manipulation of IV… so study cases can’t be manipulated
variables.

Disadvantages of Case Study Method

 Can’t make valid Causal inference… Susan SMT Treated by Krisch).
 Observer Bias and bias in data collection
 Interpretation of outcome is based on subjective impression of
observer. Dramatic demonstration of findings.
 Although generalize on population from which the patient was
selected some personality characteristics vary across individuals.
 Evidence from case study is persuasive.
 Findings are often accepted uncritically.

Single subject (small n Experiments

 Simple contrast of condition within an individual whose behavior is
being monitored.
 Focus on examination of behavior change in one individual or at
most few individuals.
 IV (treatment) is manipulated for one individual so have control and
important alternative in contrast to uncontrolled case study method

Parts of small n research

 Baseline stage A
 Manipulation of independent variable B
 Independent variable removed C

Baseline Stage
 Situation or state prior to administration of IV is called a baseline.
 Selection of some behavior …dependent variable
 Repeated recording of behavior to obtain a stable measure prior to
administration of IV….
 DV should have following characteristics
 Objectively measurable
 Easily emitted by subject or emitted at length without fatigue.
 Sensitive to change that occur in situation
 E.g., bed wetting; in no of sec eye contact ; no of headaches reported
by sufferer.
 Using baseline researcher s are able to predict what will be like in
future with treatment

2nd Stage Manipulation of IV

 Treatment is administered when stability in behavior is seen or there
are little fluctuations are seen. Criteria of stability is being set before hand.
 Manipulation of IV remains until DV stabilizes ( criteria being settled
before hand)
 Same measures are used to record behavior at A and B stage.

Removal of IV
 When DV during 2nd stage stabilize the IV is removed so subjects
experienced the same situation as baseline stage.
 LOGIC OF THE DESIGN
 To demonstrate cause and effect relationship.

 Baseline A Treatment B

Advantages
 Scientifically sound investigation with one and few subjects
 Allows control
 Less chance of secondary variation
 No statistical tests required in initial designs
 Flexible
 Easily interpretable
Disadvantages
 Hard to determine causality
 Not appropriate for certain types of researches as ex post facto
researches.
 Much time in studying most individuals.
 Due to excessive variability at baseline difficult to interpret at
treatment stage.
 Limited external validity

AB design
 Simplest strategy to study an individual behavior with and without
experimental treatment.
 DV…Operationally defined and measured throughout the study.
 A… baseline of behavior (DV)
 B…. Treatment that cause change in behavior (cause change in DV).
 E.g., a teacher is interested in decreasing students time out of chair.
A (Baseline)… No of times student is not sitting for 1 hour. B (treatment)…
delivery of reinforcement every time he sits)
 Lacks internal validity
Confounding variable.
Individual might be
abnormal with respect to
behavior being measured

ABA Design A B A
 An addition of A after B
 To see whether behavior reverses. Student sitting behavior.
 Improves to identify confounding variable.
 To determine cause and effect
 2 opportunities to determine effect of treatment variable ….AB ___ BA
transition
 Weaknesses includes as study ends at A individual deprived of
treatment.

Extensions of ABA Design

 ABAB Design:
 Another B after 2nd Baseline.
 Also called reversal design.
 Researcher gains considerable evidence for behavior change.
 Also called equivalent time samples design.

 A B A B Horton (1987) study.

Advantages
 Ends at Treatment
 3 transitions to strengthens conclusions about treatment AB BA AB
 T test ANOVA can be used.
 Disadvantages:
 If behavior during 2nd treatment does not reverses
 or improvement is seen at 2nd treatment stage.
 Or confound between treatment and attention.

AB AB AB Design
 Used to reduce chance of coincidence.
 ABCB Design
 Variation of ABAB design
 See effects of two different variables
 ANOVA Can be used.

Multiple Baseline Designs

 Use Baseline and treatment and no withdrawal.

