[tor-talk] mildly updated statistical analysis of Tor metrics data

[stn]

hi virgil
my skills are old but i might be of use.

are you using a single dependent/outcome variable (eg multiple 
regression) model or are what i think of as a "real" multivariate model? 
(multiple independent variables and multiple dependent variables) ?

if multiple regression i can likely be of use quickly.  if the latter 
... i'd have to refresh my memory and use a suggested reference text.

is the dataset small enough to use on a desktop stats package or does it 
require something beyond a typical desktop?


On 15/09/2014 8:57 PM, Virgil Griffith wrote:
> I just realized there's a better way to control for some of these
> effects using multivariate models.  Another update will come later.
>
> -V