Reading...
Play button
Play button
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;
89 Cards in this Set
- Front
- Back
|
primary goals of survey research
|
identify/describe attitudes or beahviors (in a given population)
examine relationships between the attitudes/behaviors variables measured - never saying causes - just seeing relation |
|
|
self-administered questionnaires
|
mail surveys; online or emailed questionnaires; handouts
difficult to get a representative sample through email relatively easy and inexpensive no interviewer influence increased privacy/anonymity BUT must be self-explanatory mail surveys have very low response rate!! increase response rate -have inducements -make it easy to complete and return -include persuasive cover letter and/or do advance mailing -send follow-up mailings |
|
|
interview surveys
|
personal/face-to-face
- more flexible (can probe for depth) higher response rate BUT -more potential for interviewer influence higher costs telephone -most common -quickest results -compared to face-to-face - reduced costs, more privacy, more efficiency -compared to mail - more detail possible, better response rate |
|
|
cross-sectional studies
|
one sample at one point in time
|
|
|
longitudinal studies
|
more than on point in time measured
-panel -trend -cohort |
|
|
panel study
|
longitudinal study
same people each time go back to the same people at least one more time |
|
|
trend study
|
longitudinal study
different random samples different groups of people every time, but randome samples from the same population |
|
|
cohort study
|
longitudinal study
different samples, but of the same cohort something that people belong to that anchors them in time (grad class, age, event in time (alumni 2010-2013) |
|
|
relating variables when both IV and DV are nominal/categorical
|
break down the percentages by category
male female yes no etc |
|
|
relating variables if IV is categorical and DV is interval/ratio data(continuous)
|
compare mean DV scores for the different IV categories
|
|
|
relating variables if both IV and DV are interval/ratio data
|
1) convert IV to categorical variable, and then compare mean scores for DV
-taking a variable, seeing what people do with it, and then divide them depending on the data -then do it the same as the previous method 2) compute a correlation -statistical value that relates two (or more) continuous variables |
|
|
correlation
|
compute an "r" value
pearson r r tells you direction and magnitude of relationship as something increase, does something else increase of decrease difference statements or continuous statements |
|
|
direction of relationship (correlation - pearson r)
|
direction is positive or negative
positive r: as X increases/decreases, Y increases/decreases ---also called a direct relationship negative r: as X increases/decreases, Y decreases/increases --also called an inverse relationship |
|
|
magnitude of relationship (correlational - pearson r)
|
strength
R ranges from zero to 1 -1.00 <-> +1.00 the further from zero, the stronger the relationship (in either direction) r=+1.00 (perfect correlation) dots would be perfectly lined up |
|
|
what can you conclude from survey/correlational research?
|
conclude that variables are related/associated
|
|
|
two causality problems with survey/correlational
|
causal direction problem (time order)- does X cause Y? or Y cause X?
third variable problem (other explanations) -other possible, outside explanations |
|
|
to help solve 3rd variable problem
|
partial correlation
- have to think in advance about what 3rd variables could exist -include them on the survey -measure potential 3rd variables -statistically partial out effects of the 3rd variables then see if X/Y relationship still holds |
|
|
the help solve cause direction problem with survey
|
need a longitudinal study
-establish that something occurred before something else -"cross-lagged panel design" ---time 1: measure X & Y -- time 2: measure X & Y variables again for the same people (same people is the panel part) compute r's for X&Y but across the items measure measure both variables for same people at different items; then see which cross relationship holds |
|
|
purpose of experimental research
|
to test hypotheses of cause and effect
-goal is to establish internal validity - willing to sacrifice external validty in order to get internal validty |
|
|
to establish causality in a study
|
variables must be related
must establish time order must rule out other explanations/causes |
|
|
elements of a true experiment
|
manipulation of causal variables, while controlling for all other variables
random assignment |
|
|
manipulation in an experiment
|
manipuation part is the part that distinguishes causality
controlling handles the alternative explanations IV: divide in to conditions- different groups, each group is a condition DV: compare measures across conditions and see if differences exist |
|
|
random assignment in an experiment
|
random assignment of participants to conditions
everyone must have an equal chance of ending up in either condition makes groups equal before manipulation |
|
|
what do you not need in experiments?
