Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: interaction graph
From
Nick Cox <[email protected]>
To
[email protected]
Subject
Re: st: interaction graph
Date
Wed, 23 Feb 2011 10:16:11 +0000
The procedure in your first paragraph is not what I had in mind, and
will at best produce averages across your prediction surface, which
given its curvature will not correspond to points on your surface. But
your question needs to be to yourself: whether a graph suits your
purposes depends on your purposes.
Official commands that are downloadable occur _only_ as part of an
-update- from StataCorp. Either way, I cannot comment on a command I
cannot see or find.
Nick
On Wed, Feb 23, 2011 at 10:07 AM, <[email protected]> wrote:
> In a 2-D graph, isn't it possible to "fix" the value of the continuos moderator at some values (i.e. creating a dummy variable, for instance for values below the first quatile and above the third quartile) and use regplot?
>
> I think I found interact long time ago on the internet, saying it was downlodable. I did not know it was not an official command.
>
>
>
>
>
> ----- Messaggio originale -----
> Da: "Nick Cox" <[email protected]>
> A: [email protected]
> Inviato: Mercoledì, 23 febbraio 2011 10:47:39 GMT +01:00 Amsterdam/Berlino/Berna/Roma/Stoccolma/Vienna
> Oggetto: Re: st: interaction graph
>
> I don't have a specific suggestion.
>
> Your model produces a prediction surface and although there are some
> user-written commands for 3-D graphics plotting the entire surface is
> a problem for which you are better off outside Stata.
>
> On the other hand, generating predictions for a few fixed values of
> your moderator might produce an adequate graph.
>
> -interact- is not an official command. It follows that you should
> explain where it comes from, as long and often requested on this list.
>
> Nick
>
> On Wed, Feb 23, 2011 at 9:34 AM, Fabio Zona <[email protected]> wrote:
>
>> time ago you suggested to use
>>
>> regplot x, separate(dummy)
>>
>> to graph interaction effect between a continuous variable and a dummy variable.
>>
>>
>> However, what command should I use if I have
>> a) a quadratic form (i.e., x + x_square)
>> b) the moderator is a continous variable (not a dummy), and it is multiplied by both x and x-square ??
>>
>> The stata command "interact" works only with linear (i.e. non quadratic) models
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/