[R] Effects - plot the marginal effect

[David Winsemius]

On Apr 8, 2011, at 4:24 AM, Tomii wrote:

> Thank you for your response, but these changes doesn't seem to change
> anything, outcomes of effect command is still the same - error.
>
> Tomas
>
> On Fri, Apr 1, 2011 at 5:03 AM, John Fox <jfox at mcmaster.ca> wrote:
>
>> Dear Tomas,
>>
>> Write the model as
>>
>> mreg01 = lm(enep1 ~ enpres * proximity1), data=a90)

A syntactic comment only. Perhaps:

  mreg01 = lm(enep1 ~ enpres * proximity1, data=a90)

--  
David

>>
>> That is, it's not necessary to index a90 as a list since it's given  
>> as the
>> data argument to lm, and doing so confuses the effect() function.  
>> Also,
>> enpres*proximity1 will include both the enpres:proximity1  
>> interaction and
>> enpres + proximity1, which are marginal to the interaction.
>>
>> Next, you must quote the name of the term for which you want to  
>> compute
>> effects, thus "enpres:proximity1" in the call to effect().
>>
>> Finally, effect() doesn't compute what are usually termed marginal  
>> effects.
>> If you want more information about what it does, see the references  
>> given in
>> ?effect.
>>
>> I hope this helps,
>> John
>>
>> ------------------------------------------------
>> John Fox
>> Sen. William McMaster Prof. of Social Statistics
>> Department of Sociology
>> McMaster University
>> Hamilton, Ontario, Canada
>> http://socserv.mcmaster.ca/jfox/
>>
>> On Thu, 31 Mar 2011 22:09:32 +0200
>> Tomii <diogenas at gmail.com> wrote:
>>> Hello,
>>>
>>> I try to plot the marginal effect by using package  
>>> "effects" (example of
>> the
>>> graph i want to get is in the attached picture).
>>> All variables are continuous.
>>>
>>> Here is regression function, results and error effect function  
>>> gives:
>>>
>>>> mreg01 = lm(a90$enep1 ~ a90$enpres + a90$proximity1 + (a90$enpres *
>> a90$proximity1), data=a90)> summary(mreg01)
>>> Call:
>>> lm(formula = a90$enep1 ~ a90$enpres + a90$proximity1 + (a90$enpres *
>>>    a90$proximity1), data = a90)
>>>
>>> Residuals:
>>>    Min      1Q  Median      3Q     Max
>>> -2.3173 -1.3349 -0.5713  0.8938  8.1084
>>>
>>> Coefficients:
>>>                          Estimate Std. Error t value Pr(>|t|)
>>> (Intercept)                 4.2273     0.3090  13.683  < 2e-16 ***
>>> a90$enpres                  0.4225     0.2319   1.822 0.072250 .
>>> a90$proximity1             -3.8797     1.0984  -3.532 0.000696 ***
>>> a90$enpres:a90$proximity1   0.8953     0.4101   2.183 0.032025 *
>>> ---
>>> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>>>
>>> Residual standard error: 2.029 on 78 degrees of freedom
>>> Multiple R-squared: 0.2128,   Adjusted R-squared: 0.1826
>>> F-statistic: 7.031 on 3 and 78 DF,  p-value: 0.0003029
>>>> plot(effect(a90$enpres:a90$proximity1, mreg01))Warning messages: 
>>>> 1: In
>> a90$enpres:a90$proximity1 :
>>>  numerical expression has 82 elements: only the first used2: In
>>> a90$enpres:a90$proximity1 :
>>>  numerical expression has 82 elements: only the first used3: In
>>> analyze.model(term, mod, xlevels, default.levels) :
>>>  0 does not appear in the modelError in
>>> plot(effect(a90$enpres:a90$proximity1, mreg01)) :
>>>  error in evaluating the argument 'x' in selecting a method for  
>>> function
>> 'plot'
>>>
>>>>
>>>
>>> Thanks in advance.
>>> Tomas
>>
>>
>>
>>
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

David Winsemius, MD
West Hartford, CT