Confidence intervals for correlations -------------------------------------
^ci2^ varname1 varname2 [^if^ exp] [^in^ range] [weight] [^, COrr SPear > man^]
^cii2^ #n #corr [^, corr^]
Description -----------
Without the ^corr^ or ^spear^ options, ^ci2^ and ^cii2^ behave as @ci@ and @cii > @.
With option ^corr^, ^ci2^ calculates the Pearson product moment correlation & produces a confidence interval, based on Fisher's transformation. As with @correlate@, ^ci2^ takes frequency and analytic weights.
With option ^spearman^, (with or without ^corr^), Spearman's rank correlation i > s used.
^cii2^ is the immediate version. There is no separate Spearman option, because > the method for finding the correlation is the same in both cases.
Examples -------- Use for correlations . ^ci2 colorime icp , corr^
. ^ci2 colorime icp , corr spearman^ . ^ci2 colorime icp , spearman^ (Note: these last two are effectively the same.)
. ^cii2 78 .52 , corr^ (78 paired observations, correlation 0. > 52)
Use as ^ci^ and ^cii^ . ^ci2 mpg^ . ^ci2 mpg price, level(90)^ . ^ci2 promoted, binomial^ . ^ci2 count, poisson^ . ^ci2 deaths, exposure(pyears)^
. ^cii2 166 19509 4379^ (166 obs, mean=19509, sd=4379) . ^cii2 166 19509 4379, level(90)^ . ^cii2 10 1^ (10 binomial events, 1 observed succes > s) . ^cii2 1 27, poisson^ (27 Poisson events observed)
Also see --------
Manual: [R] correlate On-line: @correlate@, @pwcorr@