Producing Data Flashcards
observational study
observe individuals and measuring variables of interest but not attempting to influence the response
experiment
deliberately imposing some treatment on individuals in order to observe their responses. An experiment can help eliminate (or at least try to minimize the effects of) lurking variables
population
entire group of individuals we are interest in
census
involves using every member of a population for an observational study
sample
subgroup of population that is being studied
sampling
studying a part in order to gain information about the whole
sampling error/variability
natural variation one would expect to see in sample statistics from sample to sample
voluntary response sample
people who choose to be part of a sample by responding to a general appeal (biased)
convenience sample
choosing the most convenient individuals from the population for your sample (biased)
Bias
occurs when the sampling method systematically favors certain outcomes
simple random sample (srs)
a sample of size n is selected in such a way if every individual in the population has an equal chance of being selected and every subset of individuals has an equal chance of being selected for the sample (unbiased)
stratified random sample
divide the population into groups of similar individuals called strata, choose a separate SRS from each stratum, and then combine all of those individuals chosen from all of the strata to make up the full sample (unbiased)
cluster sample
divide the population into groups (or clusters), then randomly select some of these clusters (completely ignoring the others). All the individuals from chosen clusters are selected to be in the sample
multistage sample design
select successively smaller groups within the population in stages, resulting in a sample consisting of clusters of individuals. Each stage may employ an SRS, or another type of sample
systematic sample
sampling with a pattern (inspecting every fifth bag of potato chips coming off an assembly line)
probability sample
each member of the population has a known chance (greater than zero) to be selected
undercoverage
some groups in population are left out of process of choosing sample, population is not same group as sampling frame
non response
an individual is chosen for the sample but cannot be contacted or refuses to cooperate
response bias
individual in sample chooses an answer to a survey that they thinks is best rather than the answer that they truly believe
wording of question
leading questions cause individual being questioned to choose one answer as opposed to another
sampling frame
group from which sample is chosen
what must you do before you trust a poll result?
insist on knowing the exact questions asked, the rate of nonresponse, and the date and method of survey
are larger samples or smaller samples better? why?
larger random samples are better because they give more accurate results that smaller samples because they decrease variability
experimental units
individuals on which the experiment is being done (if units are people they are called subjects)
treatment
specific experimental condition applied to units. level of treatment is measured by explanatory variable and level of variable we’re interested in is measured by response variable
factors
the explanatory variables in an experiment (experiments may have several factors)
3 basic principles of experimental design
control, randomization, replication (all three principles must be present to be considered a well-designed experiment)
control
controlling the effect of lurking variables on the response variable
control group
group in experiment not given treatment to compare to afterwards
randomization
ensures that individuals are assigned to treatment groups by chance
replication
each treatment applied to multiple experimental units (ensures that results are not just due to chance variation)
randomized comparative experiment
an experiment with random assignment and a control/comparison group
statistically significant
an observed effect too large to attribute plausibly to chance
placebo effect
a subject receiving a placebo reacts favorably to it (even though there is no medicinal benefit in the pill, for example) this is a psychological response
block
group of experimental units known before the experiment to be similar in some way that is expected to affect the response to treatments (homogeneous group)
matched pairs design
special type of block design
blind
subjects involved in study do not know which treatment they are receiving
double-blind
neither the researchers (people measuring the response variable) or subjects know who is receiving which treatment but the experimenters (people running the experiment) are aware of who is given what
lack of realism
most serious potential weakness of experiments
quote about how to set up groups
“control what you can, block on what you can’t control, and randomize the rest”
matched pairs experiment
one of two set ups:
- subjects matched with themselves and given two different treatments (in a random order). this is more commonly used form of matched pairs design
- similar subjects matched and each assigned a different treatment