|
representative sample
this gets external validity, dont need external validity to get internal validity |
|
|
threats to internal validity
|
if not a true experiment or if do experiment improperly, then alternative explanations become possible
|
|
|
design notation
|
X: manipulation/treatment
O: observation (measure for DV) R: random assignment |
|
|
pre-experimental designs
|
manipulation of IV, but no RA, thus many threats to internal validity
-one-shot case study -one-group pretest-posttest design -static group comparison (posttest only, non-equivalent groups) |
|
|
one-shot case study
|
pre-experimental design
X O1 (group 1) -theres a group of people that all get the manipulation -no groups to compare -no third variables controlled |
|
|
one-group pretest-posttest design
|
pre-experimental design
O1 X O2 (group 1) third variable not controlled for |
|
|
static group comparison
|
pre-experimental design
posttest only, non-equivalent groups X O1 (group 1) O2 (group 2) compare a group who recieved manipulation with a group who did not receive any manipulation no random assignment, not controlling for third variables |
|
|
threats to internal validity
|
reactivity effects
effects related to pre-testing (or measures over time) problems with procedure |
|
|
reactivity effects
|
hawthorne effect
placebo effect demand characteristics |
|
|
hawthorne effect
|
threat to internal validity
reactivity effect study done about producitivity in a factory (increase of lighitng equals increase of productivity, etc) people just reacting to the attention that was being given to them |
|
|
placebo effect
|
threat to internal validity
reactivity effect curing yourself! think you are getting something beneficial, so you make it happen |
|
|
demand characteristics
|
threat to internal validity
reactivity effects people responding on demand subjects have figured out the manipulation, and potentially even figured out the hypothesis -subject could be a good subject or a bad (subject that does the exact opposite of researchers hypothesis) protect against this by asking at the end what they think the purpose of the study was --proper control group |
|
|
effects related to pre-testing ( or measures over time)
|
testing effect
maturation instrumentation mortality(attrition) |
|
|
testing effect
|
sensitization effects
threat to internal validity idea thatj ust taking a pre-test makes participants start thinking about the topic being studied that influences their answers control by having two groups, both get tests only one is manipulated and both are tested a second time |
|
|
maturation
|
skills, opinions change over time- maturing
a natural change not due to your treatment just a natural change |
|
|
instrumentation
|
measure changing from time 1 to time 2
procedural error |
|
|
mortality (Attrition)
|
participants drop out, stop doing the study between time 1 and time 2
|
|
|
problems with procedure
|
selection bias
history effect statistical regression (to the mean) contamination of conditions |
|
|
selection bias
|
scores are due to the people that are there, not the manipulation
due to the qualities of the people that are there |
|
|
history effect
|
something happens outside of the experiment that influences scores
cause a change from time 1 to time 2, or is different between the groups |
|
|
statistical regression
|
extreme scores the next time around will go towards the mean if participant does nothing
only a threat to a study if participants have extreme scores |
|
|
contamination of conditions
|
something else got into the conditions
somes placebo people got the actual medications, etc. |
|
|
experimenter effect/bias
|
experimenters behavior or attributes, rather than treatment (IV) influences (DV)
--if you ask differently to the two groups then that could create to the internal validity |
|
|
how to control experimenter effects
|
automate the experiemt (script)
have ignorant experiment- doesnt know about the study, and oesnt know which subjects are in which group have blind experiment - knows about the study, but they are blind to which subject are in which condition double blind is best - the subjects dont know which group they are in, dont know if they are in manipulation or control group |
|
|
how to remove/control all these threats?
|
conduct a true experiment
-random assignment to proper conditions be sure to treat groups equally -all groups should be treated equally |
|
|
true experiments
|
posttest only
(control group) design pretest - posttest (control group) design solomon four-group design |
|
|
posttest only (control group) design
|
variations: more groups, several different treatments
|
|
|
pretest - posttest (control group) design
|
both groups are equally thinking about it
treated equally possible problem: iteraction of manipulation with pretest -there is priming with the pretest, so they might be thinking differently hen they see the manipulation -so it could be the manipulation, or the pretest and manipulation --to fix, ask random questions that arent related to the study -separate observations out in time so that pretest isn't priming for the posttest |
|
|
solomon four-group design
|
2 groups have only posttest design, and 2 groups have pretest-posttest design
compare both manipulation groups --if scores are different, then you know that the priming of the pretest is why they are different compare control groups |
|
|
pretesting
|
useful to check on random assignment, to get information
BUT not necessary to establish causality bad idea if treatment/pretest interaction is likely |
|
|
quasi-experimental designs
|
not true experiments (no random assignment), but have decent "comparison groups)
nonequivalent control group design (pretest-posttest, with quasi-equivalent groups) - same as pretest-posttest design but dont have random assignment -use pretest scores to match groups before manipulation |
|
|
purpose of factorial designs
|
to examine the effects of two or more IV's simultaneously
can only make causal conclusions about maniuplated variables (not subject variables) if no manipulation, then its justa survey (w/factorial-type set-up) |
|
|
simplest factorial design possible
|
two levels, a 2 X 2 design (2 levels of of something with 2 levels of something else)
each box is known as a cell |
|
|
more than two factors in factorial designs
|
