Week 4 Flashcards
observational study designs
investigator measures or records events but does not intervene
descriptive study designs
ecological and case study designs
analytical study designs
cross sectional
case control
cohort
ecological studies
at least one variable (exposure/outcome) measured at group not individual level
S ecological studies
useful for generating hypotheses
compare same population at different times or at different places in same time
L ecological studies
ecological fallacy
associations not causations
link between exposure and effect at the individual level can’t be made
ecological fallacy
inappropriate conclusions are made about individuals from aggregated group data
cross sectional
defined population is observed at a single point in time or time interval and exposure and outcome are determined simultaneously
S cross sectional
data on all variables collected only once
can investigate multiple exposures at once
measures prevalence
L cross sectional
unable to measure incidence, only prevalence
unable to determine temporality
not suitable for rare diseases
difficult to identify causation
case control
two samples are selected, one case one control
look back and determine how many from each group have the risk factor
retrospective
S case control
investigates causes of disease (especially rare disease)
easy, fast, cheap
L case control
identifying and enrolling control subjects can be difficult
increased probability of recall bias, selection bias and measurement error since retrospective
prospective cohort study
start with people free from disease who are exposed/unexposed to factors
baseline data collected
see how many people develop the disease in each group
retrospective cohort study
investigators jump back in time to identify a useful cohort which was initially free of disease and at risk
then use whatever records are available to determine each subject’s exposure status at beginning of observation period and ascertain what happened to the subjects in these exposure groups
S prospective cohort study
identify temporality
good for rare exposures and common outcomes
L prospective cohort study
selection bias
loss to follow up
insufficient for rare diseases
double dummy approach
participants receive both active drug and placebo at different times in study
field trials
aimed at disease free people in population but are at risk
aim to evaluate interventions and prevention to reduce specific exposures
community trials
community or clusters are defined by treatment groups and examined
allocation concealment
those responsible for recruiting participants must not know what groups they will be allocated to; 3rd party required
smaller the SE
the more precisely the population mean is being estimated
SE AKA
value for the spread for the sampling distribution
df
number of independent pieces of information available to estimate another piece of information
CLT
For a large sample size, the distribution of the sample mean is normally distributed, even when the population distribution from which the sample has been drawn is decidedly non-normal, with mean equal to the true mean of the sampled population and the SD equal to the SE of the sample mean
sampling- unknown true SD
use t score and t table
sampling- known true SD
use z score and normal distribution table
pyramid of evidence from best to worst
meta analysis RCT cohort study case control cross sectional ecological case series case report
if describing the spread of observations in a study
use SD
if using a sample to infer the population results
use SE
If you draw a large sample from this population what should be the shape of this variable (Y) in the sample
same as sample population
For large sample what will be the shape of the sampling distribution for the sample mean
normally distributed