Special Research Designs Final Flashcards
internal validity
-does the IV variable actually causes the DV
or something else at worl
-a continun
3 steps to identifying a causal relationship
1.temporal procedence
-cause come first
2.covariation of the cause and efffect
-presence cause =presence effect
-absencete cause=abesent effect
3.elimination plausable alternative explanation
key features of true experiments that helps achive intenal validity
-manipulated IV
-measure DV
-experimental control over confonding variables
ex:random assigment
the next best thing when a true experiment isnt possible
1.single case
2.quasi-experiments
3.developmental design
1.single case
a) ABA reversal design
b)ABAB reversal design
c) Multiple baseline design
single case design
-a reserch design in which the effect of the IV is assessed using data from a single participant
-the participant’s behaviour is first meaured during a baseline control time (the treatment is not used)> the manipulation is then intrudeced during a treatment period ,and the participants behaviour is anaylised
a) ABA/reversla design
A(baseline period)>B(treatment period)>A(baseline period)
b)abab
A(baseline period)>B(treatment period)>A(baseline period)>B(treatment)
it may be better then ABA because it prevents random situations to flunctuate the results (what is the chance that u got a puppy twice)
c)Multiple baseline design
-a few participants
-basiline -treatment
-key feature:stagger start of treatment
-can have more then one participants
-does it work across situation
c)Multiple baseline design
observing behaviour before and after manipulation under multiple circustances(across individuals,behaviours,settings)
cons of single case design
-questionable generalability
2.Developmental reserch
interested in age related trends throught the life span
-age as an IV
-its not assigned
age effect
any differences caused by underlying processes,such as biological or psychological changes that occur with aging
Cohort effects
Differences caused by experiences and circumstances unique to the generation to which one belongs
ex: WW
Time of measurement effect
differences stemming from sociocultural,environmental,historical or other events at the time of data collection
ex: measuring levels of happiness while there is a recession going on
effects are confound variables
development reserch
a)cross-sectional
b)longitudinal
c)sequential
a)cross-sectional design
compare groups of participants of differing ages at a single point in time
benefits
-less expensive
- immediate results
cons
-must infer developmental changes(problametic)
-differences may be due to cohert effect
b)longitudinal design
observe one group of participants repredetly over time
benefits
-evidences for developmental changes
cons
-loss of participants
-expensive and long
-measures,interests change
-results may not generalize to other cohorts
-possible time of measurment effects
c) sequential design
longitudinal designs with more than one cohort
-it is almost like a cross sectional comperasion across time (where new ages are edded)
if average across all three to see age effect
can control cohort and time measurment effect
pros and cons
pro
-can investigate cohort effect
-allows for cross sectional comperasions
-allows for longitudinal
cons
- similar cons to logitudinal
3.Quasi-Experiments
a)one-group posttest only
b)one-group pretest/postets design
c)nonquivalant control group
d)nonquivalent control group pretest-posttest
multiple repeted measures:
e) interrupted time series
f)control series
quasi-experiments
an attempt to get at causation when you can’t use a full experimental design
still attempt at IV>DV
true experiment vs quasi-experiment
true experiment:
-experimental manipulation ,
-random assigment*
-experimental control
Quasi-experiment:
-often no direct manipulation
-no random assigment*
-limited control
why quasi-experiment
wants to study an effect of Iv to Dv
but
cant manipulate or control variables
cant use random assigment (maybe for ethical reasons)
ex:impact of drugs on babies,abuse on depression,concussions impact on cognitive
quasi-experiment
subjects are not randomly assigned to conditions
subjects are selected based on the values of the independent variable,rather than having the experimenter assign values of the independent variable to subject
have less internal validity then true experiments
a)the one-group postets only design
poor design,it dosnt really tell u if they got better.
ex:offer of tutor progrem>students sign up>tutoring program>final exame (posttest) score:70
b)one-group prettest-postesst deign
slightly better.
ex:tutoring offered>students sign up> midterm score(pretest) 50>tutoring ptogram>final exam(postest) 70
threats to internal validty=types of alternative explenations
-history:something happen in the world at the same time as the study
- Maturation:participants change between pretest and postest
-Testing:taking the pretest changes responses to postest(order effect)
-Instrumental decay:treatment or measures changes
c)non-equivilant control group posttest only
ex;offered program> students who sign up vs students who dont sign up > tutoring program vs no treatment > final exam score 70 vs final exam score 60
d) non-equivilant control group pre-test post-test design
ex: program offered>students who sign up vs who didnt sign up>midterm 50 Vs midterm 50 >treatment vs non-treatment> exam score 70 vs 60
best design for a quasi experiment
regression to the mean
extreme groups differences reduce over repeted meauseres
must rule out regression to the mean as alternative explenation for change in pretest and protesst -that use extereme pre-test score to sort post-test into groups
ex; without anny effect of treatment,expected extreme high scores to drop and extreme low score to increse from T1 to T2
c)interrupted time series
a quasi-experimental design in which a treatment investigated by examening a series of measurements made over an extended time period,both before and after the treatment
f)control series
an extension of the interrupted time series design in which there is a comparison or control group
