EPSY 8334 Class 1 and 2 (Spring 2015) Heppner et al 3 -4 Pyrczak & Bruce 1&4.
Post on 19-Jan-2016
215 Views
Preview:
Transcript
EPSY 8334
Class 1 and 2 (Spring 2015)Heppner et al 3 -4
Pyrczak & Bruce 1&4
Science in Counseling Psychology
Are our interventions effective? Empirically Validated Treatments
Experimental studies- randomized clinical trials
Relations about psychological phenomena
2
Class Goals Identify a research question Develop a research proposal that includes
An introduction/literature review that provides conceptual framework for research question based on theoretical/empirical literature and/or reasoning
Methods to conduct study Participants Instruments to collect the data Procedures Analyses
3
Class Goals Demonstrate skills necessary to
understand results sections of journals correlation, T-Test, ANOVAS MANOVAS
Regression, moderation/mediation analyses
Ability to critique articles that report results of randomized clinical trials of psychotherapy efficacy (Goal 2 Competency B.1a)
4
Conceptualizing a Research Question
Three steps What do we know about the issue.What we do not know about the issue.How will the study bridge the gap.How what we know leads to the
questions in my study
5
6
Identify Research Topics 4/8
1. Read, Read, Read Think…..
2. Identify/discuss ideas
3. Build on previous research
4. Identify research questions (Kiplinger
& Lee, 2001)
7
Identify Research Topics 4/8
Identify:
5. Measurable constructs
6. Instruments to assess constructs
7. Participants
8. Statistical analyses
Literature: Replication vs. Guide
Use different operational definition of variable Add moderator or control variables Develop different way to test hypotheses Combine questions/strategies from several
studies Examine similar questions with a different
population Identify gaps weaknesses in previous work
that you will address in your study
8
How process unfolds: Describe research interests in general, broad
terms. Identify testable research question(s) Identify the variables/constructs embedded in
the question Provide conceptual definitions for variables. Provide operational definitions for variables
How to assess/measure the construct Observations, Self-Report
9
10
Assumptions Research Paradigms
Ontology
Epistemology
Methodology
10
11
Assumptions Research Paradigms
Ontology
Epistemology
Methodology
Branch of metaphysics that deals with the nature of being & reality
Branch of philosophy that investigates the origin, nature, methods, and limits of human knowledge
Relationship of Knower to Known
Procedures to gain knowledge about the world (reality – others)
11
12
Assumptions Paradigms
Ontology
Epistemology
Methodology
Positivism
Post-Positivism
Constructivism
Critical Theory
12
13
True or False??
1. The scientific method allows us to observe facts without bias.
2. Observations of psychological states measured with scores on an instrument always will include “true” and “error” measurement.
13
14
True or False??
3. In research, the observed cannot be separated from the observer.
4. There is privileged and non-privileged knowledge.
14
PositivismOntology
Epistemology
Methodology
15
PositivismOntology Truth is universal, absolute; it can be
grasped/discovered
Human nature follows cause-effect laws
Epistemology Discover universal laws that govern nature and human behavior
Theories are reductionist
Methodology Scientific Method
Scientific method allows individuals to observe facts without bias
The observer is a “nuisance”16
Post- PositivismOntology Universal truths exist
Human nature follows laws
Nature of universe cannot be known
Probabilistic statements about truth: p value
Epistemology
Observations always = “truth” + “error”
Successive studies to approximate truth
Methodo
logy
Scientific Method
Observations = true + error measurement
Researchers cannot be 100% objective
Scientific community arbitrates knowledge
17
ConstructivismOntology Truth - purely objective reality does not exist
“Reality” & knowledge are socially constructed
Meanings given to an experience are more important than the experience itself
Epistemology
Knower and known cannot be separated
Knowledge is based on interpretations of observed interactions within a context
Methodo
logy
Hermeneutics – Interpretations- Qualitative
Dialectics – relation of investigator and participant- to understand constructions
18
Constructivism--Critical TheoryOntology A purely objective reality does not exist
Knowledge is socially constructed – shaped by those in power
Epistemology
Knower and known cannot be separated
Observer’s location in social context biases observations
Methodology
Hermeneutics - Interpretations
Dialectics – relation of investigator and participant- to change constructions
Social Change --- Social Justice in CP
19
20
True or False??
1. The scientific method allows us to observe facts without bias.
2. Observations of psychological states measured with scores on a test always will include “true” and “error” measurement.
21
True or False??
