Research Methods in
I/O
Psychology
Chapter
2
What is Science?
 A process or method for generating a body
of knowledge
 Represents a logic of inquiry
 Primary objective is theory building
 Is psychology a science?
 Yes! It relies on formal, systematic observation
to answer questions about behavior
Goals of Science
 Description
 Accurate portrayal/depiction of
phenomenon
 Explanation
 Gathering info about “why” phenomenon
exists
 Prediction
 Anticipate an event prior to its occurrence
 Control
 Manipulation of conditions to affect
behavior
Assumptions of Science
Empiricis
m
Determinis
m
Discoverabilit
y
What Is Theory?
 Interrelated constructs (concepts),
definitions, and propositions that present
a systematic view of a phenomenon by
specifying relations among variables, with
the purpose of explaining and predicting
the phenomenon
 What makes a theory good?
What Makes a Good Theory?
 Parsimonious
 Explains a lot, yet simple
 Precision
 Specific and accurate in its
wording
 Testability
 Verifiable by
experimentation/study
What Makes a Good Theory?
(continued)
 Useful
 Practical, helpful in describing, explaining,
and predicting important phenomenon
 Generativity
 Stimulates additional research
 Efficient
 Produces value and saves resources
Is the best theory the one that
is most right?
 All theories (or models) are wrong!!!
 But some are useful
 Theories are “simplifications” or
“abstractions”
 They are important because they allow us
to do work…
 Not trying to build an exact “replication”
but rather an efficient “model”…
Cyclical Inductive–Deductive
Model of Research
 Induction: data  theory
 Deduction: theory  data
 Most research is driven by the
deductive process, but it is also
possible to start with data
 There is no perfect way to “do
science”
Cyclical Inductive–Deductive
Model of Research
Research Ethics
 Ethical Principles of Psychologists and Code
of Conduct of the APA
 Research must be approved by the
supporting institution (e.g., university);
Institutional Review Board (IRB)
 Informed consent: providing study info,
right to decline, risks/benefits, questions
 Confidentiality of participant data
 Avoiding the use of deception
 Requirements for care/use of animals
Nazi Experiments
 Nazi scientists and physicians performed
many horrific experiments on humans
without consent. For instance, they
inflicted wounds and infections to observe
the effects. They also administered
experimental drugs without permission.
Many of these experiments were made
public in the Nuremberg trials (1945-
1949).
Research Terminology and
Basic Concepts: Overview
 Causal inference can be made when
data indicate that a causal relationship
between two variables is likely
 Can never prove causality due to
other variables
 Key Terms:
 Independent variable (IV)
 Dependent variable (DV)
 Extraneous variable
Variables
 Independent variable: anything that
is systematically manipulated
 Predictor, precursor, or antecedent
 Dependent variable: what an experiment
is designed to assess
 Criteria, outcome, or consequence
 Extraneous variable: any other variable
that can contaminate results
Internal and External Validity
 Internal validity: extent to which
causal inferences can be drawn about
variables
 Ruling out alternative explanations
 External validity: extent to which
results generalize to other people,
settings, time
 Student participants and “real world”
applicability
*Important trade-off between internal
Control
 Important to ensure that a causal
inference can be made about the effect
of the IV on the DV
 Ways to control:
 Hold extraneous variables constant
 Systematically manipulate different levels of
extraneous variables (make part of
experimental design)
 Statistical control
Stage Model of the
Research Process
Formulate the hypothesis
Design the study
Collect data
Analyze the data
Report the findings
Types of Research Designs:
Overview
 Experimental methods
 Lab experiments
 Field and quasi-
experiments
 Observational methods
Types of Research Designs:
Experimental Methods
Two factors characterize experimental
methods:
 Random Assignment
 Each participant has an equally likely chance
of being assigned to each condition
 Manipulation
 Systematic control of an independent
variable
*These two techniques increase internal validity
Types of Research Designs:
Lab Experiments
 Random assignment and manipulation
of IVs are used to increase control
 Take place in a contrived setting for
control
 Very high internal validity
 External validity is questioned
Types of Research Designs:
Field and Quasi-Experiments
 Field experiments
 Random assignment and manipulation
in a realistic field setting
 Quasi-experiments (very common in
I/O)
 Field experiment w/o random
assignment
 Not always practical to randomly
assign participants; use of intact
groups
Types of Research Designs:
Observational Methods
 Observational methods: also called
correlational designs, descriptive research
 Do not involve random assignment
