validity Flashcards
what is validity
the extent to which we measured what we set out to measure
internal validity
the extent to which we can be sure the researcher has measured what they set out to measure
whether extraneous and confounding variables were controlled and sources of error were eliminated
mundane realism
if a task has low mundane realism it could mean that natural behavior not measured properly therefore the results are not genuine, therefore not representative of the population and so cant be generalized to outside the experiment.
factors that effect internal valisity
participant effects researcher effects situational variables extraneous variables confounding variables experimental design experimental method demand characteristics hawthorn effect screw you please you effect bias
how to maximize internal validity within designs
in independent groups design make sure to allocate groups randomly using lottery method eg names in hat to control for bias
repeated measures use counterbalancing ABBA to control for order effects or practice
blinding
using sigle and double blind trials increase validity
single the pps don’t know what group they are in and so there are little demand characteristics
in double blind the researcher and Ps don’t know what group they are in this controls demand characteristics and investigator effects.
external validity
the extent to which the conclusions for the study can be generalized to other contexts
types of external validity
settings : ecological validity
people : population validity
time : temporal validity
ecological validity
whether the study measures real life behavior and whether the conclusions can be generalized in other settings such as real life
population validity
whether the results can be generalized to other populations
temporal validity
whether the results can be generalized in other time periods
face and concurrent validity
face validity is whether the ‘face’ of the study looks as if it has measured what it set out to measure , this can be checked by an expert
concurrent validity is when your research is close to well known research on the same study topic , a close correlation between both sets of data would indicate concurrent data
there should be control over what two aspects of the experiment by the researcher
the way it is carried out and the way it is designed
otherwise there is no control for investigator effects which lead to demand characteristics which lead to low mundane realism and low internal validity which leads to low external validity and therefore not generalizable
how is error created
error is created by variables that interfere with measuring the variables we set out to measure
how to control for error
unambiguous (not open to more than one interpretation) hypothesis
standardization
randomization
