Re: How to discover the effect of X1, X2, ... , Xn on Y (Y is dependant on the X variables)?

From: David Jones (dajxxx_at_ceh.ac.uk)
Date: 12/06/04 

Date: Mon, 6 Dec 2004 11:03:31 -0000

Richard Ulrich wrote:
> On Mon, 6 Dec 2004 00:16:41 +0200, "Roland" <roland@nospam> wrote:
>
>> I have a variable Y which depends on variables X1, X2, ... , Xn. I
>> have many observations of both Y and X1...Xn, for example:
>>
>> 1. Y=1.1, X1=2.1, X2=3.1, ...
>> 2. Y=2.1, X2=2.2, X2=3.2, ...
>> 3 .......
>>
>> (those are not the actual numbers of course)
>>
>> The question is: what is the appropriate method to discover the
>> effect of each of the X variables on Y? (the equivalent of a simple
>> correlation if I had only a single X...)
>>
>
> Are you trying to discover "multiple regression < see
> also, correlation> " ? That would be where there are
> several values of Xs which are used to fit a linear equation
> to 'predict' the corresponding values of Y.
>
> You can probably do better by googling on words from
> your own application area, instead of posting a query this vague.

The most direct equivalent to "a simple correlation" in this context
is a "partial correlation", which might well be googleable. However,
"the appropriate method to discover the effect of each of the X
variables on Y" may well be "multiple regression".

Relevant Pages

  • Re: Partial correlation, confused
    ... "A simple way to compute the partial correlation for some ... data is to solve the two associated linear regression problems, ... get the residuals, and calculate the correlation between the ...
    (sci.stat.math)
  • Re: How to determine if one-way or two-way in pipeline
    ... I don't remember seeing those in the context. ... communication-direction and not the communication-pattern. ... > OneWay scenarios if it is configured to returned the correlation token to ...
    (microsoft.public.biztalk.general)
  • Re: Examining the premises underlying legalization
    ... correlation (constant conjunction & constancy while varying other ... Drug prices, in constant dollars, have been falling steadily, while ... Opiates also decrease ... context like duration & frequency of use, ...
    (talk.politics.drugs)
  • Re: Comparing predictors
    ... > correlation would be meaningful, ... the best I can do still seems to be the predictor correlation ... Or am I misunderstanding partial correlation? ... seems somewhat irrelevant what happens on the extremes, ...
    (sci.stat.math)
  • Re: Correlation ratio = eta-squared, or eta?
    ... one-way ANOVA design. ... <correlation ratio>, I get some pages that tell me correlation ratio = ... This coefficient was developed in the context ...
    (sci.stat.consult)