Experimental design Flashcards
What method is the best for identifying causal relationship?
- experimental method
- where groups systematically differ by only one variable
What internal validity threats are there to an experimenter’s results?
- Time
- Group threats
- Participant reactivity threats
What are more specific time related internal validity threats?
- history
- maturation
- Selection-maturation interaction
- repeated-testing
- instrument change
What are more specific Group related internal validity threats?
- selection: initial non-equivalence of groups
- regression to the mean
- Differential mortality
- Control group awareness of its status
What are more specific participant reactivity related internal validity threats?
- experimenter effects
- reactivity
- evaluation apprehension
Explain how history can threaten internal validity
- Extraneous events between pre-test and post-test affect ppt post test performance
SO add control
What in internal validity?
Having confidence that the changes to the dependent variable IS DUE to the independent variable
Explain how maturation threatens internal validity
- longitudinal experiment
- ppt can change during the course of the study
- so difference due to natural development/ things happening
SO have a control at each developmental point + try and control the natural programs
Explain how selection-maturation interaction threatens internal validity
- groups of different ages
- might have different maturation rates = response to manipulation
SO make sure they only differ by only one variable (eg match age)
Explain how repeated testing can threaten internal validity
- pre-test may improve their performance?
SO avoid repeated testing/ add control group who only get post-test
Explain how instrument change can threaten internal validity
Experimenter getting bored, not good = another systematic difference between conditions
SO having highly standardised procedures, random allocation + lots of familiarisation for the experimenter
Explain how initial non-equivalence of groups can threaten the internal validity
Cohort effect: groups differ on many variables other than the one interest
Explain how regression to the mean can threaten the internal validity
- testing once can result in extreme score and picking people from this score, their scores can be vastly different next time
SO select people randomly
Explain how differential mortality can threaten internal validity
- motivation to recover from something = making the treatment more effective
Explain how experimenter effects can threaten internal validity
- experimenters have expectations which may affect performance
- pygmalion effect
- Placebo effect
SO “double-blind” procedures if possible
What is the pygmalion effect?
Teacher’s expectations affecting a pupil’s IQ
Explain how reactivity can threaten the internal validity
Hawthorne effect
- productivity affected by other factors; Draper, 2006
What is the Hawthorne effect?
ppt’s response may have been affected by their awareness of being studied/ experimental manipulatoins
What alternative reasons did Draper in 2006 give for why productivity may be affected in an experiment?
- Material factors
- Motivation - rewards?
- Learning - practice?
- feedback on performance
- Attention and expectations of observers
What are the different types of experimental design?
- Quasi-experiemental designs
- no control over allocation of subjects to groups or timing manipulation of IV - True experimental designs
What are the different quasi-experimental designs?
- One group pre-test/ post-test design
- One-group post-test design
- Interrupted time-series design
- Static group comparison design
Describe the One-group post-test design
- after an event
EG: 9/11 lvls of anxiety - prone to time effects
- no baseline against which to measure effects - pretty useless
Describe the one-group pre-test/ post-test design
- does have a baseline against which to measure effects of treatment (so like a control)
- still prone to time effects
Describe the interrupted time-series design
measurements are taken several times before and after the treatment
- still prone to time effects
Describe Static group comparison design
2 already pre-existing groups and compare them
- subjects are not allocated randomly to groups - so differences may be due to pre-existing group differences
What are the different true experimental designs?
- Post-test only/ control group design
- Pre-test/ post-test control group design
- Solomon four group design
Describe the post-test only/ control group design
- RANDOM allocation
- -> ensure observed differences are not due to pre-existing groups but can’t be certain
- 2 different groups
- 1 = control
Describe the pre-test/post-test control group design
- RANDOM allocation
- measurement in the groups
- treatment
- measurement
- -> ensures groups are indeed comparable before the experimental manipulation was administrated
Describe the Solomon four group design
- RANDOM allocation
- 4 groups
- A + B = pretest
- C + D = no pretest
- then either treatment or no treatment
- then measurement
- -> ensures groups comparable before experimental manipulation
- -> an pre-test hasn’t affected performance
- -> rarely used since need lots of people
What is the different between Within-subjects (repeated measures) and between-groups (independent measure) designs?
Between-groups (independent measures)
–> each subject in one condition only
Within-subjects (repeated measures)
–> Each subject does all of the conditions in a study
What is a potential problem with within-subjects group design?
- order effects!!
–> practice, fatigue, boredom - fixed order = vary systematically with conditions
SO randomise order of condition - counter balance order of conditions
What are the ad of between groups design therefore the opp for within subjects?
- straightforward
- need lots more subjects
- no carry over effects between conditions
- low sensitivity to experimental effects
- reversibility of conditions is NOT important
What type of method is used to study phenomena involving long time-periods?
- cross-sectional
- longitudinal
What is the difference between cross-sectional and longitudinal methods?
C: different groups of ppt for each time-phase of the study
–> separate groups 5, 10 etc each given the same test
L: each ppt is measured repeatedly over time
–> at 5yrs, 10 yrs etc
What are the disadvantages of the experimental method?
- intrusive, they know they are being watched, may affect behaviour
- Experimenter effects
- Not all phenomenas suitable for this method
- -> eg age/ sex = quasi-experimental strictly speaking
