Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
67 Cards in this Set
- Front
- Back
types of observation
|
naturalistic
participant contrived |
|
sampling procedures
|
random
stratified cluster convenience quota |
|
type of observation:
naturalistc |
observe them in their own environment
|
|
type of observation:
participant |
get into group, be one of them
|
|
type of observation:
contrived |
bring them to your observation location to watch them unnoticed
|
|
problem w/ disguised vs. nondisguised observation
|
social desirablity
|
|
sampling procedure:
random |
random assignment
|
|
sampling procedure:
stratified |
randomly sample from groups of interest
ex. race or gender |
|
sampling procedure:
cluster |
sample group rather than individual
ex. university or city |
|
sampling procedure:
convenience |
sample participants who are readily available
ex. teacher has class as participants |
|
sampling procedure:
quota |
representative proportions
|
|
sampling
|
selection of participants
*must be representative of population |
|
correlational studies
|
-see if variables are related at all
-relation of variables to eachother -two continuous variables -can not determine causal effects |
|
structure of correlational hypotheses
|
___ is (positively/negatively) associated with ___.
|
|
factors that influence magnitude of correlations
|
-restricted range/variance of scores
-low reliability of measures -outliers -sample size influences significance |
|
determining correlations
|
-scatterplots
-Pearson's r |
|
Pearson's r
|
-range +1 to -1
-r^2 (coefficient of determination): % of variance shared b/w variables -influenced by sample size -.10 is weak, .30 is moderate, .50 is strong - +/- doesn't matter - only gives direction |
|
pros and cons of correlational designs
|
pros
-can determine the unique relation of 2 variables -mediation and moderation effects cons -can't determine direction of causation -effects of outliers on magnitude |
|
structure of multipredictor hypothesis
|
(variable1), (variable 2), and (variable 3) all predict unique variance in (variable 4).
(variable 1) predicts (variable 2) controlling for (variable 3) and (variable 4). |
|
controlling for
|
-variance explained by competing variable is held constant
ex. violent crimes and ice cream sales -controlling for regional temp. -are ice cream sales related to violent crimes when the temp. is always the same? |
|
conceptual construct
|
-abstract definition
-psychological construct ex. aggression |
|
operational construct
|
-concrete definition
-how you will measure the construct ex. blasting another person w/ loud noises -measurement issues |
|
reliability
|
the consistency or dependablility of a measuring technique
|
|
ways to test reliability
|
-test-retest
-interitem -interrater |
|
test-retest reliability
|
consistency over time
|
|
interitem reliability
|
consistency of items to each other
|
|
interrater reliability
|
agreement of observations b/w raters
|
|
measuring reliability
|
-cronbach's alpha >.80
-item-total correlations >.30 -T1-T2 correlation >.70 -intraclass correlation for raters |
|
increasing the reliability of a measure
|
-standard administration
-reword or discard bad items -add items -train your coders |
|
face validity
|
-extent to which an instrument APPEARS to measure what it is supposed to measure
-often clear-cut items -problem: social desirability |
|
construct validity
|
the degree to which the study measures and manipulates the underlying psychological elements that the researcher claims to be measuring and manipulating
|
|
testing construct validity
|
-how does the measure relate to other measures?
-convergent validity -discriminate validity |
|
convergent validity
|
should correlate positively with similar measures
|
|
discriminate validity
|
should correlate negatively or not at all with dissimilar measures
|
|
criterion validity
|
the extent to which the measures allow researchers to distinguish among participants on the basis of a particular criterion
|
|
ways to test criterion validity
|
-concurrent validity
-predictive validity |
|
concurrent validity
|
measure distinguishes b/w people in the present
|
|
predictive validity
|
measure distinguishes b/w people on a relevant behavior in the future
|
|
two types of problems in measurement
|
bias
-systematic errors that "push" scores int he desired direction -needs to be resolved immediately random error -unsystematic errors due to chance -typically avg. to zero w/ repeated use |
|
utilitarian perspective
|
all research should be conducted as long as it serves a useful purpose for society as a whole
|
|
deontological perspective
|
no research should be conducted that violates the rights of the participants
|
|
current ethics principle
|
-no research should be conducted in which the costs outweigh the benefits
-determined by the Institutional Review Board (IRB) |
|
cost in conducting research
|
-time
-money -effort -discomfort -embarrassment -pain -stress -threat to health -boredom -death (non-human) -recalling trauma -lack of significant impact |
|
benefits of research
|
-knowledge
-improving assessments -improving quality of life -improving an established procedure -resolving a societal problem |
|
informed consent
|
informing the participants of the nature of the study and obtaining their agreement to participate
|
|
informed consent must...
