stats Flashcards
one way ANOVA
tests for mean differences in the DV across multiple levels of an IV
applied behavior analysis
translational research, frequency, both basic and applied, no control group
Archival research
involves carefully observing what has happened in the past, uses data that already exists
between subjects variables
will contrast two conditions of an IV between two groups of people, can’t be in both groups
within-subjects variables
will contrast two conditions of an IV within the same group of people
external and internal validity
external: generalizability, internal: describes the extent to which the experiment is free of errors and differences in the DV are due only to the IV
regression
a procedure used to predict a score on one variable from a score on another variable and multiple regression performs the same procedure using multiple variables for a single outcome
reliability
exists when researchers obtain the same results with consistency, interrater reliability is the ability for different experimenters to get the same result
validity
exists when researchers are accurately measuring something
construct validity
measures constructs should correlate with things if they’re related to and shouldn’t correlate with things they’re not related to
empirical questions
must be answerable with quantitative or qualitative data, have operational defined variables
operational definition
example a researcher wants to determine if depression is related to grades, a detailed explanation of technical terms and measurements used during data collection
converging operations
means that our confidence in a result increases when experiments with different operational definitions and procedures converge the same conclusion
confound
an extraneous variable that correlates with the IV, can control confounds by designing your exp. to eliminate them or statistically controlling them
complete counterbalancing
means all possible sequences of conditions are used ABC CBA etc.
Partial counterbalancing
means only some possible sequences of conditions are used
reverse counterbalancing
presents the conditions in one order and then the opposite order more than once
block randomization (type of counterbalancing)
presents the condition in random order within a block and this procedure is repeated a number of times
participant bias
can occur when participants have expectations and beliefs about their roles in the study
Hawthorn effect
unrealistic results b/c participants are behaving knowing their being studied
Good subject effect
persist through tasks when normally they would stop but they want to be a good subject
Evaluation apprehension
participants respond differently b/c they wish to be evaluated positively
Experimenter bias
the researcher does something that leads the participant to behave in certain ways
Ways to control experimenter bias
eliminate all interaction with the researcher, script/protocols, or use blinding
