EXPERIMENTAL DESIGN Flashcards
what are the different types of unrelated design?
between-subjects
between-ppts
independent groups
unrelated groups
uncorrelated groups
what are the different types of related design?
within-subjects
within-ppts
repeated measures
dependent groups
related groups
what is the difference between between-subjects and within-subjects?
between-subjects = random assignment to conditions -> each ppts experiences one out of two conditions
within-subjects = random assignment to conditions -> condition experienced first (control/experimental) -> swap conditions second time
2 reasons for using research lab -
practicalities - equipment/apparatus too heavy/expensive to move
experimental control - all other factors constant apart from the ones manipulated
holding extraneous variables/environmental factors depends on how strong effect of manipulated variable is
characteristics of true randomised experiment
experimental manipulation
standardisation of procedure
random assignment to conditions/order
experimental manipulation
manipulation creates and has effect on IV
manipulation has 2 levels/more
easy to manipulate one thing by accident, be careful
effectiveness and exp manip should be checked, questions at debrief
standardisation of procedures
differences between exp and control condition should be exp manipulation
experimenter interactions should be standardised
random allocation prevents collision of groups, lack leads to time of day effects (diff cog functions shows optimal performance at diff times of day)
uncontrolled variation = analysis is less statistically significant
unintended differences in treatment of diff groups leads to wrong conc
randomisation
used for allocating ppts to exp/control condition in unrelated design
randomisation in order of exp condition in related designs
matched pairs = related design, must be done on dependent variable
pre and post test sensitisation effects
pre-test way of checking if random assignment is equated in exp/control groups before manipulation, or else x know if diff due to pre-existing differences
measurement after manipulation = post-test
post-test mean - pre-test mean
however, alerts ppts to purpose of research so disguise in other task or do pre-test some time after
within subjects design provides..
more sensitive test of difference between condition as it controls differences for individuals
what can occur if we don’t counterbalance?
fatigue/boredom
practice
carryover, asymmetrical transfer or differential transfer
what is counterbalancing?
when one set of ppts engage in the exp condition first and control second and the other vice versa
means conditions occur equally frequently
key points
causal relationships through random allocation, manipulated of variables and standardised procedure
between vs within subjects design requires diff statistical tests, within subjects allows fewer ppts but needs pre-test due to sensitisation
experimental condition means level of manipulated variable is higher
statistical diff at post-test due to manip variable if x differ to conditions at pre-test