or manipulation
 Make use of available resources
 Can draw conclusions about relationships
but NOT causality
 Common in field settings
Data Collection Techniques:
Overview
 Naturalistic
observation
 Case studies
 Archival research
 Surveys
Data Collection Techniques:
Naturalistic Observation
 Observation of someone or something in
its natural environment
 Participant observation – observer tries to
“blend in” completely with those who are
observed
 Unobtrusive observation – observer
objectively observes individuals without
drawing attention to him/herself; does not
blend in completely
Data Collection Techniques:
Case Studies
 Examination of single individuals,
groups, companies or societies
 Main purpose is description; explanation is
also a reasonable goal
 Not typically used to test hypotheses
 Provide details about a typical or exceptional
firm or individual
Data Collection Techniques:
Archival Research
 Answering a research question using
existing (secondary) data sets
 Lack of control over quality of data is a
concern
 Minimizes time developing measures
and collecting data
 Include both:
 Cross-sectional data: one point in time
 Longitudinal data: multiple time periods
Data Collection Techniques:
Surveys
 Selecting a sample of respondents
and administering a questionnaire
 Most frequently used method of data
collection in I/O
 Two approaches:
 Self-administered
 Interviews
Data Collection Techniques:
Surveys
 Self-administered questionnaires
 Completed by respondents in absence
of researcher
 Used in both lab and field settings
 Useful for 3 reasons:
 Ease of administration
 Can be administered to large groups at
one time
 Provides respondents with anonymity
Data Collection Techniques:
Surveys
 Self-administered questionnaires
(continued)
 Drawbacks of using surveys
 Low response rates (mail surveys)
 Difficult for respondents to ask
questions
Data Collection Techniques:
Surveys
 Interviews (Investigator-administered)
 Usually conducted face-to-face, can be done
over the phone
 More time consuming than self-
administered questionnaires
 Clear benefits:
 Higher response rates
 Ambiguity about questions can be resolved
Data Collection Techniques:
Surveys
 Technological Advances
 Web-based surveys – valuable alternative
 Experience sampling methodology (ESM)
 Captures momentary attitudes and
psychological
states
 PDA “signals” participants to answer questions
at a predetermined time
 Popular for the study of emotions at work
Measurement: Overview
 Measurement – assignment of numbers
to objects or events using rules in a way
that represents specified attributes of
the objects
 Attribute – dimension along which
individuals can be measured and
along which they vary
Measurement: Overview
 Because accuracy of measurement is
important there are two major
concerns:
 Reliability
 Validity
Reliability and Validity
 Accuracy is similar to Validity
 Precision is similar to
Reliability
Reliability
 The consistency or stability of a measure
 Predictors must be measured reliably
 Measurement error renders
measurement inaccurate or unreliable
 We cannot predict outcomes with variables
that are not measured well
Validity
Summary of Reliability Types and
Approaches to Construct Validity
Statistics: Overview
 Statistic: summarizes in a single number
the values, characteristics, or scores
describing a series of cases
 Measures of central tendency
 Measures of dispersion
 Shapes of distributions
 Correlation and regression
 Meta-analysis
Statistics:
Measures of Central Tendency
 Characterize a typical member of the
group
 Mode: most frequent single score
in a distribution
 Useful with categorical data
 Median: score in the middle of a
distribution
 Extreme scores do not affect the
median
 Mean: arithmetic average of a group
of scores
Statistics:
Measures of Dispersion
 Inform us how closely scores are
grouped around the measure of
central tendency; “spread-outedness”
of the data
 Range: spread of scores from the
lowest to the highest
 Variance: most useful measure
of dispersion
 Standard deviation: square root of
the variance; retains original metric
of score
Statistics:
Shapes of Distributions
 Normal Distribution
 Bell-shaped curve
 Most observations are clustered around the
mean with fewer outside in either direction
 Many psychological variables are
distributed this way (e.g., job attitudes,
performance, intelligence)
 Normal distribution can be used to
calculate percentile scores: individual
score relative to the population
Statistics:
Correlation and Regression
 Correlation coefficient (r) : strength of
the relationship between two variables
 Provides information about the direction
and the
magnitude of the relationship
 Direction can be positive (elevator) or
negative (teeter totter)
 Magnitude: 0 to 1.00
Correlation Examples
The Correlation Coefficient (r)
 …indicates how much two
variables covary or “go together”
(regardless of units of
measurement!).
