After studying Chapter 4, you should know and understand the following key points:

Sampling Behavior

When a complete record of behavior cannot be obtained, researchers seek to obtain a representative sample of behavior.

The extent to which observations may be generalized (external validity) depends on how behavior is sampled.

Time Sampling

Time sampling refers to researchers choosing time intervals for making observations either systematically or randomly.

When researchers are interested in events that happen infrequently, they rely on event sampling to sample behavior.

Situation Sampling

Situation sampling involves studying behavior in different locations and under different circumstances.

Situation sampling enhances the external validity of findings.

Within situations, subject sampling may be used to observe some people in the setting.

Classification of Observational Methods

Observational methods can be classified as "observation with intervention" or "observation without intervention."

Methods for recording behavior can be classified in terms of how much of behavior is recorded.

The goals of naturalistic observation are to describe behavior as it normally occurs and to examine relationships among variables.

Naturalistic observation helps to establish the external validity of research findings.

When ethical and moral considerations prevent experimental control, naturalistic observation is an important research strategy.

Most psychological research uses observation with intervention. The three methods of observation with intervention are participant observation, structured observation, and the field experiment.

Participant Observation

Undisguised participant observation is often used to understand the culture and behavior of groups of individuals.

Disguised participant observation is often used when researchers believe individuals would change their behavior if they knew it was being recorded.

Participant observation allows researchers to observe behaviors and situations that are not usually open to scientific observation.

Participant observers may sometimes lose their objectivity or may unduly influence the individuals whose behavior they are recording.

Structured Observation

Structured observations are set up to record behaviors that may be difficult to observe using naturalistic observation.

Clinical and developmental psychologists often use structured observations.

Problems in interpreting structured observations can occur when the same observation procedures are not followed across observations or observers, or when important variables are not controlled.

Field Experiments

In a field experiment, researchers manipulate one or more independent variables in a natural setting to determine the effect on behavior.

The goals of observational research determine whether researchers seek a comprehensive description of behavior record or a description of only selected behaviors.

How the results of a study are ultimately summarized, analyzed, and reported depends on how behavioral observations are initially recorded.

Qualitative Records of Behavior

Narrative records in the form of written descriptions of behavior, audiotapes, and videotapes are comprehensive records of observed behavior.

Researchers classify and organize data from narrative records to test their hypotheses about behavior.

Narrative records should be made during or soon after behavior is observed, and observers must be carefully trained to record behavior according to established criteria.

Quantitative Measures of Behavior

Researchers often obtain quantitative measures such as frequency or duration of occurrence when they seek to describe specific behaviors or events.

Quantitative measures of behavior use one of the four levels of measurement scales: nominal, ordinal, interval, and ratio.

Rating scales, often used to measure psychological dimensions, are frequently treated as if they are interval scales even though they usually represent ordinal measurement.

Electronic recording devices may be used in natural settings to record behavior and pagers sometimes are used to signal participants to report their behavior (e.g., on a questionnaire).

Data Reduction

Observational data are summarized through the process of data reduction.

Researchers quantify the data in narrative records by coding behaviors according to specified criteria, for example, by categorizing behaviors.

Data are summarized using descriptive measures such as frequency counts, means, and standard deviations.

Observer Reliability

Interobserver reliability refers to the extent to which independent observers agree in their observations.

Interobserver reliability is increased by providing clear definitions about behaviors and events to be recorded, by training observers, and by providing feedback about discrepancies.

High interobserver reliability increases researchers' confidence that observations about behavior are accurate (valid).

Interobserver reliability is assessed by calculating percentage of agreement or correlations, depending on how the behaviors were measured and recorded.

Influence of the Observer

If individuals change their behavior when they know they are being observed ("reactivity"), their behavior may no longer be representative of their normal behavior.

Research participants may respond to demand characteristics in the research situation to guide their behavior.

Methods to control reactivity include unobtrusive (nonreactive) measurement, adaptation (habituation, desensitization), and indirect observations of behavior.

Researchers must consider ethical issues when attempting to control reactivity.

Observer Bias

Observer bias occurs when observers' biases determine which behaviors they choose to observe and when observers' expectations about behavior lead to systematic errors in identifying and recording behavior.

Expectancy effects can occur when observers are aware of hypotheses for the outcome of the study or the outcome of previous studies.

The first step in controlling observer bias is to recognize that it may be present.

Observer bias may be reduced by keeping observers unaware ("blind") of the goals and hypotheses of the study.