Interpreting and Generalizing your Results Psych 231: Research Methods in Psychology
Dec 21, 2015
Announcements
This week in labs: POSTER SESSIONS Jen has final papers available at her office
hours (not in class today) Amanda’s may be done by labs this week
Research Results
Goal of Research– Experiments:
• To establish a cause and effect relationship between independent and dependent variables
– Correlational studies• To determine potential relationships between different
variables
Interpretation of your results
Statistical significance– Tell us if differences or (relationships with
correlations) go beyond what we’d expect by chance
– Don’t tell us anything about what the result really means in theoretical terms
Interpretation of your results
Theoretical Significance– Return to your theory(ies)
• Reject them? • Support (not prove) them?• Revise them?
– What are the limitations of your interpretations?• Are there alternative theories still to test?
– Yes, there are always going to be alternatives (not always good ones)
• Do your results generalize well?
Generalizing results External validity revisited
– What limits/constraints were imposed because of:• the use of experimental control• availability of resources
– Participant factors (who did you test?)• Volunteers • College students• Sex• Culture
– Location/setting factors (where did you test?)• College• Region• Laboratory
Generalizing results– Experimenter factors
• Personality, sex, experience as experiementer
– Experimental items• Do the results only apply to the stimuli that you used, or would
they extend to others– E.g., memory for:
» cat, dog, truck, car, … all concrete objects– Would the same pattern of results be seen for:
» call, dig, truth, cut, … more abstract concepts» cath, dob, trush, caf, … all non words
Threats to external validity as interactions
Generalization as interactions– Does your theory (or anyone else’s) predict an
interaction?– If you collect the appropriate data/controls, you
can do some statistical tests
In defense of college students
Convenient? – Yes, but not because researchers are lazy– Researchers do have limited resources
Criticisms about generalizability need to be backed up with theory and/or data
College students are after all humans Replications with other samples provide a safegaurd
against limited generalizability
Replications Importance of replications
– Demonstrate that the results weren’t a “one time” thing
• Single failure to replicate - don’t panic• Repeated failure to replicate - original results probably
wrong (maybe a Type I error)
Replications Types of replications
– Exact replications• Replication of the research as precisely as possible
– Conceptual replications• Same independent variables, but measured in a
different way– Same conceptual variables, different operational variables
Increasing External Validity Aggregation
– Grouping together data• Over participants• Over stimuli or situations• Over trials or occasions• Over measures (converging operations)
Increasing External Validity Non-reactive measurements
– Unobtrusive measures• Observation without participant awareness
– Naturalistic observation• To validate experimental findings
– Field experiments• Experimentation outside of the lab
Example
Results from Lab studies– Self awareness reduces likelihood of people
engaging in socially undersirable behaviors (e.g., cheating or lying)
The big picture
Generalizing your to the outside world– The bottom of the hourglass– How do these results impact …– Not always stated in the report– But, should always be in the back of your mind.
• why are you doing the research? • why is it important? • who will benefit from this research?