8/6/2025

Research Methods for the 3.1 An Introduction to Measurement

Behavioral Sciences, 5e
GRAVETTER | FORZANO
• First step in the research process
– Find an unanswered question that will serve

3 as a research idea
• Second step: form a hypothesis
• Next steps: use the hypothesis
Defining and Measuring – Develop an empirical (observable) research
Variables study that will either support or refute the
hypothesis

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

1 2

Variables 3.2 Constructs and Operational Definitions

• Characteristics or conditions • Types of variables involved in research

– Change or have different values for different – Well-defined, easily observed, and easily
individuals measured
• Evaluating differences or changes in • Examples: height and weight
variables – Intangible, abstract attributes
– Must be able to define and measure the • Examples: motivation or self-esteem
variables being studied • Measurement is more complicated

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

3 4

1
 8/6/2025

Theories and Constructs Operational Definitions

• Theory: a set of statements • Methods of defining and measuring

– About mechanisms underlying a particular constructs
behavior – Variables that cannot be observed or
• What are constructs (hypothetical measured directly
constructs)? • Examples: intelligence can be operationally
defined as a score on an IQ test; and “hunger” can
– Hypothetical entities created from theory and be operationally defined as the number of hours of
speculation food deprivation
• Cannot be seen, but are assumed to exist
• Help explain and predict behavior in a theory

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

5 6

Limitations of Operational Definitions 3.3 Validity and Reliability of Measurement

• An operational definition is not the same • Researchers have developed two general
as the construct itself criteria for evaluating the quality of any
• Concerns about the quality of operational measurement procedure
definitions and measurements produced – Validity and reliability
– It is easy for operational definitions to leave
out important components of a construct
– Operational definitions often include extra
components that are not part of the construct
being measured
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

7 8

2
 8/6/2025

Consistency of a Relationship Relationship Among Variables

• Demonstrating consistency • Determine the consistency of a

– Two scores (one score from the new measure relationship by computing a correlation
and one score from a well-established between the two measures
measure) are represented for each person in – Consistent positive relationship: a correlation
a scatter plot near +1.00
• Positive relationship: the two measurements
– Consistent negative relationship: a correlation
change together in the same direction
near -1.00
• Negative relationship: the two measures change in
opposite directions – Inconsistent relationship: a correlation near
zero

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

9 10

Scatter Plots Showing Different

Validity of Measurement
Relationships
• To establish validity
– Measurement procedure must accurately
capture the variable that it is supposed to
measure
• Example: Does an IQ test measure intelligence
accurately?

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

11 12

3
 8/6/2025

Six Common Definitions of Validity Six Common Definitions of Validity – II

• Face validity: simplest and least scientific • Predictive validity

– Whether a measure superficially appears to – Scores obtained from a measure accurately
measure what it claims to measure predict behavior according to a theory
• Concurrent validity • Construct validity
– Scores obtained from a new measure – Scores obtained from a measurement behave
• Directly related to scores obtained from a more exactly the same as the variable itself
established measure of the same variable

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

13 14

Six Common Definitions of Validity – III Reliability of Measurement

• Convergent validity • Stability or the consistency of the

– Strong relationship between the scores measurements produced by a specific
obtained from two (or more) different methods measurement procedure
of measuring the same construct • Concept of reliability
• Divergent validity – Inconsistency of a measurement comes from
– Showing little or no relationship between the error
measurements of two different constructs – Measured Score = True Score + Error

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

15 16

4
 8/6/2025

Sources of Error Types and Measures of Reliability

• Observer error • Successive measurements

– The individual who makes the measurements – Test-retest reliability compares scores of two
can introduce simple human error successive measurements of the same
• Environmental changes individuals and correlates the scores

– It is difficult to attain the ideal of identical • Simultaneous measurements

circumstances – Inter-rater reliability: agreement between two
• Participant changes observers who simultaneously record
measurements of the behaviors
– The participant can change between
measurements (e.g., degree of focus)
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

