exam 4 Flashcards
within subjects design (paired/repeated measures)
the same participant gets all the treatments that everyone else got
advantages to within subjects design
-eliminates worry about individual differences because you have the same people being studied twice
-requires fewer participants
-each subjects is their own control
which is more powerful - within subjects or between
within subjects design
order effects
performance in one condition is influenced by participation in another condition
carry over effect
lingering after effects of one of the earlier treatment conditions
-specifc treatment condition produces a lasting change in the participants
progressive error
changes in behavior that are related to experience in the project (but not to one specific condition)
-fatigue, practice, retro memory
order effects as confounds
if all subjects relieve the treatment conditions in the same order, then order effects can potentially be a confounding variable
counterbalancing
distributes the order effects across the treatments evenly
complete counter balancing
subjects are run in all possible sequences of treatments
partial counter balancing
s a way of ordering the presentation of levels of the independent variable to minimize some of the effects of sequential confounding variables.
latin square
design that helps to control for sequencing effects in within-subjects designs
contrast effects
wondering if being exposed to all levels of the independent variable changed behavior in one or more conditions
oppenihiemer
-belief in a just and orderly world
-compared dutch children and adults
-belief decreases w age and more w social inequality
langer and rodin
two groups of elderly people on different floors
a. responsible group
b. no responsibility group
— found that group a was happier, more alert and lived longer
quasi experiment
a design that is like an experimental design but lacks the key ingredient of random assignment to groups.
quasi experiment lack
-random assignment (between subject)
-counter balancing (within subjects)
-IV cant be manipulated
-has some protection against validity threats
non experimental
research design not having protection from the threats to internal validity provided by experimental or quasi-experimental designs.
between subjects
subject grouped together by factors other than random assignment and then given different treatments
between subjects concern
lack of random assignment
differential research design
comparing pre existing groups defined by some participant variable
-no manipulation of IV or random assignment to groups
postest only non equivalent control design
experimenter administers treatment to one pre-existing , not to the other
-problem: its measured after treatment
pretest -postest non equivalent control group design
both groups are measured before the treatment as well as after
-protects against possibility that groups were different to begin with
pre-post
one group is observed before and after a treatment is put into place
-no counter balancing
-might be used if a treatment is permanent or long lasting
one group pretest postest design
consists of measure, treatment, measure
time series design
multiple measurements before and after treatment
-useful if treatment takes a while to start working
interrupted time series design
determining effect of a naturally occuring event (cant control) in a time series
-ptsd and catastrophe
second hand smoke of montana
smoking ban and then the ban got taken away
-interrupted time series design
developmental research
no manipulation/ random assignment
developmental research
no manipulation/ random assignment
cross sectional design
people of different ages (cohorts) examined at the same time
strength = short term
weakness - cohort effects - differences may be due to different in histories
longitudinal
testing the same group over time
strength - avoid cohorts
weakness - long term, attrition and or mortality
rodman and berger one factor
2 levels of IV or quasi IV
factorial design
an experimental design containing more than one independent variable in which every level of each independent variable is combined with every other level
