Naturalistic Observation and CaseStudy Research Graziano and Raulin

Naturalistic Observation and Case-Study Research Graziano and Raulin Research Methods: Chapter 6 Graziano & Raulin (2000)

Challenge of Low-Constraint Research n Usually involves careful observation of participants in their natural surroundings – – – Can be very difficult to observe behavior in natural surroundings Often we are not sure what behaviors are important until after we have observed for a while Without the controls of the laboratory, participants are free to do what they want to do, and not what we are hoping to observe Graziano & Raulin (2000)

Naturalistic Observation Examples Charles Darwin’s voyage on the HMS Beagle (the basis for his theory of natural selection) n Jane Goodall's study of chimpanzees n Dian Fossey’s study of the mountain gorilla n Adeline Levine’s study of the Love Canal n Rosenhan’s study of psychiatric hospitalization n Graziano & Raulin (2000)

Case-Study Examples Sigmund Freud’s study of patients, which formed the basis for his psychoanalytic theory n E. L. Witmer’s study of children in the first psychology clinic in North America n Jean Piaget’s study of the development of children, which led to numerous theories of child development n Graziano & Raulin (2000)

Value of these Methods n When to use low-constraint research – – n For questions involving the natural flow of behavior When first studying a new research area When testing the feasibility of a procedure To test the generalizability of laboratory findings Information gained from these methods – – – Provides new descriptive information Can negate a general proposition Provides information about contingencies Graziano & Raulin (2000)

Problem Statements and Hypotheses n Problem statements are often general and flexible in low-constraint research studies – n Problem statements and hypotheses may evolve (i. e, start out general and become more specific) as the study progresses, building on new observations Unable to test causal hypotheses with lowconstraint research Graziano & Raulin (2000)

Making Observations Unobtrusive observation: observing behavior without participants’ knowledge n Participant observation: observing behavior while participating in the situation n Want to reduce measurement reactivity n – People behaving differently when observed n Reactive measures: enhance reactivity n Nonreactive measures: minimize reactivity Graziano & Raulin (2000)

Archival Records n Exist independent of a research study – – Kept for purposes other than research, but they may be valuable in some research studies May include government records, school and hospital records, census data, etc. Access to such archival records are restricted by legal and ethical constraints n Valuable data source for some studies n Graziano & Raulin (2000)

Ethical Issues n Use of unobtrusive measures (including archival records) raises ethical issues – – n Participants are not given the right to consent Some of the archival records contain sensitive data Researchers need to show the necessity for unobtrusive measures and safeguards to protect the rights of the participants – Must have IRB approval Graziano & Raulin (2000)

Sampling of Participants n Try to obtain a representative sample – n Representative samples allow us to generalize findings to the larger group Sampling is often not under the control of the researcher in low-constraint research – – Therefore, caution is required in interpreting the results Generalize only to similar participants and NOT to the general population Graziano & Raulin (2000)

Sampling of Situations n People (and animals) behave differently in different situations – To get an adequate picture of behavior, we need to sample the behavior in many situations Sampling many situations will give us an idea of how consistent behavior is n It also gives clues about what factors may be affecting the behavior n Graziano & Raulin (2000)

Sampling of Behaviors n Even in the same situation, people may behave differently on different occasions – n Repeated sampling of behavior in a specific situation will indicate the consistency of the behavior Repeated observation (essentially a replication) prevents us from developing theories based on a single, unusual behavioral response of the organism Graziano & Raulin (2000)

Evaluating the Data n The data from low-constraint research is a rich set of information – – n Data usually needs to be coded to provide simplification and organization The analyses will depend on the questions and the level of data produced after coding Must be cautious in interpreting data from low-constraint research Graziano & Raulin (2000)

Limitations Poor representativeness n Poor replicability n Ex post facto fallacy n Limitations of the observer n Going Beyond the Data n Graziano & Raulin (2000)

Poor Representativeness Most low-constraint studies have small samples that were not randomly selected from the population n Rarely do the samples represent the population n Consequently, it is dangerous to generalize your findings to the population n Graziano & Raulin (2000)

Poor Replicability n Studies can be replicated only if – – n In low-constraint research – – – n The procedures are clearly specified The procedures were followed exactly Procedures are often not specified They may change as the study continues They are often unique to the observer Therefore, replication is very difficult Graziano & Raulin (2000)

Ex Post Facto Fallacy n Interpreting an observed contingency as if it represented a causal connection – n Low-constraint observation will never provide the controls for such strong conclusions If ex post facto conclusions are interpreted as hypotheses for further research, and not as established facts, then they serve a useful scientific purpose Graziano & Raulin (2000)