17 18

The Relationship Between Reliability and

Types and Measures of Reliability (cont’d.)
Validity
• Internal consistency • Reliability is a prerequisite for validity
– Split-half reliability: splits the test in half – A measurement procedure cannot be valid
computing a separate score for each half, and unless it is reliable
then calculating the degree of consistency • It is not necessary for a measurement to
between the two scores for a group of
participants
be valid for it to be reliable
– Consistency of measurement is no guarantee
of validity

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

19 20

5
 8/6/2025

3.4 Scales of Measurement Four Types of Measurement Scales

• Measurement is a procedure for • Nominal scale

classifying individuals – Represents qualitative differences in the
– Scale of measurement is the set of categories variable measured
used for classification • Ordinal scale
• The process of measurement involves two – Represents differences in a series of ranks
components: • Interval and ratio scales
– Set of categories – Consist of a series of equal intervals like the
– Procedure for assigning individuals to inches on a ruler
categories – A ratio scale has a true zero value
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

21 22

Dealing With Equivocal Measurements Selecting a Scale of Measure

• Interval or ratio scale scores are: • What is the scale’s ability in comparing
– Compatible with basic arithmetic different measurements?
– Suitable for statistical analysis (e.g., mean – Nominal: reveals whether a difference exists
and variance) ► allow hypothesis testing – Ordinal: indicate the direction of the difference
• Scores from an ordinal scale are not: (which is more and which is less)
– Compatible with basic arithmetic or – Interval: determines the direction and the
appropriate for statistical analysis magnitude of a difference

• Distinction of a measure as an interval and • Affects the ability to describe relationships

ratio is not always clear cut between variables
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

23 24

6
 8/6/2025

3.5 Modalities of Measurement Modalities of Measurement (cont’d.)

• Self-report measures ask participants • Behavioral measures are based on overt

direct questions to study behavior behaviors that can be observed and
– Provides a direct measure; validity is dubious measured
• Physiological measures are based on – Provide researchers with a vast number of
manifestations of the underlying construct options
• Possible to select the behavior(s) that seems to be
– Ex.: galvanic skin response (GSR) and MRIs best for defining and measuring the construct
– Provide objective measures – Behaviors may be only a temporary or
– Require expensive equipment; may not situational indicators of an underlying
provide valid measures of constructs construct
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

25 26

3.6 Other Aspects of Measurement Sensitivity and Range Effects

• Multiple measures provide more • Range effect: a measurement that is not

confidence in the validity of the sensitive enough to detect a difference
measurements – Ceiling effect: clustering of scores at the high
– Can introduce problems end of a measurement scale
• Complex statistical analysis and interpretation • Allowing little or no possibility of increases in value
• Desynchrony: lack of agreement between two – Floor effect: clustering of scores at the low
measures end of a measurement scale
– Method for limiting problems • Allowing little or no possibility of decreases in
• Combine the measures into a single score or each value
individual
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

27 28

7
 8/6/2025

Artifacts: Experimenter Bias and

Limiting Experimenter Bias
Participant Reactivity
• Artifact: a nonnatural feature accidentally • Standardize or automate the experiment
introduced into something being observed • Single-blind study
– Experimenter bias: measurements are – The research study is conducted by an
influenced by the experimenter’s expectations experimenter who does not know the
regarding the outcome of the study expected results
• Double-blind study
– Neither the experimenter nor the participants
know the expected results of the study

© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

29 30

Results Distorted by Actions and

Study Subjects May Adopt Roles
Responses
• Demand characteristics refer to any of the • Good subject role
potential cues or features of a study that: – Supports the experimenter’s hypothesis
– Suggest to the participants the purpose and • Negativistic subject role
hypothesis of the study, and
– Acts contrary to the hypothesis
– Influence the participants to respond or
behave in a certain way • Apprehensive subject role
– Presents himself in a “good light”
• Reactivity
– Participants modify their natural behavior in • Faithful subject role
response to knowing they are in a study – Follows instructions to the letter (ideal
participant)
© 2016 Cengage Learning. All Rights Reserved. © 2016 Cengage Learning. All Rights Reserved.

31 32