-
(Quasi experimental) variance is necessary...
in any research
-
(Quasi experiental) 2 forms of variance
-systematic between groups variance
-nonsystematic within groups variance
-
(Quasi experiental) Systematic between groups variance
- -experiemnetal variance (due to IV)
- *differences will be because of IV
- -extraneous variance (due to confounding variable)
- *differences will be other factors involved
-
(Quasi experiental) Controlling variance (3)
- -maximizing experimental variance
- *make sure there are real differnces between groups
- -controlling extraneuous variables
- *groups are similar a start of experiment
- -minimizing error variance
- *control with careful measurement or with specific designs
-
(Quasi experiental) < designs
-do not include the critical controls of experimental designs
-may still be used but caution
-
(Quasi experiental) 4 designs of quasi exp.
-single-group, posttest only design
-single group, pretest and pottest design
-pretest and posttest, natural control group design
-ex post fact design
-
(Quasi experiental) Ex post facto design
-a very weak design
-not good science
- does not control confounding variables
*what we do when we try to figure out, after the fact, what caused somehting to happen
N.O people > Hurricane > how depressed were they? they were already depressed before the hurrican
-
(Quasi experiental) Single group posttest only
-no control over confounding variables
-weak design, stronger than de facto
*we tend to use an implciti control group (what we think would have happened if there had been no manipulation)
-
(Quasi experiental) Single group pretest and postest design
- -pretest documetns change but factors other than the treatment cold have accounted for the change
- *history, maturation, regression to the mean etc.
- ex) people may have adjusted to their stress
-
(Quasi experiental) Pretest and postest, natural control group design
-like an experiment except that participants aer not randomly assigned to groups
- -it does not control selection but has a strong design
- *selection can play a major role in certai experiments
-
(experimental design) characterteristics of exp. design
-more than 2 levels of the IV
-random assingment
-controls threat to validity
-
(experimental design) Randomized pretest only control group designs
-random assignment controls selection
-other confoudning variables are controlled by comparing the treatment and no treatment group
-
(experimental design) randomized, pretest-postest, control group design
- -adding pretest allows us to quantify the amount of change following treatment
- *allows us to verify groups were equal at start of experiment
- -strong basic research design with excellent control over confounding variables
-
(experimental design) multilevel, randomized, between subjects design
- -may or may not include a pretest
- *posttest is necessary
-multigroup extension of the basic experimental design
- -controls virtually all sources of confounding variables
- *we can have as many groups possible
- **most common used in drug companies
- ***how much dosage works best
-
(experimental design) Solomon 4 group design
- -combines 2 basic experimental designs
- *randomized postest only, control group
- *randomized, pretest-postest, control group design
-allows the assesment of an interaction between the pretest and the treatment
*uncommon research design
-
Research is design to...
measure and control sources of vriance
-
experienteal design have 2 elements
-random assingments of participants to conditions
-at least one control group
|
|