lecture 5 Flashcards
(15 cards)
demand characteristics
- participants behave a certain way because they know they’re being studied
1. guess the hypothesis and behave that way
2. guess the hypothesis and behave the opposite way
3. Not behaving they way they normally would because they want to behave how they think they should
4. Being studied makes people nervous so they behave differently
expectancy effects
- experimenter interacts differently with participants ex) horse and math example
- experimenter records children’s data differently ex) observing children altruistic behavior
ceiling and floor effects
- all their scores are really high, there will always be outliers. The test is too easy and there is no variability (ceiling)
- same thing if the test is really hard (floor)
how to avoid these pitfalls
- Demand characteristics:
1. Deception- not specific about what you want to test
2. Post experiment interview- ask them if they figured out the hypothesis
3. Observation- they don’t know they’re being watched - Expectancy effects:
1. Blind studies- person performing test doesn’t know the hypothesis
2. Scripts- don’t give away ques, tell people the same stuff
3. Automation- don’t give away ques
4. Training- practiced - Floor and ceiling effects
1. Ensure task is moderately challenging
populations and samples
populations= all the individuals that make up a group ex) normally functioning adults sample= smaller “representative” group selected from the population ex) intro psych students
probability vs non probability
access to whole population is with probability, if you don’t then its non-probability
probability sampling techniques
systemic- every nth subject (can only do this if you have access to whole population)
stratified- sperate population into sub groups ex) ethnicity, education
cluster sampling- using existing “natural” groups
non-probability sampling techniques
convenience sampling- the “easy way” ex) intro psych students
quota sampling- making “groups” that are proportional ex) males and females
WEIRD
-western, educated industrial, rich, democratic
why is WEIRD a problem
- Conclusions may represent extreme end of a continuum of responses
- policies made based on conclusions from WEIRD samples may not be appropriate for all circumstances (economic behavior, self-esteem, conformity, moral reasoning)
sample sizes- what to do
- Power analyses: need ¾ sample size, effect size, significance level, power
- Precedent: what have other people done
sample sizes-what not to do
- Haphazardly collect data- defensible decision
2. Run subjects- check data-run subjects-check data: not good science
other considerations for sample size
Very large samples makes it hard to get statistical significance: you will get an effect but doesn’t mean its significant
large sample size can make it easier to reach statistical significance
statistical significance does not always mean practical significance
descriptive research
- goal of psychology: describing behavior: what does “x” look like
- what kinds of variables/measurements can you use? Nominal, ordinal, ratio/interval
- goal of psychology: describing and or/predicting behavior: is there a relationship between X and Y
- what kinds of variables/measurements can you compare? Nominal, ordinal, ratio/interval
experimental research
goal of psychology: cause and effect relationship: does X cause Y? ex) does drinking beer cause GPA to go down
