STAT 121 Commonly Tested Questions Verified Answers
What are the good samples for probability sampling design and what makes them good?
-They all use some random device to select
-Representative of the population
-Uncertainty can be assessed
-Ex: Simple Random Sampling (SRS), Cluster, Stratified Random Sampling, and multistage
sampling.
Define Simple Random Sampling
(SRS) Taking SOME individuals from the population. Equivalent of selecting names out of a hat.
Every possible option has an equal chance of being selected
Define Cluster Sampling
ALL individuals from SOME groups. A random sample of clusters (groups that population are
naturally divided into) is taken and all the individuals in the selected clusters are included in
the sample.
Define Stratified Random Sampling
SOME from ALL (or EVERY) group. Select a SRS (simple random sample) from each stratum
(groups that share a similar characteristic, ex; gender), then combine all together
Multistage Sampling
SOME from SOME groups. Takes sample at each hierarchical level of the population.
Define Non-Sampling Bias and give examples
-Other sources of error (not measurable, so must be eliminated or minimized)
Undercoverage
Non-Response
,Misleading
Interviewer
Question Order
Open and Closed Questions
Question Wording
Define Undercoverage
no possibility to be selected for the sample. Ex-homeless or no phone
Define Non-Response
Individuals choose not to be in the sample because they refuse to provide information
or cannot be contacted Ex-Too long of a survey and so some refuse to complete it
Define Misleading Response
People lie/inaccurate answers (sensitive information)
Interviewer Bias
Person asking questions influences responses
Question Order Bias
The order that questions are asked promotes certain responses. Ex-debt before levels
of happiness or happiness before debt question
Open Questions & Closed Questions in Sample Surveys
Open: unlimited responses, short answer, more difficult to analyze Closed;
limit responses, MC, easier to analyze, should include other/unsure
Question Wording
The way a question is asked may lead on to an answer, mislead, or confuse
Observational Studies
,Study where individuals are NOT assigned to treatments, instead they self-
select. -Observe or measure variables from undisturbed subjects -No cause and
effect conclusions
-Lurking variables are confounded with treatments
Experiment
Study where individuals are ASSIGNED to
treatments -Treatments imposed on subjects by
researchers -Can be used to establish conclusions -
No confounding lurking variables
Response Variable
What we want to measure. Characteristic measured n each subject, what we really want
to know.
Explanatory Variable
Used to predict or explain changes in response variable.
-WHAT IS BEING MANIPULATED BY THE RESEARCHER
Lurking variable
A variable that has an important effect on the relationship among the variables in a study
but was not measured or included as a planned factor in the study
Control
Reduces effects of lurking variables (Ex-placebo)
Confounding Variable
When effects of lurking variable cannot be distinguished from effects of treatment
What 3 things must valid experiments incorporate?
, control/comparison
-randomization
-replication
*double blinding is good if possible but not required
-valid experiments allow us to make cause and effect conclusions
Randomization
Randomly assign subjects. Reduce effects of lurking variable
Control/Comparison
two or more groups. Control lurking variables by including comparison treatments
Replication
Two or more subjects in each group. Assign more than one subject to each treatment group
Placebo Effect
Response due to psychological effect of thinking you're being treated. This psycho effect is
counfounding lurking
Diagnostic Bias
diagnosis of subjects biased by preconceived notions about effectiveness of treatment
Lack of realism
realism is compromised by condition of the study (study set-up is not real life)
Hawthorne Effect
Behavior of subjects is not like real life
Non-Compliance
Subjects fail to submit assigned treatment or refuse
Randomized Controlled Experiments
What are the good samples for probability sampling design and what makes them good?
-They all use some random device to select
-Representative of the population
-Uncertainty can be assessed
-Ex: Simple Random Sampling (SRS), Cluster, Stratified Random Sampling, and multistage
sampling.
Define Simple Random Sampling
(SRS) Taking SOME individuals from the population. Equivalent of selecting names out of a hat.
Every possible option has an equal chance of being selected
Define Cluster Sampling
ALL individuals from SOME groups. A random sample of clusters (groups that population are
naturally divided into) is taken and all the individuals in the selected clusters are included in
the sample.
Define Stratified Random Sampling
SOME from ALL (or EVERY) group. Select a SRS (simple random sample) from each stratum
(groups that share a similar characteristic, ex; gender), then combine all together
Multistage Sampling
SOME from SOME groups. Takes sample at each hierarchical level of the population.
Define Non-Sampling Bias and give examples
-Other sources of error (not measurable, so must be eliminated or minimized)
Undercoverage
Non-Response
,Misleading
Interviewer
Question Order
Open and Closed Questions
Question Wording
Define Undercoverage
no possibility to be selected for the sample. Ex-homeless or no phone
Define Non-Response
Individuals choose not to be in the sample because they refuse to provide information
or cannot be contacted Ex-Too long of a survey and so some refuse to complete it
Define Misleading Response
People lie/inaccurate answers (sensitive information)
Interviewer Bias
Person asking questions influences responses
Question Order Bias
The order that questions are asked promotes certain responses. Ex-debt before levels
of happiness or happiness before debt question
Open Questions & Closed Questions in Sample Surveys
Open: unlimited responses, short answer, more difficult to analyze Closed;
limit responses, MC, easier to analyze, should include other/unsure
Question Wording
The way a question is asked may lead on to an answer, mislead, or confuse
Observational Studies
,Study where individuals are NOT assigned to treatments, instead they self-
select. -Observe or measure variables from undisturbed subjects -No cause and
effect conclusions
-Lurking variables are confounded with treatments
Experiment
Study where individuals are ASSIGNED to
treatments -Treatments imposed on subjects by
researchers -Can be used to establish conclusions -
No confounding lurking variables
Response Variable
What we want to measure. Characteristic measured n each subject, what we really want
to know.
Explanatory Variable
Used to predict or explain changes in response variable.
-WHAT IS BEING MANIPULATED BY THE RESEARCHER
Lurking variable
A variable that has an important effect on the relationship among the variables in a study
but was not measured or included as a planned factor in the study
Control
Reduces effects of lurking variables (Ex-placebo)
Confounding Variable
When effects of lurking variable cannot be distinguished from effects of treatment
What 3 things must valid experiments incorporate?
, control/comparison
-randomization
-replication
*double blinding is good if possible but not required
-valid experiments allow us to make cause and effect conclusions
Randomization
Randomly assign subjects. Reduce effects of lurking variable
Control/Comparison
two or more groups. Control lurking variables by including comparison treatments
Replication
Two or more subjects in each group. Assign more than one subject to each treatment group
Placebo Effect
Response due to psychological effect of thinking you're being treated. This psycho effect is
counfounding lurking
Diagnostic Bias
diagnosis of subjects biased by preconceived notions about effectiveness of treatment
Lack of realism
realism is compromised by condition of the study (study set-up is not real life)
Hawthorne Effect
Behavior of subjects is not like real life
Non-Compliance
Subjects fail to submit assigned treatment or refuse
Randomized Controlled Experiments
