Experimental research Flashcards
true experimental design
manipulation of variable
random assignment of participants to different experimental groups
pottest only randomized control group design
most basic true experimental design
randomly assign participants to different treatment groups, implement the different treatments,
and then observe and measure their participants’ behavior.
pretest posttest randomized control group
randomly assign participants to different treatment groups, observe, implement the different treatments,
and then observe and measure their participants’ behavior.
aka mixed design
within-subjects
repeated measures of the same aprticpants
between subjects
treatment - no treatment comparison involves measures from different participants
Solomon Randomized Four-Group Design
4 groups, two receive treatment and two have post pre and post test. .. lots of participants, and time consuming
switching replications
intervention research
randomly assigned to treatment and control groups to start, after first post test the participants switch roles
this addresses the major problem of needing to deny the program to some participants.
factorial design
researchers manipulate two or more independent variables simultaneously.
ex) two IV and two levels of each
Internal validity
the extent to which researchers’ conclusions about cause and effect
relationships are accurate
history
some outside influence occurred during the course of your
study, and this outside influence could account for the changes in outcome you observed at the
end of your study
maturation
increases in performance that are due to the participants growth and development over time
statistical regression
very high or low scores will score closer to the mean on retest
Instrumentation
Equipped meant or tester, prob for pre/post
Selection
when two or more groups differ in a systematic way, rather than in a
random way, prior to a study
mortality
when some participants
drop out before the end of a study.
quasi experimental design
researcher manipulation of a variable but lack random assignment
nonequivalent control group design
two pre-existing or intact groups, assigning one of the groups to an experimental treatment and other to control or alternative treatment
usually pretest-post test because there is a greater need to determine if the groups were similar on the relevant variables before treatment
N
non random assignment
double pretest postest
some control over selection threat associated with different rates of maturation or learning.
repeated measures group design
one group of participants
two more measurements from same participants
serve as their own control
not need as many participants and all well matched
order effect
counterbalancing
1/2 do X1 and
1/2 do X2
single subject
still act as their own control but the results are reported for individuals
treatment withdraws design
A1-B1-A2
not ideal for intervention w lasting effects
multiple treatment design
serial order problem
A1-B-A2-C-A3
sometimes counter balance it
A1-C-A2-B-A3
