Chapter 3

Chapter 3: Defining and Measuring Variables

Introduction to Variables

• Variables: Characteristics or conditions that can change and differ among individuals.

  • Well-defined Variables: Easily observed and measured (e.g., height, weight).

  • Abstract Variables: Intangible and more complex (e.g., motivation, self-esteem) require more sophisticated measurement techniques.

Types of Variables

• Independent Variable (IV): Factor manipulated by the experimenter to observe its impact.

  • Levels: Different values or conditions of the independent variable, at least two required.

• Dependent Variable (DV): Measured outcome reflecting the effect of the independent variable.

• Subject Variables: Characteristics inherent to subjects that influence outcomes.

• Confounded Variables: Variables that can interfere with the relationship between IV and DV.

• Extraneous Variables: Other variables that could affect the outcome if not controlled.

Measuring Behavior

• Frequency: Count of how often a behavior occurs.

• Rate: Frequency of behavior relative to a specified time frame.

• Duration: Length of time a behavior is observed.

• Latency: Time taken from instruction to performance.

• Topography: The specific form or pattern of behavior.

• Force: Intensity or strength of the behavior.

• Locus: The specific location in the environment where the behavior takes place.

Measurement Examples

• To measure different behaviors (like kicking furniture or writing), the appropriate measuring methods from frequency, rate, duration, etc., must be selected.

Theories and Constructs

• Theory: Set of principles explaining mechanisms behind behaviors; includes concepts not directly observed.

• Constructs: Hypothetical entities derived from theories; cannot be observed but help to predict behaviors (e.g., intelligence, hunger).

Operational Definitions

• Defines and measures constructs through observable behavior:

  • Specify measurement procedures for external behaviors.

  • Examples: intelligence may be operationally defined as IQ scores; hunger may be measured by hours of food deprivation.

Limitations of Operational Definitions

• Operational definitions do not equate to the constructs themselves.

  • Poorly defined operational definitions may omit vital aspects or include irrelevant components.

General Criteria for Evaluating Measurement

• Two key criteria: Validity and Reliability.

• Good measurements are consistent across different testing scenarios.

Consistency of a Relationship

• Relationship analysis involves plotting scores to observe correlations:

  • Positive Relationship: Scores move in the same direction.

  • Negative Relationship: Scores move in opposite directions.

Correlation and Consistency

• Correlation calculations help assess the strength and direction of relationships:

  • Values near +1.00 indicate a strong positive relationship.

  • Values near -1.00 indicate a strong negative relationship.

  • A value near 0 indicates no discernible relationship.

Validity of Measurement

• Measurement must accurately reflect what it intends to measure.

  • Face Validity: Superficial appearance of measurement.

  • Concurrent Validity: Correlation with established measures.

  • Predictive Validity: Ability to predict future behaviors.

  • Construct Validity: Alignment with theories.

  • Convergent Validity: Agreement across different measurement methods.

  • Divergent Validity: Lack of correlation between different constructs.

Reliability of Measurement

• Consistency of measurements over time:

  • Measured score = True score + Error.

  • Sources of Error:

    • Observer Error: Human error in measurement.

    • Environmental Changes: Variability in conditions.

    • Participant Changes: Variations in participant states.

Types and Measures of Reliability

• Test-Retest Reliability: Consistency across successive measurements.

• Inter-Rater Reliability: Agreement between different observers.

• Internal Consistency: Consistency of score across different parts of the same test.

Relationship of Reliability and Validity

• Reliability is necessary for validity; without reliability, validity cannot be guaranteed.

Measurement Procedures

• Measurement involves classifying individuals into categories:

  • Components include a set of categories and procedures for assignment.

Types of Measurement Scales

• Nominal Scale: Qualitative categories.

• Ordinal Scale: Ranks data and establishes order.

• Interval Scale: Equal intervals without a true zero.

• Ratio Scale: Includes a true zero and equal intervals, enabling various statistical methods.

Sensitivity and Range Effects

• Range Effect: Measurement inability to detect differences.

  • Ceiling Effect: Scores cluster high with no room for variance.

  • Floor Effect: Scores cluster low, restricting measurable decreases.

Artifacts in Measurement

• Artifact: External factors influencing data validity.

  • Experimenter Bias: Influences outcome expectations.

Limiting Experimenter Bias

• Standardization or automation of procedures increases validity.

  • Use of Single-Blind Studies: Researcher unaware of expected results.

  • Use of Double-Blind Studies: Neither researcher nor participants know the expected outcomes.

Demand Characteristics and Participant Reactivity

• Demand Characteristics: Cues that suggest study expectations, resulting in biased responses.

  • Reactivity: Participants altering behavior knowing they are studied.

  • Subject roles: Good, Negativistic, Apprehensive, and Faithful.

Selecting a Measurement Procedure

• Review existing literature when selecting methods;

  • Ensure chosen methods appropriately address the research question and sensitivity needs.