3. The observed cannot be separated from the observer.
4. There is privileged and non-privileged knowledge.
2005 Nobel Prize for Physiology or Medicine: Australia Dr. Barry J. Marshall (gastroenterologist) and Dr. J. Robin Warren (pathologist)
2005
1984
bacterium Helicobacter pylori
•Causes stomach inflammation, ulcers and cancers (80-90% of ulcers)•Paradigm: stress gastric acid ulcers•Old Treatment: drugs to block gastric acid•Pharmaceutical Industry: Tagamet and Zantac were biggest profit earners.•MDs: Stm. Ulcers was a chronic disease
22
Goals of Science Advance knowledge Make discoveries Increase our understanding of human
behavior, and Acquire facts about counseling (behavior)
Science vs. Practice-- ??
23
Science Practice
Knowledge for its own sake
Academic freedom
Seeks to understand populations: probability
Individual differences: nuisance/confound
Knowledge for what it can accomplish
Public service
Seeks to understand individuals – 0 or 100%
Individual differences: focus of attention
24
25
Testable Research Question (Kerlinger & Lee, 2000)
(a) Asks a Question
(b) About ……
(c) That …………
25
26
Testable Research Question
(a) Asks a Question
(b) About the relation of two or more constructs (correlation/experimental)
(c) That can be measured in some way
26
Types of Research Questions Descriptive
Describe what a phenomenon is like: Collect/Categorize Information- Survey Designs
Difference Compare between two or more groups
Experimental: manipulation of IV; random assignment to groups
Non-experimental: no manipulation of IV; non- random assignment to groups
Correlation (relation) Extent to which 2 or more continuous variables
vary together 27
28
Research Questions (which analyses?)
Descriptive, Difference, Correlation
1. What are the levels of anxiety and career indecision reported by college men and women?
1. Do college women report higher levels of anxiety than college men?
2. Do college women report higher levels of career indecision than college men?
Research Questions (which analyses?)
Descriptive, Difference, Correlation
4. Among college students, are anxiety scores related to career indecision scores?
5. Do college students diagnosed with an anxiety disorder report higher levels of career indecision than college students without an anxiety disorder ?
29
30
Research Questions Descriptive, Difference, Correlation
6. Among college students, what is the combined and unique contribution of anxiety, self-esteem and optimism to career indecision?
7. Among college students, are there gender differences in the relation of anxiety to career indecision?
Moderation Question
7. Among college students, are there gender differences in the relation of anxiety to career indecision?
Among men, is anxiety related to career indecision? Is the relation + or -?
Among women, is anxiety related to career indecision? Is the relation + or -?
31
Research Questions Descriptive, Difference, Correlation
8. Is career indecision –CDS-- a multifactorial or a one-dimensional construct?
32
Report to
Types of Variables
Categorical Discrete groups:
gender, diagnosis (depressed vs. not depressed)
Continuous Ratio Scale (equal intervals)
money, # of days, test score (0-100) Continuous Interval
Likert Scales
33
Types of Research Questions Descriptive
Describe what a phenomenon is like Collect/Categorize Information - Typically used in SurveysRequires large samples that are
representative of the population of interestDo not involve generation of hypothesisNo use of inferential statistics
34
Analyses: Descriptive Collect/Categorize Information - Survey
Means, SD Mean score in standardized test by groups Mean score in opinion survey by groups
Proportions of populations/groups Percentage of high school graduates who attend
college, by ethnic group, Percentage of college students who graduate in 6
years or less, by gender
35
Types of Research Questions Difference
Compare Means Between – groups Descriptive vs. Hypothesis testing purpose
Independent variable: categorical Experimental orStatus (cannot be manipulated)
Dependent Variable: continuous
36
Stat. Analyses: Difference Compare means in continuous variables
between two or more groups: t-test, ANOVA, MANCOVA-Experimental- IV is categorical and
manipulated Stauts - DV is categorical and not
manipulated
The IV is always continuous
37
Difference questions vs.Descriptive Questions?
Difference questions for descriptive purpose do not require rational/hypothesisEg. Proportion of men and women who favor
the republican vs. the democrat nominee
Difference questions for hypothesis testing require rational for expected differences
38
Analyses: Difference African American depressed adults who
receive culturally infused CBT will obtain lower post-test scores in the Beck Depression Inventory(BDI) than their counterparts who receive traditional CBT.
Experimental or Non-Experimental ?
IV DV
39
Analyses: Difference African American depressed adults who
receive culturally infused CBT will obtain lower post-test depression scores (Beck Depression Inventory-BDI) than their counterparts who receive traditional CBT.