|
-be presented in a way that is understandable
-be obtained from a person who can give valid, legal consent -mention that participation is voluntary -include a signature line -describe the study and the potential risks involved -encourage participation w/o coercion |
|
what if you use deception?
|
-make sure it is necessary
-you must metion it during the debriefing |
|
confidentiality
|
the data participants provide may only be used for the purposes of the research and may not be divulged to others
|
|
assuring confidentiality
|
-make the responses completely anonymous
-assign a unique number to each particpant -keep all data in a locked cabinet |
|
debriefing
|
the procedure through which research participants are told about the purpose of a study after it is completed
|
|
debriefing should...
|
-take place as soon as the study is complete
-clarify the purpose -remove any discomfort -obtain participant's reactions -involve thanking the participant -give the participant a "script" if you use deception |
|
professionalism
|
all researchers should treat the particpants with respect and attend to any discomfort as soon as it becomes apparent
|
|
violations to professionalism
|
-failing to show up on time to conduct the study
-unable to run the study properly -being disrespectful -being ungrateful for participation -falsifying data or results |
|
types of variables
|
-categorical
-discrete -continuous -all analyses involve defining variables as either categorical or continuous |
|
categorical variable
|
-yes or no
ex. male or female, did it or not, true or false |
|
discrete variable
|
-ranges from high to low
-set increments (1,2,3,4,5) -whole # and in betweens ex. 2 or 2.5 |
|
continuous variable
|
-ranges from high to low
-infinite increments -whole # and in betweens ex 2 and 2.5 |
|
types of scales
|
-nominal
-ordinal -interval -ratio |
|
nominal scale
|
-number assigned to characteristics
-aka categorical -gender, race nominal sounds like "name" - assign # to name "1"-male taking survey "2"-female taking survey |
|
ordinal scale (rank)
|
-rank-ordering
-best performance, second best, etc spaces between 1, 2, 3 not necessarily even ex. race 1, 2, 3, 4 sounds like "order" |
|
interval scale
|
-continuous scale with no zero point
ex. stress -no point where there is an absolute zero point -impossible to have none |
|
ratio scale
|
-continuous scale with zero point
ex. weight, tempurature -descrete variable ex. 2 is 2x of 1 4 is 2x of 2 zero in ratio = scale has a zero in it |
|
self-report items
|
-specific
-simple -relevant -proper and consistent response format *answer same way for all questions, same format -need some reverse-coded items *other extreme "i like myself" and "i don't like myself" ex. Rosenberg Self-Esteem Scale |
|
Rosenberg Self-Esteem Scale
|
1-strongly agree
2-agree 3-neutral 4-disagree 5-strongly disagree |
|
interview
|
-friendly atmosphere
-conceal personal reactions -ask questions EXACTLY as worded -do NOT lead the interview |
|
problems in self-report
|
-social desirability
-leading questions -double-barreled questions *have 2 phrases in which participants don't know which to answer ex. "i don't like to go out, I'd rather read a book" - they could say yes to one half but say no to other half -response sets *responses hang around 3 or only answer with 1 or 5 -length *don't make it too long or participants won't answer truthfully if they are bored |
|
things to remember
|
-don't "reinvent the wheel"
*if it already exists, don't make another one -pretest your measure before you use it |