Limitations of the Observer Low-constraint studies often rely on the observational skills of the researcher n Detailed procedures are not specified n – – n Specific procedures would decrease flexibility However, detailed procedures could constrain the observer in a way that would limit experimenter biases Give up some control for the flexibility Graziano & Raulin (2000)

Experimenter Bias n Without the controls found in higher constraint research, it is difficult for the researcher to avoid influencing participants – n Experimenter reactivity is the term used to describe this unwanted influence on the participants’ behavior Experimenter effects can be controlled in higher-constraint research Graziano & Raulin (2000)

Going Beyond the Data Low-constraint data are often intriguing n Nevertheless, one must be careful in drawing strong conclusions n Rosenhan has been criticized for broadly over-interpreting his data, even distorting the data n Interpretation should take into account other information we know about a phenomenon n Graziano & Raulin (2000)

Survey Research n A widely used research technique – – n Provides information about people’s attitudes, experience, and knowledge Used extensively today by researchers, politicians, and news organizations First introduced in the 1830 s in England – Study the impact of the Industrial Revolution on people’s lives Graziano & Raulin (2000)

Types of Surveys n Status surveys – – n Descriptive survey about the current status of the population sampled Descriptive information can guide policy and inform policy makers Survey research – Seeks to identify relationships among the variables studied in the survey Graziano & Raulin (2000)

Steps in Survey Research Determine what you want to study n Define the population to be studied n Design, construct, pilot, and refine the survey instrument n Select a representative sample n Administer the survey n Analyze, interpret, and communicate the results n Graziano & Raulin (2000)

Form of the Survey Instrument n Group-administered survey – – n Written survey with instructions included Administered in groups or through the mail Individual survey schedule – – – Administered in person or over the phone Can be rather complex, provided there is a clear procedure for the interviewer to follow Allows clarification more easily than in groupadministered surveys Graziano & Raulin (2000)

Developing a Survey Instrument Use clear language and explicit instructions n Types and number of questions will depend on the purpose and the type of survey planned n Types of items n – – – Open-ended items Multiple-choice items Likert-scale items Graziano & Raulin (2000)

Sample Open-Ended Questions How do you typically handle interpersonal difficulties with your co-workers? n What are the most important values to instill into today’s children? n If you were president, what issues would you make your top priority? n What situations are particularly stressful for you? n Graziano & Raulin (2000)

Sample Multiple-Choice Questions n How frequently do you take a sick day from work? a) never b) once or twice a year c) 3 to 5 times a year d) 6 to 12 times a year e) at least once a month n Identify the issue that you believe is most critical to this country’s future. a) the economy b) education c) integrity in government d) national defense e) some other issue Graziano & Raulin (2000)

Sample Likert-Scale Questions Rate each item on the scale shown to indicate your level of agreement: n I believe in the Bill of Rights. strongly agree n uncertain disagree strongly disagree I think that everyone should vote. strongly agree n agree uncertain disagree strongly disagree Most politicians cannot be trusted. strongly agree uncertain disagree Graziano & Raulin (2000) strongly disagree

Issues in Sampling n Need to clearly define the population and then develop a strategy for adequately sampling from the population – – n Identify each member of the population Sample from the comprehensive list The more representative the sample, the more valid our conclusions from the survey Graziano & Raulin (2000)

Sampling Procedures n Nonprobability samples – n Convenience sample gathered quickly, but with little interest in drawing strong conclusions Probability samples – Simple random sampling n Every – person has an equal chance of being included Stratified random sampling n Random sampling within clearly defined strata (subdivisions of the population) Graziano & Raulin (2000)

Sample Size n Sample size is based on several factors – – Costs and time constraints Degree of precision needed n Larger samples will provide more precise estimates of population parameters n More heterogeneous populations require larger sample sizes n The larger the sample, the more narrow the confidence intervals Graziano & Raulin (2000)

Research Design of Surveys n Cross-sectional designs – n Administer the survey once to a sample Longitudinal designs – – – Repeatedly survey the same participants Allows us to see changes in attitudes, experience, and knowledge over time Difficult to get participants to submit to such long-term research Graziano & Raulin (2000)

Summary Low-constraint research methods provide valuable information n Types of low-constraint research n – – – n Naturalistic observation Case-study research Surveys Appropriate caution must be used because of the inherent limitations of these methods Graziano & Raulin (2000)