Test 3 Flashcards
In a survey every person surveyed is exposed to the same set of questions, same wording, phrasing, no variation
(exposed to standard stimulus) True/false
true
what is a probability sample
a random sample from a larger population
surveys are what? they have no manipulation of any variable
observational
surveys can test knowledge of individual on a certain subject T/F
T
Surveys do what in disease, health, behavior in the larger population, may suggest causal links that cannot be done in experiment
track trends
a survey at a single point in time, but emphasis on past behavior
cross sectional survey
multiple cross sectional surveys over time are called
trend study
Longitudinal studies are done how
over time
Follow subjects into future for specified
period of time (social science research) is what kind of survey
panel survey
Follow subjects into future until a condition develops (epidemiological research) is what type of survey
cohort study
study looks at the exact same people at two different times is what
panel study
identifying subjects with a clinical condition and is retrospective study is called
case controlled study
Trend studies equal or dont equal panel study
dont equal
in surveys it is difficult to prove causes of the effect, not sure if there are other variable to the effect, hard to make strong causal inference
is true or false problems with surveys
True
people tend to do what when reporting effects
under report negative, over report positive
surveys can have poor external validity true or false
true
is a convience sample representative of overall population
no
what type of sample is representative of overall population
probability sample
what are four probability sample types
simple random sample, multicluster sample, stratified sample, random digit dialing
what type of probability sample used when you take table of random numbers or randomly selected using computer
simple random sample
taking random samples of multiple level (cities to counties to towns to households)
multicluster sampling
what is an advantage to multicluster sampling
easily obtainable sampling frames
what is a disadvantage to multicluster sampling
complex sampling design
what probability sample occurs when you over estimate sample in order to get proper results, when a sub group is small in a population
stratified sampling
Computer randomly assigns 4 digits to complete phone number is called
random digit dialing
if its not a probability sample what do you want the researcher to do
give enough info to prove valid results
which are the worst result rate of surveys
what is respondent bias
people who fill out survey may be different than overall population
Can the researcher demonstrate that non-respondents are not substantially different from respondents? example of
respondent bias
jargon slang, technical language, mutually exclusive( each respondent fits in one and only one category) are what
poorly constructed questions
Watch out for survey questions with ambiguity, confusion, vagueness
Watch out for surveys which ask questions that are beyond
R’s capabilities T/f
Both true
what time frame is better for less important events
short time frame
longer time frame is good for what events
important events
looking for a specific answer from a person in the survey-like for a fundraiser is called
leading bias
surveys should never assume what
respondent knowledge
coupling survey with figures and well known people it enhances results-even when there is no such thing is this true/false and what is it called
true, prestige bias
surveys use what two thing which are a set of questions when combined yield info about respondent
scales and indexes
combine a survey with a physical examination increases external validity (these are non mail surveys) True or false
True
according to chronbach’s index what indicates the survey is more reliable
closer to 1
which surveys have highest rate of validity
face to face
can respondents be influenced by interview
yes
cheap and fast, convenient, innovative(imbed pictures etc), have become quite common is what kind of survey
web based
what is major problem with mail and web based surveys
not sure who actually fills them out
which survey must be synchonous
telephone
which surveys are asynchronous
mail and web based
A multi stage cluster sample is a type of probability sample of a larger population used in survey research. In probability samples each case has an equal chance of inclusion in the sample which are true or false
both true
Surveys typically use convenient samples drawn from a larger population. Typically surveys rely on a respondent’s willingness to participate which may affect the validity and accuracy of their survey which is true or/and false
false, true
is info derived from a cross sectional study related to surveys
no
Information derived from cross sectional surveys allow a researcher to make causal statements about the IV and the DV, is this statement about surveys true or false
false
Typically surveys are dependent on willingness of ppl to participate, can mess up validity and accuracy of the responses. Survey samples are typically probability samples drawn from a larger pop
both true
A specific statistic that is calculated to indicate the magnitude (strength) and direction of the degree of relationship between two variables (X, Y) is what
correlation coefficient
a strong correlation is when
when data points are close to trend line
relationship between an interval/ratio dependent variable and an interval/ratio independent variable
bivariate regression
is the best line to minimize error (relationship between two variables)
Regression line, trend line
in multiple regression what does the trend line become
a plane
one dependent variable with multiple independent variables is
multiple regression
How does an independent variable increase the odds (likelihood or probability) of an event occurring, relative to some baseline category?
logit regression
if linear regression line is flat than what
no difference in odds
comparing odds of something happening compared to baseline category
logit regression