class 8-9 Flashcards
research study design
blueprint or detailed plan for conducting a study
what is a design
-addresses the research problem in an appropriate way
-determines the degree of bias and/or control
-guides the researcher in planning and implementing a study
experimental (interventional) design
-experimental
-quasi-experimental
non-experimental (observational) design
-correlational
-descriptive
-comparative
design: time dimension
cross-sectional design
data collected at one point in time
design: time dimension
prospective design
data collected at one point and then again at future point(s)
may be longitudinal
design: time dimension
retrospective design
data collected and compared against previously collected data or in relation to something that occurred in the past
causality - design concepts
A causes B
relies on temporality being able to determine the independent variable occured/changed before the outcome
multicausality - design concepts
same as causality but A,C,D… causes B
manipulation - design concepts
-intentionally controlling/altering a condition or treatment (independent variable)
-must be consistent
bias - design concepts
distorts findings away from the truth
-researchers must attempt to eliminate or substantially reduce randomness & blinding
control - design concepts
keeping study conditions constant
-reduce bias
-improves accuracy of findings
-impacts generalizability
control can be applied to:
sample
setting
treatment/intervention
measurement
extraneous variables
design validity
-measure of truth or accuracy of a study
statistical conclusion validity
are the conclusions drawn from statistical analyses an accurate reflection of the real world
type 1 error statistical conclusion validity
incorrectly concluding that a difference or relationship exists
type 2 error statistical conclusion validity
incorrectly concluding that a difference or relationship does not exist
construct validity
-congruency between operational definitions and conceptual definitions
-to what degree do the instruments sued actually measure the variable being studied