Examples…
 “Need to achieve” & GPA
 Anger & heart disease
 Psychopathy & hockey penalties
 Self-esteem & body weight
Effect Size
 Correlation between being married and satisfaction with
life?
 .10
 Correlation between being in a good mood and being
willing to help others (i.e., give time or money)
 .30
 Correlation between being in a bad mood and
being perceived as aggressive (in lab interactions
and team exercises)
 .50
 r = 1 
perfect
covariation
(higher x 
higher y)
r = 0  no
covariation (higher x
tells nothing about y)
 r = -1  perfect
negative covariation
(higher x  lower y)
Statistics: Regression
 Allows us to predict one variable
from another
 How much variance in a criterion
variable is accounted for by a predictor
variable
 Coefficient of determination (r2):
percentage of variance accounted for by
the predictor
Coefficient of Determination
Meta-Analysis
 Methodology used to conduct
quantitative literature reviews
 Previously only narrative reviews
were conducted
 Used to combine the results of multiple
studies to arrive at the best estimate of
the true relationship

Research Methods in Organizational Psychology pptx

  • 1.
  • 2.
    What is Science? A process or method for generating a body of knowledge  Represents a logic of inquiry  Primary objective is theory building  Is psychology a science?  Yes! It relies on formal, systematic observation to answer questions about behavior
  • 3.
    Goals of Science Description  Accurate portrayal/depiction of phenomenon  Explanation  Gathering info about “why” phenomenon exists  Prediction  Anticipate an event prior to its occurrence  Control  Manipulation of conditions to affect behavior
  • 4.
  • 5.
    What Is Theory? Interrelated constructs (concepts), definitions, and propositions that present a systematic view of a phenomenon by specifying relations among variables, with the purpose of explaining and predicting the phenomenon  What makes a theory good?
  • 6.
    What Makes aGood Theory?  Parsimonious  Explains a lot, yet simple  Precision  Specific and accurate in its wording  Testability  Verifiable by experimentation/study
  • 7.
    What Makes aGood Theory? (continued)  Useful  Practical, helpful in describing, explaining, and predicting important phenomenon  Generativity  Stimulates additional research  Efficient  Produces value and saves resources
  • 8.
    Is the besttheory the one that is most right?  All theories (or models) are wrong!!!  But some are useful  Theories are “simplifications” or “abstractions”  They are important because they allow us to do work…  Not trying to build an exact “replication” but rather an efficient “model”…
  • 9.
    Cyclical Inductive–Deductive Model ofResearch  Induction: data  theory  Deduction: theory  data  Most research is driven by the deductive process, but it is also possible to start with data  There is no perfect way to “do science”
  • 10.
  • 11.
    Research Ethics  EthicalPrinciples of Psychologists and Code of Conduct of the APA  Research must be approved by the supporting institution (e.g., university); Institutional Review Board (IRB)  Informed consent: providing study info, right to decline, risks/benefits, questions  Confidentiality of participant data  Avoiding the use of deception  Requirements for care/use of animals
  • 12.
