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;
118 Cards in this Set
- Front
- Back
empirical analysis
|
how and what we know
|
|
normative analysis
|
how we should use our knowledge
|
|
quantitative
|
statistical comparisons
|
|
qualitative
|
based on informed understanding
|
|
normative
|
developing and examining subjective goals, values, and moral rules to guide us in applying what we learned
|
|
scientific research
|
effective, explicit, systematic, controlled. method of testing theories and hypothesis by applying certain rules of analysis to the observation and interpretation of reality under strictly delineated circumstances
|
|
research process
|
1. theory
2. operationalization 3. select research techniques 4. observe behavior 5. analyze data 6. interpret results |
|
formulation of theory
|
think of a research question. should fulfill a scientific or societal need.
|
|
operationalization
|
put theory into concrete terms so you can measure it. decide what socioeconomic status means - for example, annual income...
|
|
selection of appropriate research technique
|
qppropriateness of a technique depends on teh proglem
|
|
observation of behavior
|
actually doing the research
|
|
generalizability
|
how representative is your sample? how far can you say your findings represent the world?
|
|
reactivity
|
observation should not effect the subject
|
|
hawthorne effect
|
presence of an observer made workers more productive = research was invalid
|
|
analysis of data
|
is there a relationship? how confident and representative are your findings?
|
|
theorizing
|
attempt to create possible explanations for events
|
|
theories
|
sets of logically related symbols that represent what we think happens in the world
|
|
exploratory research
|
background research to understand more
|
|
induction
|
process of generalizing from what we have observed to what we have not observed
|
|
empirically grounded
|
theories built through inductions from observations
|
|
deductions
|
opposite of induction.. moving from generalizations to specifics
|
|
theory construction
|
1. use induction to make assumptions
2. use deduction to make predictions 3. test predictions against observations 4. revise assumptions |
|
theoretical import
|
important in theory to explain
|
|
proposition
|
statement of relationship between concets
|
|
covariation
|
as one goes up or down, the other goes up or down. does not tell why, just that it happens
|
|
causal
|
one directly influences the other.
|
|
testing causation
|
1. must change together
2. cause must precede effect 3. identify causal linkage or process by which it happens 4. covariance must not be due to a 3rd factor |
|
spurious relationship
|
when A and B vary together only because they are both caused by C
|
|
theorizing
|
attempt to create possible explanations for events
|
|
theories
|
sets of logically related symbols that represent what we think happens in the world
|
|
exploratory research
|
background research to understand more
|
|
induction
|
process of generalizing from what we have observed to what we have not observed
|
|
empirically grounded
|
theories built through inductions from observations
|
|
deductions
|
opposite of induction.. moving from generalizations to specifics
|
|
theory construction
|
1. use induction to make assumptions
2. use deduction to make predictions 3. test predictions against observations 4. revise assumptions |
|
theoretical import
|
important in theory to explain
|
|
proposition
|
statement of relationship between concets
|
|
covariation
|
as one goes up or down, the other goes up or down. does not tell why, just that it happens
|
|
causal
|
one directly influences the other.
|
|
testing causation
|
1. must change together
2. cause must precede effect 3. identify causal linkage or process by which it happens 4. covariance must not be due to a 3rd factor |
|
spurious relationship
|
when A and B vary together only because they are both caused by C
|
|
causal model
|
shows all relationships in a theory
|
|
hypothesis
|
a statement of what we believe to be true
|
|
intervening variables
|
link independent and dependent variables
|
|
antecedent variable
|
comes before independent variable.
parents strong party ID leads to child's strong party ID leads to more likely to vote |
|
alternative rival hypothesis
|
another hypothesis that only one can be true. typically a spurious relationship hypothesis can be tested against original hypothesis
|
|
research design
|
the scheme that guides teh process of collecting, analyzing and interpreting data. it is a logical model of proof that allows teh making fo valid causal inferences. what you intend to do, how, and why in this particular way
|
|
exploratory research
|
to learn more in order to make a specific question
FLEXIBLE |
|
descriptive research
|
better explain or describe a phenomenon in order to make a better hypothesis
ACCURATE |
|
explanatory research
|
to test a hypothesis. when we already know a lot about a phenomenon. if you can use the results of the study to argue that one thing causes another.
BASIS FOR CAUSAL INFERENCE |
|
coping with alternative rival hypothesis
|
you can have a control group. show that the difference in the DV between the two groups is the IV.
|
|
adequacy
|
a research design needs to be logically support the inferences that you want to make. you need to look for alternative rival hypotheses.
|
|
rival hypothesis
|
means that both it and the original hypothesis cannot be true
|
|
experimental design
|
experimental group with subjects exposed to the IV stimulus and a control group not exposed to the stimulus
|
|
pretest
|
measure DV before experiment
|
|
posttest
|
measure DV after experiment
|
|
test effect
|
when subjects react to the pretest
|
|
solomon two-control-group research design
|
removes test effect from scores
|
|
solomon three-control-group research design
|
removes changes due to external factors
|
|
randomization
|
helps have similar groups. rules out alternative rival hypotheses that say that differences in groups made the results.
