Chapter 1 Flashcards
generalization
only appropriate when based on samples that are:
-randomly selected or otherwise representative of a pop
-selected from that pop specifically
confounding factors
factors that, when analyzing data, are hard to parse out if the cause is one, both/all, or none of them
observational study
-taken w/o imposing treatments and a cause and effect canNOT be inferred
-observes characteristics of a sample and then generalizes to the pop
-does NOT determine who will be in what group
experiment
-treatments are imposed and a cause and effect relation can be inferred
-a response variable behaves under differential conditions
-determines who will be in what group and what the experimental conditions will be like
population characteristic
a # that describes the ENTIRE pop
statistic
a # that describes a SAMPLE
census
collects data from all individuals in a pop
simple random sample
sample of size n is selected from the pop in a way that ensures that every different possible sample of the desired size has the same chance of being selected
variable ānā
used to denote sample size; # of individual or objects in the sample
sampling frame
a list of all individuals/objects in the sample
sampling with replacment
sampling in which an individual/object, once selected, is put back into the sample before the next selection
sampling without replacement
sampling in which an individual/object, once selected, is removed from the sample before the next selection
when can you treat sampling with replacement as the same as sampling without replacement
when sample size ānā is relatively small compared to the pop size (no more than 10% of pop)
convenience sampling
selecting individuals/objects that are easy/convenient to sample
voluntary response
a type of convenience sampling which relies solely on individuals volunteering to be part of a study (ppl motivated to respond are often HIGHLY motivated and respond with extremes)
variable āNā
used to denote an entire population
cluster random sample
when a pop is divided into cluster, and an simple random sample of clusters is taken, with all individuals within each cluster sampled
-grouping is ideally heterogeneous
stratified random sample
pop is divided into strata based on a similar characteristic, with a simple random sample being taken with in each stratum and all selected individuals combined into larger sample
-strata are ideally homogeneous
systematic random sample
randomly chose a start point, and the sample at a fixed periodic interval based off said start point
-aka 1-in-k systematic sample
-k=N/n
variation
a measure of precision
bias
the tendency for samples to differ from the corresponding pop in some systematic way
undercoverage bias
when part of a pop has a reduced chance of being included in a sample
voluntary response bias
occurs when an invitation is sent to all individuals in a pop to participate and those who choose to participate may differ from individuals who dont chose to participate
question wording bias
when survey questions are confusing and/or leading
self-reported response bias
when individuals inaccurately report their own traits (often happens in embarrassing or incriminating situations)
selection bias
when samples are taken from a source where some groups are excluded
measurement/response bias
when the method of observation tends to produce values that systematically differ from the true value in some way
-ex. improperly calibrated scale
-ex. wording in survey
-could occur from appearance/behaviors of those giving the survey
when does an increase in sample size NOT reduce effects of bias in a study?
when selection method is flawed and when high nonresponse
nonresponse bias
when responses are NOT obtained from all individuals selected for inclusion in the sample
confounding variable
another variable that is related to the explanatory variable and influences the response variable
-may create a false perception of association
what does a well designed experiment include?
1) comparison of 2+ treatment groups, with one control
2)random assignment/allotment of treatments to experimental units
3) replication
4) control of potential confounding variables, where appropriate
single blind
where subjects in a study do NOT know which treatment they are receiving, but researches do (or vise-versa)
double blind
neither subjects nor the researchers interacting with eh subjects are aware of the treatments being administered
response variable
variable not controlled by experimenter, and is measured as part of the experiment
explanatory variable
variables that have values that are controlled by the experimenters
(AKA factors)
experimental conditions
any particular combination of the explanatory variables
(AKA treatments)
experimental units
the smallest unit to which a treatment is applied
(AKA subjects)
direct control
holding the other variables constant so that their effects are not confounded with those of the treatments
replication
ensures that there is an adequate # of observations for each experimental condition
statistical significance
that observed changes/differences occurred NOT merely by chance
-evidence that a treatment caused an effect
variable āpā
the probability (chance) of observing a specific outcome due to chance variation
-how likely an outcome is to occur
completely randomized design
an experiment in which experimental units are randomly assigned to treatments
randomized block design
an experiment that incorporates blocking by dividing the experimental units into homogeneous blocks and then randomly assigning the individuals within each block
control group
an experimental group that does NOT receive any treatment (AKA baseline)
-allows to assess how the response variable behaves when treatment is NOT used
placebo
something that is identical to the treatment received, except it contains NO active ingredient(s)
COMMON MISTAKES
1) drawing cause and effect from an observational study
2) generalizing results of an experiment that uses volunteers as subjects (only valid if provable that volunteers are representative)
3) generalizing data based on data from a poorly designed study
4) generalizing conclusions based on an study that used voluntary response/convenience sampling to a larger pop