    Nazi Experiments  Naziscientists and physicians performed many horrific experiments on humans without consent. For instance, they inflicted wounds and infections to observe the effects. They also administered experimental drugs without permission. Many of these experiments were made public in the Nuremberg trials (1945- 1949).
  • 13.
    Research Terminology and BasicConcepts: Overview  Causal inference can be made when data indicate that a causal relationship between two variables is likely  Can never prove causality due to other variables  Key Terms:  Independent variable (IV)  Dependent variable (DV)  Extraneous variable
  • 14.
    Variables  Independent variable:anything that is systematically manipulated  Predictor, precursor, or antecedent  Dependent variable: what an experiment is designed to assess  Criteria, outcome, or consequence  Extraneous variable: any other variable that can contaminate results
  • 15.
    Internal and ExternalValidity  Internal validity: extent to which causal inferences can be drawn about variables  Ruling out alternative explanations  External validity: extent to which results generalize to other people, settings, time  Student participants and “real world” applicability *Important trade-off between internal
  • 16.
    Control  Important toensure that a causal inference can be made about the effect of the IV on the DV  Ways to control:  Hold extraneous variables constant  Systematically manipulate different levels of extraneous variables (make part of experimental design)  Statistical control
  • 17.
    Stage Model ofthe Research Process Formulate the hypothesis Design the study Collect data Analyze the data Report the findings
  • 18.
    Types of ResearchDesigns: Overview  Experimental methods  Lab experiments  Field and quasi- experiments  Observational methods
  • 19.
    Types of ResearchDesigns: Experimental Methods Two factors characterize experimental methods:  Random Assignment  Each participant has an equally likely chance of being assigned to each condition  Manipulation  Systematic control of an independent variable *These two techniques increase internal validity
  • 20.
    Types of ResearchDesigns: Lab Experiments  Random assignment and manipulation of IVs are used to increase control  Take place in a contrived setting for control  Very high internal validity  External validity is questioned
  • 21.
    Types of ResearchDesigns: Field and Quasi-Experiments  Field experiments  Random assignment and manipulation in a realistic field setting  Quasi-experiments (very common in I/O)  Field experiment w/o random assignment  Not always practical to randomly assign participants; use of intact groups
  • 22.
    Types of ResearchDesigns: Observational Methods  Observational methods: also called correlational designs, descriptive research  Do not involve random assignment or manipulation  Make use of available resources  Can draw conclusions about relationships but NOT causality  Common in field settings
  • 23.
    Data Collection Techniques: Overview Naturalistic observation  Case studies  Archival research  Surveys
  • 24.
    Data Collection Techniques: NaturalisticObservation  Observation of someone or something in its natural environment  Participant observation – observer tries to “blend in” completely with those who are observed  Unobtrusive observation – observer objectively observes individuals without drawing attention to him/herself; does not blend in completely
  • 25.
    Data Collection Techniques: CaseStudies  Examination of single individuals, groups, companies or societies  Main purpose is description; explanation is also a reasonable goal  Not typically used to test hypotheses  Provide details about a typical or exceptional firm or individual
  • 26.
    Data Collection Techniques: ArchivalResearch  Answering a research question using existing (secondary) data sets  Lack of control over quality of data is a concern  Minimizes time developing measures and collecting data  Include both:  Cross-sectional data: one point in time  Longitudinal data: multiple time periods
  • 27.
    Data Collection Techniques: Surveys Selecting a sample of respondents and administering a questionnaire  Most frequently used method of data collection in I/O  Two approaches:  Self-administered  Interviews
  • 28.
    Data Collection Techniques: Surveys Self-administered questionnaires  Completed by respondents in absence of researcher  Used in both lab and field settings  Useful for 3 reasons:  Ease of administration  Can be administered to large groups at one time  Provides respondents with anonymity
  • 29.