the key to successful laboratory experiments |
|
precision matching
|
identify characteristics that may influence subjects responses and make sure the groups have matched people with the same characteristics
|
|
frequency distribution control
|
make groups with same average characteristics and distribution fo characteristics. same number of males, same average age, more practical than precision matching
|
|
political communication experiments
|
expose subjects to newspaper or TV ads, newscasts
|
|
field experiments and non-experimental designs
|
when you change IV and exposure but not other factors
|
|
quasi-experimental design
|
most common. cannot control Iv exposure or conditions but can gather data or analyze data and stimulate an experimental design. "proceed as if we had exercised all control characteristic of a true experiment and they provide a basis for causal inferences"
|
|
ex-post facto experiment
|
= collect data and analyze. survey randomly to see if college education affects voting behavior. it's as if they set up an experiment years ago and exposed some to college and some not
|
|
time series design p. 110
|
researcher makes several observations before and after the introduction of a causal phenomenon and compares DV before and after.
shows if IV changed the trend |
|
regression toward the mean
|
extreme DV measurements tend to return to teh average value regardless of IV exposure. you can control this by using a controlled time-series design. compare to a data set like ours that is not exposed to the IV (like seeing if a policy worked to reduce crime...)
|
|
research papers
|
empirical but not normative = do not suggest policy initiatives...
|
|
confirm a hypothesis if...
|
1. find covariation between X and Y
2. find covariation was not due to chance 3. disconfirms alternative explanations |
|
quasi experiments
|
only mediocre at deciding on alternative explanations.
|
|
omitted variable bias
|
when z could be related and you ignore it
|
|
intervening variable
|
when X causes Y but it goes through something - how X leads to Y
|
|
spurious relationship
|
when A and B vary together only because they are both related to C
|
|
antecedent variable
|
comes before independent variable. like parents strong party ID - kids stong ID - more likely to vote
|
|
coding
|
process of assigning numerical values to observations
|
|
codes
|
allows the researacher to come to conclusions...numerical manifestations of operational definitions
|
|
nominal
|
categories
|
|
ordinal
|
ranking
|
|
interval ratio
|
dollars, years, numbers.
|
|
codebook
|
listing of variables, values, and codes associated with the experiment
|
|
coding sheet
|
data recording device
|
|
operationalization
|
selecting observable phemonena to represent abstract concepts
|
|
instrumentation
|
specification of steps to take in making observations
|
|
measurement
|
application of an instrument to assign numerical values to cases
|
|
indicator
|
to measure a variable (like height in inches)
|
|
observation
|
process of using a measuring instruement to assign values to a phenomenon in cases
|
|
multidimensional
|
using multiple indicators of a variable
|
|
operational definitions
|
explicitly define valid and reliable measures of concepts studied
|
|
NOIR
|
nominal - ordinal - interval - ratio
|
|
measurement error
|
differences between cases that result because of a flawed measuring process, not because of actual differences.
|
|
systematic errors
|
from confusion of variables of the nature of teh instrument
|
|
random errors
|
due to characteristics/variations by chance
|
|
validity
|
extent to which our measures correspond to the concepts they are intended to reflect. are you measuring what you want to measure?
|
|
internal validity
|
are we measuring what we want to measure?
|
|
external validity
|
generalizablity of our results
|
|
pragmatic validation
|
degree to which a measure allows us to predict behavior and events
|
|
predictive validity
|
the better we are at predicting, the more valid the measure
|
|
construct validation
|
infer validity. determine the extent to which a variety of measures are consistent with what our theory predicts
|
|
discriminant validation
|
whether using it as an indicator of a given condept allows us to distinguish that conept form other concepts
|
|
face validation
|
it may seem valid, but you should do other tests
|
|
reliability
|
how stable the values are. do you consistently get the same answer?
|
|
representative sample
|
reflects teh population
|
|
sample
|
a subset of a population
|
|
sampling bias
|
nonrandom differences between sample and population
|
|
sampling error
|
chance difference between sample statistic and population parameter
|
|
what's the difference between sample bias and sample error?
|
sampling bias is your fault. sampling error is simply due to chance
|
|
simple random sample
|
number population 1-N
choose a sample size n randomly choose n cases |
|
cluster sample
|
divide population into clusters
choose clusters at random |
|
bad sampling techniques:
convenience sampling volunteer sampling quota sampling |
response bias, self-selection bias
|
|
aggregate data
|
individual data - single human beings based on surveys
aggregate data - based on characteristics of an entire group average level of US education |
|
confidence interval
|
40-60%
confidence interval is a concrete measure of sampling error - random chance |
|
confidence interval shrinks...
|
as sample size grows
|
|
confidence interval grows...
|
as level of confidence grows.
49-51% grows as 95% goes to 99% |
|
sampling error
|
due to chance.
even if you have a valid, unbiased measure, you can still have sampling error |
|
bigger sample size means...
|
less sampling error
|
|
52, 48, 52
|
sampling error. not reliable
|
|
70, 70, 70,
|
sampling bias. not valid
|