Experimental methods Flashcards
Confounding variables
vary systematically with the IV
- may not affect the P in the same way
Extraneous variables
affect the DV if not controlled for
- temperature of day, weather, lighting in room , time of day
Conducting research online
- questionnaires
- conduct experiments through the internet
> large samples
> diverse Ps
> cheaper
X ethical issues
X EVs n CVs
X biased sample
X social desirability
X dishonest people
LAB experiments
- controlled artificial setting
- standardised procedures
- deceive Ps on true aims
- variables carefully controlled
- manipulate variables and measure effects
- looks for difference in conditions
> establish cause n effect
> EVs can be controlled
> quantitative data
> replicable/ reliable due to SPs
X low ecological validity
X artificial conditions
X demand characteristics
X experimenter effects
FIELD experiments
- manipulate at least 1 IV
- natural environment
- IV directly manipulated by researcher
- street, school, hospital, their own home
> high ecological validity/ mundane realism
> reduced demand characteristics
> still in control of IV to measure DV
> cause n effect discovered
X harder to replicate = less reliable
X less control of EVs and CVs = reduce validity
X risk of demand characteristics = reduce realism
X ethical issues- no fully informed consent/ privacy/ deceived
Types of QUASI experiment
NATURAL
- IV not under direct control of the experimenter
- IV is naturally occurring
- impractical/unethical- separating child from mum
DIFFERENCE
- naturally occurs/ exists in a person
- different IQs, men, women, age
- Ps randomly assigned
Strengths and weaknesses of QUASI experiments
> research phenomena that cannot be studied in any other way- both practical and ethical
study real problems e.g divorce
X cannot draw definite cause n effect
X many EVs beyond control
X replication- impossible`
Matched pairs
Matched on key characteristics
- age, IQ, personality
Pre tested - 1 control, 1 experimental
> low chance of individual differences
> no order effects
> low demand characteristics
X more Ps needed
X matching is difficult
X time consuming
Repeated measures
P takes part in both conditions
- experimental and control
> less Ps needed
> no individual differences
X more materials needed
X order effects and boredom
X high demand characteristics
= COUNTERBALANCING
Independent groups/measures
Participant only takes part in one condition
- allocated randomly to either group
> less materials needed
> no order effects
> low demand characteristics
X more Ps needed
X individual differences- P variables
= random allocation
= names in a hat
= allocate alternately
What causes RELIABILITY issues?
1) variables not clearly defined
2) research only conducted once
3) instructions not given in the same way
4) Ps are not asked Qs in the same way
5) Only 1 researcher- no inter-rater reliability
3 issues in RELIABILITY
1) LACK OF OPERATIONALISED VARIABLES
= replication impossible
+ inter-rater reliability
2) ORDER EFFECTS
= repeated measures design
+ counterbalancing
3) LACK OF STANDARDISED PROCEDURES
= inconsistent measuring tools
= uncontrolled environment
PREDCITIVE validity
> accurately forecasts future outcome on more broadly related topic
-varies situations?
compare results of out measure with other measures to see if it predicts what we expect
ACCURATE PREDICTIONS?
FACE validity
> based on intuition
does test measure what it claims to?
- objective = purpose is clear
most simple method
CONCURRENT validity
> does it agree with existing measures?
compare with an established measurement
Ps do new and old test
if there’s high agreement on both it has CONCURRENT VALIDITY
