non-experimental designs. outline 1.experimental vs. non-experimental research 2.four kinds of...
TRANSCRIPT
Non-experimental designs
Outline
1. Experimental vs. non-experimental research
2. Four kinds of non-experimental research:– Observational research– Archival research– Case studies– Surveys
1. Experimental vs. non-experimental research
• Experimental research requires exerting control.
• Experimenters exert control through manipulation and assignment
1. Experimental vs. non-experimental research
• Manipulation • researcher systematically varies conditions under which participants perform task
• “treatments”
1. Experimental vs. non-experimental research
• Manipulation• Assignment
• researcher decides who gets which treatment
1. Experimental and non-experimental research compared
• Non-experimental research allows researcher much less control
• No assignment• In many cases, no
manipulation
1. Experimental and non-experimental research compared
• Four approaches:– Observational
• Recording ongoing behavior without trying to influence it.
1. Experimental and non-experimental research compared
• Four approaches:– Observational– Archival
• Testing a hypothesis using data that the researcher did not collect
1. Experimental and non-experimental research compared
• Four approaches:– Observational– Archival– Case study
• Researcher carries out very detailed examination of individual cases
Experimental and non-experimental research compared
• Four approaches:– Observational– Archival– Case study– Survey
• Researcher collects information on beliefs, attitudes, preferences, behaviors, and their correlations.
Observational Research
• The researcher observes behavior without influencing it.
• Goals:
– to describe behavior as it naturally occurs
Observational Research
• The researcher observes behavior without influencing it.
• Goals:
– to describe behavior as it naturally occurs– to assess relationships among variables present
Observational Research
• Four approaches to non-experimental research:– Observational– Archival– Case study– Survey
• Four approaches to observation:
i. Naturalistic observationii.Participant-observer
researchiii.Structured observationiv.Field experiments
Observational Archival Case Study Survey
NaturalisticParticipant-
observerStructuredobservation
Fieldexperiments
Naturalistic Observation• Observing phenomena that
cannot be created in lab, for practical or ethical reasons
• Effects can be observed when such events occur naturally
Naturalistic Observation• Physical trace measures
– e.g., how “well-thumbed” is a book? – Where do paths through the snow go?
• Reactivity– subjects react to the presence of the observer
Naturalistic Observation
• Example: effect of early childhood isolation on later psychological development.
• We can’t isolate children to study them
• But we can use naturalistic observation when we discover such cases
Naturalistic observation• Candland (1993) –
descriptions of feral children (raised outside human cultures)
• Curtiss (1977) case studies of children subjected to unusual isolation by parents (e.g., Genie)
• Spitz (1965) – observation of institutionalized children– Showed effects of deprivation
of stimulation during infancy and early childhood
Observational Research
• Four approaches to observation:
i. Naturalistic observationii.Participant-observer
researchiii.Structured observationiv.Field experiments
Participant-observer research
• Observer joins a group for the purpose of studying group members
• Undisguised vs. disguised
• Why use disguised observation?– Access to behavior and situations
Participant-observer research
• Observer joins a group for the purpose of studying group members
• Potential cost to objectivity– Stockholm syndrome
Observational Research
• Four approaches to observation:
i. Naturalistic observationii.Participant-observer
researchiii.Structured observationiv.Field experiments
Structured observation
• Researcher exerts some control
– Eleanor Gibson’s visual cliff studies
– Piaget’s studies
• Replication depends upon following exactly the same procedures
Observational Research
• Four approaches to observation:
i. Naturalistic observationii.Participant-observer
researchiii.Structured observationiv.Field experiments
Field experiments
• Researcher manipulates one or more variables in a natural setting to determine effect on behavior
• One end of the intervention – non-intervention continuum
Field experiments example
• Crusco & Wetzel (1984)
• effect of touching on restaurant customers
• waitresses worked as confederates
• tip amount was dependent variable
Field experiments example
• Crusco & Wetzel (1984)• Compared No Touch
condition with Fleeting Touch and Shoulder Touch conditions
• Men tipped more than women
• Both men and women tipped more after being touched at some point during their meal.
Observational Research
• Four approaches to non-experimental research:– Observational– Archival– Case study– Survey
• Testing a hypothesis using data that the researcher did not collect
b. Archival Research
• Archival records are a rich source of data– No possibility of reactivity– Often very inexpensive
approach
• Government files• Corporations• Universities• Newspapers• Google cache• Internet wayback machine
b. Archival Research – an example
• Lau & Russell (1980)
– Tested external validity of laboratory findings on causal attributions
• People make internal attributions for success and external attributions for failure
Lau & Russell (1980)
• Sports pages in 8 daily newspapers
• Found 594 explanations for success and failure involving 33 sports events
• Proportions of internal attributions:
• success – 75%• failure – 45%
Case Studies
• Intensive studies of individual cases.– Strength: you learn a lot about the case studied– Weakness: results may not generalize
• We’ll come back to this topic when we look at Single-Subject Experiments.
Surveys - Definition
• A procedure for systematically collecting data on attitudes, preferences, knowledge, or behavior by asking people questions.
• The answers provide information about the group(s) that those people represent.
d. Surveys
• Use surveys when:– You want data regarding a large group of people (a population)– Measuring whole population is too expensive in time, money or other resources
• Population = all the cases of interest
Surveys
• We’ll look at surveys in greater detail in the next lecture…