Please enable JavaScript.
Coggle requires JavaScript to display documents.
SAS chp 1-descriptive stats - Coggle Diagram
SAS chp 1-descriptive stats
sample is a representative of population so that we can infer about the population from the sample
cant use population to test because a population may be infinite, so its impossible to test
standard deviation: how
spread out
a set of data is
usually used for a normally distributed data, or a bell curve(most common distribution, core of all statistics
part of methods used to measure spread: range, interquartile range, standard error etc.
range is easily affected by extreme values
interquartile range value is substraction of 1st quartile value(25%ile) from 3rd quartile value(75%ile)
standard error-the standard deviation of multiple means
eg. 4 sample means grouped together, then averaging them(produce another mean)
mean of the means, grand means
very close to true population mean
why so important? because of the central limits theorem!
68% of data points fall under +/- 1 SD of the mean
percentage is the confidence interval
formula(for 95% etc): mean+/- 2x(SD)
the more confidence u want
the larger the range
, but no 100%
to find confidence interval of average data, instead of using SD u use SE(standard error)
96%~ +/-2SD
99.8%~ +/- 3SD
central limits theorem conveys that as
sample size gets larger,
the distribution of sample means from any number of samples will be
normally distributed
how to calculate SE:divide SD by sq root of sample size
on excel: stdv.p(...)/sqrt(count(range))
if calculate for SE of sample, use stdv.s
variance is the sq of standard deviation
types of data: nominal, ordinal, interval, ratio
nominal-cannot be averaged
discrete
ordinal-can be arranged, can still do some averaging
discrete
interval-range
continuous
ratio-every value has different meaning to it
continuous
mean, median mode are methods to measure central tendency-the usual value