Experimental or Non-Experimental ? Random assignment to treatment groups
IV Treatment Groups: CICBT --- TCBT DV Scores in BDI
40
Analyses: Difference - T-Test
Treatment(Random assignment)
BDI Post test Mean SD
Culturally Infused CBT
10.3* 5.3
Traditional CBT 15.1 6.2
*p<.05; Possible range of scores BDI = 1-20
41
Analyses: Difference Women will obtain higher BDI scores at
post-test than men (regardless of treatment group)
Experimental or Non-Experimental ? IV DV
42
Analyses: Difference Women will obtain higher BDI scores at
post-test than men (regardless of treatment group)
Experimental or Non-Experimental ? Gender – status variable, no random assignment
IV Gender DV BDI scores
43
Analyses: Difference - T-Test
Client’s Gender
BDI Post-Test Mean SD
Men 11.5ns 5.2
Women 12.3 6.5
Possible range of scores BDI = 1-20
44
2X2 ANOVA: Treatment and Client Gender Differences in Depression
IVsDV
Post Test BDI Scores F statistic
Treatment T I vs. T2 (CI-CBT vs. Trad CBT)
Main Effect*
Client Gender Male vs. Female Main Effect
Treatment x Gender
MeanT1/Male MeanT2/Male T1/Female T2/Female
Interaction Effect*
45
Does gender Moderate the treatment effects?
Gender
BDI Post-Test CI CBT
BDI Post-Test Trad-CBT
Male 11.8* 16.5
Female 12.5ns 13.1
*p<.05; Possible range of BDI = 0-20
46
2X2 ANOVA: Treatment and Client Gender Differences in Depression
Independent Variable
Parenting Rating Scale
Treatment CI-CBT Trad-CBT
10.3* 15.1
C Gender Male Female
11.5 12.3
Treat. X Gen* Sig
Treatment X Cl Gender Interaction
Cl Gender Treatment
CI-CBT Trad.-CBT
Male 11.8* 16.5
Female 12.5 13.1
* p<.05
47
Analyses: Difference Mothers who are diagnosed with clinical
depression (BDI>16) will obtain lower scores in parenting skills (Parenting Skills Rating Scale: PSRS) than mothers who are not clinically depressed (BDI<16).
Experimental or Non-Experimental? IV DV
48
Analyses: Difference Mothers who are diagnosed with clinical
depression (BDI>16) will obtain lower scores in parenting skills (Parenting Skills Rating Scale: PSRS) than Mothers who are not clinically depressed (BDI<16). .
Experimental or Non-Experimental?
Status variables- no randomnization IV Depression group: Yes vs. No DV Score in Parenting SRS
49
Analyses: Difference - T-Test
Clinically Depressed
Parenting Skills Rating Scale Mean SD
Yes 2.3* 1.4
No 3.8 1 .1
*p<.05; Possible range of PSRS = 1-5
50
Analyses Differences: 2 IVs Are there differences in scores in the Parenting
Skills Scale among mothers (a) who are depressed or non depressed (b) and who have a partner or not?
Does partner status moderate (makes a difference in) the association of mothers’ depression diagnosis to scores on parenting skills ratings?
IVs DV
51
Analyses Differences: 2 IVs Are there differences in scores in the Parenting
Skills Scale among mothers (a) who are depressed or non depressed (b) and who have a partner or not?
Does partner status moderate the relation of
mothers’ depression status to scores on parenting skills ratings?
IVs partner status (Mod)- depression diagnosis DV parenting skills
52
2X2 ANOVA: Depression Diagnosis and Partner Status in Parenting Skills (n=100)
IVsDV
Parenting Skills F statistic
Depression Status (n=100)
Yes vs No (n=50 n=50)
Main Effect*
Partner Status (n=100)
Partner vs No Partner (n=50 n=50)
Main Effect NS
Depression x Partner Status
Partner ---- DY vs DN
No Partner ---- DY vs DN
Interaction EffectNS
53
Depression Diagnosis, Marital Status and Parenting Skills m
Independent Vs. Parenting Rating Scale
Depression Yes No
2.3* 3.8
Marital Status Partner No Partner
3.1 2.9
Dep X Partner S NS
Depression X Marital Status Interaction
Partner S Mom Depression
Yes No
Partner 2.5* 3.6
No Partner 2.4* 3.4
* p<.05
Types of Research Questions• Correlation
– Extent to which 2 or more continuous psychological constructs vary together
– Pearson correlation coefficient r- bivariate correlation: strength and direction
– Regression - multi IV’s• test combined contribution of several Ivs to one DV• test unique contribution of each IV to the DV while
controlling for the other IVs
55
56
External and Internal Validity
E I
Experimental
Field
E i
Correlational/
Descriptive
Field
e I
Experimental
Laboratory
e i Correlational/
Descriptive
Laboratory
Clinical Trials
57
Fill in the Blanks
The best research design is:
What type of design is used in the randomized controlled clinical trials used to examine therapy outcome?
top related