Karteikarten
stats-test one
Karten in diesem Stapel (64)
- 1descriptive inferenceusing observations from the world to learn about characteristics of populations
- 2causal inferenceusing observations about the world to learn about causal relationships
- 3null hypothesiswhat you think you'd observe if the theory was wrong
- 4name the five steps of the scientific method (in order)1.causal question
2. theory
3.hypothesis
4.test against data
5. evaluation - 5what is the independent variable?cause
- 6what does it mean to say correlation doesnt equal causation?jdfja
- 7what is the dependent variable?the effect or outcome
- 8what is a time series variation?spatial dimension is constant for all observations represent time units
ex. minutes, years etc - 9what is a cross section variation?time dimension is constant for all observations represent spatial units
ex.countries,people etc - 10what are the components of a good theory?1. has real world significance
2. novel: help us look at things in a new way
3. generality
4. parsimonious
5. causal mechanism
6. contain empirical claims, not normative
7. falsifiable - 11tautologysomething that is true by definition
- 12unobservable elementsno way to prove it is wrong
- 13why can't we prove theories?always potentially another test that could refute it
- 14is poli sci a real science?if you develop and test falsifiable statements= a science
(clark and golder) - 15"God exists" tautology or unobservable?unobservable element
- 16"democracies are more likely than non-democracies to hold free and fair elections." tautology or unobservable?tautology
- 17We should reject every theory that fails empirical tests. true or false?false:
1. test may invalid
2. falsification could have been due to chance
3. a theory may apply in some conditions but not these present during the test - 18to say that X causes Y is to say that X generates change in Y. true or false?true
- 19individual treatment effectthe experimental/treatment condition
- 20average treatment effect-mean impact of change in x on y across the population
-basically average impact of experimental condition
ex. in tutorial average guesses of african countries in the UN was 70 percent - 21what is a counterfactual?means counter to facts or reality
- 22What is the problem with counterfactuals?we cannot go back and change it to see if its true
- 23ceteris paribus conditionmeaning all else equal or all else constant
- 24deterministicif x occurs y is certain
- 25probabilisticif we see a change in x it changes the probability of Y occuring
- 26"Going to Queens does guarantee success but raises your chances of it." deterministic or probabilistic?probabilistic
- 27"democracies never go to war with one another" probabilistic or deterministic?deterministic
- 28What are the four hurdles to establishing causal relationships?1. logical causal connection
2. reverse causality
3. covariation between X and Y
4. Confounding variables - 29what are the consequences of ignoring confounding variables?1.could make it seem like a relationship exists where it doesnt
2.hide a correlation between 2 variables where one does exist
3. reverse direction of the relationship (seem + when its - and vice versa) - 30omitted variable biasconfounding or omitted variables influence both X and Y causing a spurious relationship between them
- 31What is an example of a spurios relationship1.correlation between shark attacks and ice cream sales
2. global warming and pirates - 32what is operationalization?turning the concepts in your causal theory into measurable variables
- 33What does reliability mean?repeatable or consistent
- 34What does valid mean?a valid measure is one that measures what it claims/is supposed to
- 35what are the four types of validity?1.face validity
2. criterion validity
3. content validity
4. construct validity - 36face validitywhether or not on its face/as it appears, it measures what it is supposed to
- 37content validitywhether or not this measure captures all the basic components of this concept or theory
- 38construct validityhow empirically well your indicator relates to another indicator as the theory would predict
- 39is it possible to have a reliable but invalid measure?yes
- 40systematic measurement errormeans that estimates are systematically too high or too low
-invalid - 41random measurement errorswhich means that on average, estimates are off
-unreliable - 42dataseta collection of data
- 43datuma singular piece of data
- 44unit of analysiswhat or whom that is under study
- 45what are the four levels of measurement?1.nominal
2.ordinal
3.interval
4. ratio - 46what is a nominal variable? give examples.there are differences between attributes but they cannot be objectively ranked
ex. religion, hair colour, gender - 47what is an ordinal variable? give examples.there are differences between attributes, they can be ranked, but we dont know that the difference between each is equal
ex. true, mostly true, mostly false, false - 48what are interval variables? give examples.there are differences between attributes, they can be ranked, and there IS equal difference between adjacent attributes
ex. IQ, celsius - 49what are ratio variables? give examples.there are differences between attributes, they can be ranked, and there IS equal difference between adjacent attributes and have a true zero point
ex. age - 50what does continuous mean? give example.can take on any value in a given interval
ex. time in minutes: 2.5 minutes - 51what does discrete mean? give example.can only take on particular values; values in an interval are finite or countable
AKA whole numbers
ex. age in years - 52what is an interactive/conditional effect?the size of the effect of one independent variable is conditional on the value of another independent variable
- 53what are the differences between experiments and observational designs?1.random assignment in experiments and not in observational
2. who assigns it: in experiments its the researcher;in observations its nature or the participants - 54what does it mean to randomly assign?wherein every participant has equal probability of receiving treatment
-treatment and control groups should be comparable across gender,race etc
-difference between them should be due to treatment - 55why does random assignment only work well in large samples?-in smaller sample sizes it becomes less likely that the sample will be representative of the population
- 56random samplingeveryone in the population has an equal opportunity of being recruited for the study
- 57what is the difference between random assignment and random sampling1.random assignment is when a person has equal opportunity of being in either the experimental or the control condition
vs
2. random sampling is when everyone in the population has equal opportunity ti be recutied to your study - 58what is the point of random assignment?-ensures that there is no systematic differences between each group that may skew results
- 59internal validity?refers to our confidence that the results of the experiment are truly a product of the treatment
- 60external validity?refers to the confidence that results/findings are generalizeable
- 61what does attrition mean? why is it a problem?a) when participants fail to respond to questionnaires or drop out of the study
b) there may be systematic similarities between those who dropped out that may skew results - 62threats to external validity?1) contrived environment, experiments may not be a true reflection of reality
2) subjects, university students seem to be the most frequently used because they are most convenient and accessible but there are systmatic differences between them and the rest of the population - 63what are some upsides of experiments?1. dealing with confounding variables
2. guarding against reverse causality - 64what are some disadvantages of experiments?1. cant always determine which IV is most important
2. cant randomly assign everything
3. problems of external validity
4.ethical concerns
Informationen
- Beschreibung
- Karteikarten am stats-test one, erstellt von n sans am 04/02/2017.
- Erstellt von
- n sans
- Aufrufe
- 0
- Bewertung