 Establishment of various baselines.
 A- B-C- BC-B- A-BC-C-A
 Interaction among variables is assessed.
 E.g., Baseline of an aggressive child is established in several
situations at home, in class room, while playing
 Then treatment is introduced only in one situation than another and
see whether behavior improves.
 Behavior change only when treatment is administered.
 Design has several variations
Multiple Baseline Design Across Individual
 Baseline is established for different individuals when behavior of each
individual stabilizes treatment is introduced first to one than to second
than to third.
 If treatment is effective change in behavior will be immediate
following application of treatment in each individual.
 Allison and Allon (1980). Evaluated effectiveness of coaching
behavior
Multiple Baseline Design Across Behaviors
 In it 2 or more baselines are established by observing different
behaviors in same individual.
 Treatment is given for one behavior than for another and so on.
 E.g., teaching socially appropriate effective behaviors
 As inappropriate behaviors limits their ability to communicate
effectively.
 Target behaviors… showing sympathy, appreciation, talking favorite
things, indicating disliking.
 Treatment… Verbal Praise & Tokens
 Visual Inspection… show effectiveness of treatment.
Multiple Baseline Design Across Situations
 In it 2 or more baselines are established across situations in same
individual.
 Treatment is administered at different time and than behavior is
recorded.
 Researches used this design in 1981 to see the way meal is served
has an effect on conversation among peers.
 During 3 Meals (breakfast, lunch and dinner ).

Steps for conducting Single Case Research

 Define research problem
 State research hypothesis
 Define variables operationally
 Define population be studied
 State individuals be selected (if possible) sample size….
 Establish research methodology
 Administer research conditions
 Observe and record appropriate data
 Compare the performance of individual with and without treatment
condition.

Ex post Facto research Design

 It is any systematic empirical inquiry in which Ivs have not been
directly controllable because they have occurred or because they are not
inherently manipulatable .
 It refers to the selection of subjects who have already been exposed
to particular treatment or exhibit particular characteristics. These are
selected for these characteristics after the fact.
 E.g., Pettigrew (1959)
Types

 Prospective Ex Post Facto Designs

 Procedure searches forward in time to see if subjects belonged to a
particular group
 E.g., we can take sample of 100 subjects with high stress and 100
with low stress to see frequency of cancer.
 Retrospective Ex Post Facto Design
 In it organism often come equipped with some particular
characteristics we wish to investigate as a treatment condition. We select
the subjects based on characteristics they have or not.
 E,g., we select a sample of 100 su

 Advantages
 Used in those situations in which IV can not be manipulated due to
physical and social restraints.
 When time and funds are limited.

 Disadvantages
 Inability to assign subjects randomly to different treatment conditions
in an experiment. Other characteristics of subjects are confounding.
(Solution is match the subject on confounding variable).
 Convenient sampling often found…. Issue of external validity.
 Detection Bias (Feinstein, 1977)
 Identification of members based on archival records threats internal
validity

Matching as a solution (selection of confounded

variables along with IV match subjects on these characteristics
Two approaches are frequently used
SUBJECT-FOR SUBJECT
MATCHING
Highly stressed and low DISTRIBUTION-FOR –
stressed person are a good DISTRIBUTION MATCHING
match.
If highly stressed person took 2 Two groups are matched in
pack of cigarettes then low terms of their average
stressed swill be taken of 2 number of characteristics.
pack smoker. There is also issue of
Difficult to find as precise external validity.
matching…drop the subject…
threats to external validity

Measuring variable as solution (search for

confounding variable and make them part of the study by
matching them)

Relative
RELATIVErisk
RISK ratio
RATION

Prospective ex post facto

research is concerned with the RELATIVE ODD RATIO
cause of disease and special
calculations are used to Good estimate of relative
describe the results …relative risk ratio;
risk ratio. Ratio calculated for
Reporting of DV as relative retrospective ex post facto.
risk ratio. Hides absolute risk ration
Hides absolute risk ratio

High STRESS
GROUP LOW STRESS GROUP
100 subjects of high 100 subjects of low stress
stress group group
Frequency of cancer= Frequency of Relative
cancer= 5risk
20 subjects subjects ratio = 20/100
Frequency of no Frequency of ÷
no5/100
cancer=
cancer= 80 95 20/100 ×
5/100= 4

Relative Odd ratio

CANCER PATIENTS GROUP OF SUBJECTS
GROUP WITHOUT CANCER
100 subjects with 100 subjects without cancer
cancer Frequency of low stress and
Frequency of high cancer= 5 subjects
stress in subjects Frequency of no cancer= 95
with cancer= 20
subjects
Frequency of no
cancer= 80
Relative odd ratio
= 20/5 ÷ 80/95
20/5× 80/95= 4.7