BBH 310 exam 3 || with Error-free Solutions.
hypothesis correct answers what is being tested
theory correct answers really broad- framework in which you situate your hypothesis- guides
thinking
regular causality correct answers A causes B
Reverse causality correct answers B causes A
Bidirectional causality correct answers A causes B and B causes A
Third variable causality correct answers C causes A and B
Case study correct answers one person with rare disease. super interesting so it makes the cost
worth it
clinical trial correct answers longitudinal experimental design always on humans
lab experiments correct answers longitudinal- involves animals
cohort correct answers participants come from same population (all have something in common)
cross-sectional correct answers recruit and test them only ONCE
case-control correct answers 2 groups (1 with disease and 1 without) look at how risk factors
differ retrospectively. what was different in the 2 groups and why did they get the disease? Look
at behavior, family history etc. Key is to recruit 2 groups, 1 group with disease
External validity correct answers all about generalizability- generalize to population at large.
Representative to population
T/F: observational studies usually have better external validity correct answers true
internal validity correct answers "can you infer causality"? if there is random assignment, and
you can control all other extraneous factors, then you can infer causality from your study
T/F: Experimental studies usually have better internal validity correct answers True
Extraneous correct answers potential confounds- any variables that you are not explicitly
watching out for your hypothesis
Between subject designs correct answers comparing 2 groups of people
, within subject designs correct answers looking at change in people. for example, looking at the
same classroom
What is the correct order?
A. population, sample, sampling frame
B. sample, sampling frame, population
C. Population, sampling frame, sample correct answers C
population correct answers everybody you are interested in
sample correct answers the people you choose to study
sampling frame correct answers all of the potential people
convenience sampling correct answers put up an ad, it is a free for all. you are asking people to
participate so you take anybody that meets your criteria
stratified random sampling correct answers specific probability at which you recruit people into
your study. probability of being selected depends on which group you are in. you are recruiting
from different groups
simple random sampling correct answers specific probability at which you recruit people into
your study. everybody has the same probability of being selected
random sampling correct answers first recruiting people at certain probability- observational
random assignment correct answers once they are in the study and have signed up to be in it,
(experimental), random assignment, they go into different groups
Random error correct answers not clear --> can be higher or lower
Systematic error correct answers always leans in one direction (either higher OR lower)
-machine calibrated incorrectly
Convergent validity correct answers comparing your test to a different test. test you're comparing
to is just another test that is supposed to get at the same thing but is also not perfect like yours
gold standard correct answers comparing your results to known answers that are correct/perfect
inter-item reliability correct answers comparing individual items of a questionnaire to one
another- consistency of items in the same survey
test retest reliability correct answers all about consistency over time. test the same test multiple
times
hypothesis correct answers what is being tested
theory correct answers really broad- framework in which you situate your hypothesis- guides
thinking
regular causality correct answers A causes B
Reverse causality correct answers B causes A
Bidirectional causality correct answers A causes B and B causes A
Third variable causality correct answers C causes A and B
Case study correct answers one person with rare disease. super interesting so it makes the cost
worth it
clinical trial correct answers longitudinal experimental design always on humans
lab experiments correct answers longitudinal- involves animals
cohort correct answers participants come from same population (all have something in common)
cross-sectional correct answers recruit and test them only ONCE
case-control correct answers 2 groups (1 with disease and 1 without) look at how risk factors
differ retrospectively. what was different in the 2 groups and why did they get the disease? Look
at behavior, family history etc. Key is to recruit 2 groups, 1 group with disease
External validity correct answers all about generalizability- generalize to population at large.
Representative to population
T/F: observational studies usually have better external validity correct answers true
internal validity correct answers "can you infer causality"? if there is random assignment, and
you can control all other extraneous factors, then you can infer causality from your study
T/F: Experimental studies usually have better internal validity correct answers True
Extraneous correct answers potential confounds- any variables that you are not explicitly
watching out for your hypothesis
Between subject designs correct answers comparing 2 groups of people
, within subject designs correct answers looking at change in people. for example, looking at the
same classroom
What is the correct order?
A. population, sample, sampling frame
B. sample, sampling frame, population
C. Population, sampling frame, sample correct answers C
population correct answers everybody you are interested in
sample correct answers the people you choose to study
sampling frame correct answers all of the potential people
convenience sampling correct answers put up an ad, it is a free for all. you are asking people to
participate so you take anybody that meets your criteria
stratified random sampling correct answers specific probability at which you recruit people into
your study. probability of being selected depends on which group you are in. you are recruiting
from different groups
simple random sampling correct answers specific probability at which you recruit people into
your study. everybody has the same probability of being selected
random sampling correct answers first recruiting people at certain probability- observational
random assignment correct answers once they are in the study and have signed up to be in it,
(experimental), random assignment, they go into different groups
Random error correct answers not clear --> can be higher or lower
Systematic error correct answers always leans in one direction (either higher OR lower)
-machine calibrated incorrectly
Convergent validity correct answers comparing your test to a different test. test you're comparing
to is just another test that is supposed to get at the same thing but is also not perfect like yours
gold standard correct answers comparing your results to known answers that are correct/perfect
inter-item reliability correct answers comparing individual items of a questionnaire to one
another- consistency of items in the same survey
test retest reliability correct answers all about consistency over time. test the same test multiple
times