    Data Collection Techniques: Surveys Self-administered questionnaires (continued)  Drawbacks of using surveys  Low response rates (mail surveys)  Difficult for respondents to ask questions
  • 30.
    Data Collection Techniques: Surveys Interviews (Investigator-administered)  Usually conducted face-to-face, can be done over the phone  More time consuming than self- administered questionnaires  Clear benefits:  Higher response rates  Ambiguity about questions can be resolved
  • 31.
    Data Collection Techniques: Surveys Technological Advances  Web-based surveys – valuable alternative  Experience sampling methodology (ESM)  Captures momentary attitudes and psychological states  PDA “signals” participants to answer questions at a predetermined time  Popular for the study of emotions at work
  • 32.
    Measurement: Overview  Measurement– assignment of numbers to objects or events using rules in a way that represents specified attributes of the objects  Attribute – dimension along which individuals can be measured and along which they vary
  • 33.
    Measurement: Overview  Becauseaccuracy of measurement is important there are two major concerns:  Reliability  Validity
  • 34.
    Reliability and Validity Accuracy is similar to Validity  Precision is similar to Reliability
  • 35.
    Reliability  The consistencyor stability of a measure  Predictors must be measured reliably  Measurement error renders measurement inaccurate or unreliable  We cannot predict outcomes with variables that are not measured well
  • 36.
  • 37.
    Summary of ReliabilityTypes and Approaches to Construct Validity
  • 38.
    Statistics: Overview  Statistic:summarizes in a single number the values, characteristics, or scores describing a series of cases  Measures of central tendency  Measures of dispersion  Shapes of distributions  Correlation and regression  Meta-analysis
  • 39.
    Statistics: Measures of CentralTendency  Characterize a typical member of the group  Mode: most frequent single score in a distribution  Useful with categorical data  Median: score in the middle of a distribution  Extreme scores do not affect the median  Mean: arithmetic average of a group of scores
  • 40.
    Statistics: Measures of Dispersion Inform us how closely scores are grouped around the measure of central tendency; “spread-outedness” of the data  Range: spread of scores from the lowest to the highest  Variance: most useful measure of dispersion  Standard deviation: square root of the variance; retains original metric of score
  • 41.
    Statistics: Shapes of Distributions Normal Distribution  Bell-shaped curve  Most observations are clustered around the mean with fewer outside in either direction  Many psychological variables are distributed this way (e.g., job attitudes, performance, intelligence)  Normal distribution can be used to calculate percentile scores: individual score relative to the population
  • 42.
    Statistics: Correlation and Regression Correlation coefficient (r) : strength of the relationship between two variables  Provides information about the direction and the magnitude of the relationship  Direction can be positive (elevator) or negative (teeter totter)  Magnitude: 0 to 1.00
  • 43.
  • 44.
    The Correlation Coefficient(r)  …indicates how much two variables covary or “go together” (regardless of units of measurement!). Examples…  “Need to achieve” & GPA  Anger & heart disease  Psychopathy & hockey penalties  Self-esteem & body weight
  • 45.
    Effect Size  Correlationbetween being married and satisfaction with life?  .10  Correlation between being in a good mood and being willing to help others (i.e., give time or money)  .30  Correlation between being in a bad mood and being perceived as aggressive (in lab interactions and team exercises)  .50
  • 46.
     r =1  perfect covariation (higher x  higher y) r = 0  no covariation (higher x tells nothing about y)  r = -1  perfect negative covariation (higher x  lower y)
  • 48.
    Statistics: Regression  Allowsus to predict one variable from another  How much variance in a criterion variable is accounted for by a predictor variable  Coefficient of determination (r2): percentage of variance accounted for by the predictor
  • 49.
  • 50.
    Meta-Analysis  Methodology usedto conduct quantitative literature reviews  Previously only narrative reviews were conducted  Used to combine the results of multiple studies to arrive at the best estimate of the true relationship