3 X2 design (3 levels of something by 2 levels of something else)
3X2X2 design -easiest way to do this by doing a 3X2 for one category (male) and then a 3X2 for another (Female) |
|
|
factorial designs test for
|
main effects
interaction effects |
|
|
main effects
|
the effect of one IV individually on the DV
|
|
|
interaction effect
|
the unique effect of the combination of IV's
- not just one variable having an effect |
|
|
identifying effects
|
to see if there are main effects, compare "marginal means" of DV
to see if there are interaction effect, graph the cell means -if the lines are not parallel, then there is an interaction |
|
|
quasi-experimental designs
|
not true experiments
nonequivalent control group design time series designs |
|
|
nonequivalent control group design
|
O1 X O2 group 1
O1 O2 group 2 -examine pre-test scores to match groups before manipulation |
|
|
time series designs
|
used w/very real world data
track many observations over time, before and after a manipulation single-group interrupted time series design multiple time series design |
|
|
single group interrupted time series design
|
a bunch of pretests of the exact same thing
the same measure over and over then an interruption (manipulation), than a bunch more posttests -designed to improve upon the one-group pretest-posttest |
|
|
multiple time series design
|
similar to single group interrupted time series design, BUT with another group that does not get the manipulation
|
|
|
within subjects designs
|
every subject is in every condition, just not at the same time
issue-> people may figure out what the study is about b/c they go into differet conditions and have to fill out surveys BUT-> sometimes they want to see how people react under on conidtion vs another subjects serve as their own control group technique increases sample size, lower error but issue of alternative explanations |
|
|
problems with within subjects designs
|
the order in which you did somthing that caused the results not the manipulation
practice effect fatigue effect first treatment/contrast effects |
|
|
practice effect
|
as conditions go on, you get better/faster
|
|
|
fatigue effect
|
if each task is arduous, they get tired and start messing up
|
|
|
first treatment/contrast effects
|
whatever you do first has a reaction to whatever you are doing after that
-basically the order that you do things is causing your results |
|
|
solution of carryover/order effects
|
counterbalance orders
-randomly assign order of treatments to each subject |
|
|
laboratory vs. field experiments
laboratory experiment |
bring subjects into a highly controlled setting- lab
high control-high internal validity artifical setting - low external validity -cant generalize and cant say that this is how it is in the rela world must watch more carefully for experimenter and reactivity effects |
|
|
laboratory vs field experiments
field experiments |
not a field study!
true experiment but in the real world manipulate IV in the real world -ex: littering study more natural setting/behavior - higher external validity less reactivity harder to maintain experimental control |
|
|
content analysis used for
|
describe how much/what kind of certain messagers there are
assess image of particualr groups ompare media content to real world examine message changes over time provide background for research on media effects method for coding/analyzing open-ended data in surveys/experiments |
|
|
important issues with content analysis
|
sampling
-define population of interest -identify unit of analysis -select representative sample (but increasingly content analysis havent been doing this) coding:transforming content into numerical categories -conceptualize categories --manifest content - visible, surface content --latent content - underlying meaning operationalize categories establish reliability |
|
|
limitations of content analysis
|
purely descriptive- cannot explain why, cannot conclude anything about effects of the messages
very reductionistic |
|
|
qualitative studies of texts
|
subjectively analyze naturally occurring texts (Conversations media messages)
this is subjective. has the researchers spin rhetorical criticism critical studies analysis conversation analysis |
|
|
rhetorical criticism
|
critique form, content, imagery, elivery of speeches/pop culture
|
|
|
critical studies analysis ("cultural studies")
|
craft arguments aboutthe cultural implications of media
-typically arguing for major social change, no pointing being objective because you ar trying to change a persons opinion to agree |
|
|
conversation analysis
|
micro-level analysis of recorded conversation
going into 40 seconds of a conversaion and writing a paper on waht happened in that 40 seconds about what happens in interactions, not anything regarding relationship of people interacting |
|
|
qualitative studies of people (interpretive, ethnographic, field research)
|
goal- to develop rich understanding of peoples subjective experience
|
|
|
important features of qualitative studies
|
natural setting
researcher is not separete from participants the subjects guide what is studied inductive theory-building (start w/observation, empirical generalizations, theory) |
|
|
participant observation
|
research participates in the events/gropus under study
natives may or may not be aware of being studied |
|
|
important issues w/participant observation
|
typically purposive types of sampling (case studies common)
-purposive-> a specific criteria about a participant -choose someone purpose -case study-> one group, one organization, etc. construction of detailed field notes and records finished when achieve saturation - no new data will add and new info |
|
|
field interviewing
|
unstructured
-open-ended questions getting depth is key ethnographic conversation indepth interviews |
|
|
ethnographic conversation
|
reserach just happens to run into somenoe and turns it into an interview
naturally occurring on the fly |
|
|
focus groups
|
group discuss an issue in presence of moderator
openended questions respond to moderator as well, facilitate |
|
|
the trustworthiness of data through qualitative research
|
no concerned with reliabilty and validity of measurement --> internal and external validity
focus is on the quality of researcher interpretations -can i convince you that my data is the correct interpretation |
