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;
46 Cards in this Set
- Front
- Back
Why case studies?
|
Regression analysis really only tells us about correlations. Sometimes we need to know about the effect of the sequence of events, and the direction of the events.
|
|
Difference between correlation and causation
|
Case studies trace a causal process that link the proposed independent variable to the dependent one
|
|
method of agreement
|
if the outcome is the same despite differences among your cases, you want to search for the common characteristics between them that could explain the similarity
|
|
method of difference
|
you choose two cases. in one case, the phenomenon of interest occurs. in the other it does not. the two cases should be chosen in a way that they almost share all other traits. if they share all except one, then this one difference can be taken as the reason for the different outcomes.
|
|
examples of universe of cases
|
world wars, nuclear weapon sites, countries
|
|
single case studies
|
emerges from empirical puzzle. usually starts by an observation that something does not fit our expectations based on available theories and/or logic
|
|
paired comparisons
|
following a single case over time - something called within-case comparison
|
|
process tracking
|
involves examining pieces of evidence in order to support or overturn competing explanations of a particular outcome.
|
|
the coding frame
|
list of all possible responses to a question and their accompanying numerical codes
|
|
rules for coding
|
1. codes must be mutually exclusive: any response can fit into one, and only one category
2. codes must be exhaustive: you must cover all possible options 3. codes must be applied consistently throughout |
|
purpose of IRB
|
protect the well-being of human subjects of medical and social research
|
|
when do you need IRB approval
|
all research involving human subjects
|
|
informal interviews
|
researcher tries to remember conversations heard during the course of the day
|
|
unstructured interviews
|
researcher wants people to open up and express themselves at their own pace
|
|
semi-structured interviews
|
write list of questions to be covered
|
|
structured
|
questionnaires
|
|
deference effect
|
when people tell you what they think you want to hear
|
|
expectancy effect
|
creating the objective results we want to see
|
|
why case study
|
regression only tells us about correlations. sometimes we need to know about effect of sequence of events and direction of effect
|
|
t-test
|
for significance of the correlation of coefficient relationship between variables
|
|
one sample z-test
|
differences between one sample and population
|
|
regression
|
relationship between more than two variables
|
|
t-test for dependent samples
|
two groups of variables being tested more than once
|
|
repeated ANOVA
|
more than two groups of variables being tested more than once
|
|
t-test for independent samples
|
two groups of variables instead test only once
|
|
simple ANOVA
|
more than two groups of variables being tested only once
|
|
what's the study of probability
|
basis for determining the degree of confidence we have in stating a particular finding or outcome is "true"
|
|
asymptotic
|
tails come closer and closer to horizontal axis but never touch it
|
|
difference between mean and 1 SD
|
34.13% of cases under curve
|
|
z score
|
allows us to compare values from different distributions, compare a sample to population, or compare two samples
|
|
central limit theorem
|
for any trait, regardless of whether it is normally distributed across the population, then a N (the size of the sample drawn from the population) becomes large, the sampling distribution of samples means takes on a normal shape
|
|
chi-square
|
test used to analyze data that comes as frequencies
|
|
r square
|
tells you amount of variation in Y that is explained by X
|
|
multivariate regression
|
allows you to assess the relationship between two variables controlling for other intervening variables
|
|
multicollinearity
|
when two independent variables are correlated
|
|
logit model/logistic regression
|
technique used when dependent variables is binary (0/1 or yes/no) and independent variable is either interval-ratio or binary
|
|
nominal-nominal
|
chi square
|
|
nominal-ordinal
|
chi square
|
|
ordinal-ordinal
|
chi square
|
|
interval ratio-nominal
|
t test (only 2) or ANOVA (more than 2)
|
|
interval ratio-ordinal
|
t test (only 2) or ANOVA (more than 2)
|
|
interval ratio-interval ratio
|
pearson's r
|
|
how do partial slopes help
|
they help us estimate the separate effects of our independent variables on the dependent variable
|
|
informed consent
|
those who are researched should have the right to know that they are being researched and that in some sense they should have actively given their consent
|
|
sugging
|
pretending to conduct an interview that turns into a sales pitch
|
|
content analysis
|
set of methods for systematically coding and analyzing qualitative data
|