Common Statistical Tests Flashcards
statistics that summarizes characteristics of a dataset
descriptive statistics
statistics that allows you to test a hypothesis
inferential statistics
study that provides details about who, when, where in relation to what (outcome)
descriptive studies
study that answers why or how; test the hypothesis
analytic study
detailed report of individual patient; typically a novel or unusual case ex. Phineas Gage
case report
detailed report on several individual patients; no comparison group; retro or prospective
case series
study that looks at data at a single point in time
cross-sectional study
study that has unit of measure being a group, not separate individuals; used to understand relationships b/t outcome and exposures at population level
ecological studies
assuming what is true of a population is true for the individual members of that population
ecological fallacy
researcher observes effect of a specific variable
observational study
researcher manipulates conditions and observes effects in controlled setting
experimental study
compare group w/ defined exposure/risk factor to group without exposure; risk factor must be present before disease development
cohort study
ex. individuals who use tanning beds have higher risk of developing skin cancer than those who do not
cohort study
compares group of people w/ disease to group without that disease; retrospective
case-control study
ex. individuals w/ skin cancer had higher odds of indoor tanning in the past than those without skin cancer
case-control study
statistical analysis that combines results of multiple scientific studies
meta-analysis study
this specific trial is prospective and measures effectiveness of a new intervention or treatment
randomized controlled trials (RCTs)
trial that compares effect of a series of greater or equal to 2 treatments on a subject; subject serves as their own control
crossover clinical trials
selecting group from whom you will actually collect data in your research
sampling
type of sampling that is randomization of a simple frame using random number generation
simple
type of sampling that divides a population into groups w/ similar characteristics; then takes random samples from each group
stratified
similar to stratified, except groups are representative of population; not always accurate
cluster
sample every nth person or group
systematic
participants in the control and experimental groups are paired up
matched sampling
characteristics required for participants to be included
inclusion criteria
prevents selection bias in the RCTs by concealing the allocation of participants into groups until the moment of assignment (think match day)
concealed allocation
occurs when some members of pop. are systematically more likely to be selected in a sample than other members
selection bias
when participants leave/withdraw during a study
attrition
bias introduced if characteristics of people lost to follow up differ from those who remain in the study
attrition bias
systematic error when participants do not remember previous events/experiences accurately or omit details
recall bias
bias that deals with error in data collection
measurement bias
participants change behavior b/c they are aware they are being observed
Hawthorne Effect
subjects in diff. groups are not treated the same
procedure bias
a distortion that modifies an association b/t an exposure and an outcome b/c a factor is independently associated w/ the exposure and outcome (ex. ice cream sales and shark attacks)
confounding bias
variable experimenter manipulates
independent variable
what is affected by independent variable
dependent variable
variable that deals with 2 or more groups being measured
categorical
type of categorical variable that deals with descriptive only and no order
nominal
type of categorical variable that deals with the sequence of categories being ordered so we can assign numbers
ordinal
variable that can be quantified as a number
numerical
any number is possible b/t 2 integers
continuous
represent measured quantities of things, allowing for degree of difference b/t two values
interval
whole integers; only some numbers possible
discrete
explores 1 variable in a set at a time
univariate analysis
explores 2 variables in a set at a time
bivariate analysis
explores multiple variables in a set at a time; reduces chance of bias
multivariate analysis
how much variability exists in a set of values, around the mean of these values
standard deviation
these both decrease as sample size increases
standard deviation and standard error
estimate of how much variability exists in a (theoretical) set of a sample means around the true population mean
standard error
measure of the spread of a dataset
variability
diff. b/t highest and lowest value
range
range of the middle half of a distribution
interquartile range
average distance from the mean
standard deviation
average of squared distances from the mean
variance
concept of how the data evens out
regression to the mean
shows data near the mean occur more frequently than data far from mean; probability distribution that is symmetric about the mean
normal distribution
mean of 1 and a standard deviation of 1
standard normal distribution
extreme values at right end
positive skew
extreme values at left end
negative skew
occurrence of new cases
incidence
number of existing cases
prevalence