PUBH5018 Flashcards
if distribution is skewed, what measure of central tendency shd you use?
median (and IQR)
if distribution is symmetrical, what measure of central tendency shd you use?
mean (and SD)
the bigger your sample size, the narrower your…what?
margin of error (confidence interval) and the more precise your estimate
the 2 ways of drawing inference from a sample are:
including confidence intervals and doing hypothesis tests
a normal distribution (not skewed) is defined by the
mean (the centre of the distribution) and SD (how spread out it is)
what is a z value?
standardised normal deviation in a normal distribution
what is an n value?
sample size
what is an x value?
variable
a low p value means the strength of evidence for rejecting the null hypothesis is..
strong (there is a likely effect)
a p value above 0.05 (5%) means that…
statistically not significant (if yr dichotomising)
which measure of uncertainty demonstrates practical (not statistical) importance?
confidence intervals (not p values)
if yr sample size is small, what would you expect of your CI?
very large (no precision)
if the CI contains a value of 0, the effect is what?
not significant
what is standard deviation?
a measure of how spread out the datapoints are - a ‘standard’ difference from the mean
what is a confidence interval?
range of possible values (effects) around a mean (best possible estimate)
what is a z score?
the number of standard deviations that a value (x) is from the mean - use it look up in Table F1
what is an x bar value?
sample mean
what’s the difference btwn categorical or continuous variables?
categorical - in groups (sex, diagnosis, a scale)
continuous - keep going (height, blood pressure)
what’s the difference btwn sampling and non-sampling errors?
sampling error - sample not large enough
non-sampling error - sample not representative enough
what’s the difference btwn simple random sampling and stratified random sampling?
simple - purely randomised
stratified - pop initially divided into groups and then randomised