Module 6 Flashcards
experiments
manipulate the IV to create levels or groupings
randomly assign people to each level
One might be a control
all subjects then measured on the same DV
confounding variable
one that is MARRIED to the IV and therefore might affect the DV
an experiment tests if tv programming affects empathy in young children. The programs are spongebob and and sesame street. Spongebob is watched in a red room and sesame street in a blue room. A counfound could be the color of the room because they are perfectly corelated with the IV
disturbance variable
looks like it affects the DV, but DV is really caused by another IV
remedies for 3rd variables
hold variables constant (all kids watch tv in the blue room)
match groups to the confounding variable
random assignment to groups equalizes people on all extraneous variables! HUGE!
between subjects design
each participant is only one comparison group (receives ONLY one level of the IV)
should have three groups:
- true control
- experimental group
- control group
within subjects design
also called repeated measures
all particpants are in all groups (they receive all levels of the IV)
order is also randomized so there are no order effects (counterbalalncing)
have more power because you need fewer subjects
controls for all extraneous variables because subject acts as their own control group
have potential effects
might introduce demand characteristics
independent samples t tests
use it when
- your IV has only two groups
- IV is between subjects in nature
- when your DV is quantitative
do these two groups come from different populations? That’s the question
cohen’s D
tells you the effect size that tells you how far apart 2 means are
t and z tests don’t tell you that. No test statistic tells you how significant an effect is
tells you how many standard deviations the two means are apart
small effect = .20 - .30
medium effect = .30 - .50
large effect = .50 +
degrees of freedom in ind samples t tests
N - 2