standard error and confidence intervals Flashcards
what is standard error
the standard deviation of the distribution of sample meaans measures the spread oif popualtion mean
where does 95% of a sample lies
within the SD
how can uf-us be small but the t test ratio is large
due to small SE
what are the three assumptions of a t test (unpaired)
- the two samples are independant
- the two samples follow normal distribution
- the population SDs are equal
what does an unpaired t etst assume
that the population is normal
what does a paired test assume
that their are differnces in normlaiity
how much of the popualtion lies between the two SDs
95%
what is the reaonsable range to see if the mean is within the 95%
(-2SE,2SE)
what are the 3 steps for setting up a t test
- make a null hypothesis
- set the signifcance level
- calcaulting the chance of falling between -2SE, 2SE
what is the short hand t test stat
ms-mf/se
what would using an unpaired t test do when they are meant to be paired
increase SE and P value and get a smaller T test statistic
if there is a significant assocaition from a t test where would you expect to see mf-ms/se
outside of -1.96 and 1.96
how are vairiables treated in regression analysis
not equally
how are variables treated in correltaion analysis
equally
are residual considered normal
yes
what coreeation coefficient measure
the strenght of the relationship between two variables
what is required in correaltion
for y or x to be normla
what does it assume when you find a confidence interval for r
that y and x are normal
