Notice: On March 31, it was **announced** that Statalist is moving from an email list to a **forum**. The old list will shut down on April 23, and its replacement, **statalist.org** is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
David Hoaglin <dchoaglin@gmail.com> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
Re: st: Question about interactions |

Date |
Mon, 4 Mar 2013 11:55:07 -0500 |

Karin, Together, the two 4-point scales would yield a total of 16 predictors (each of which is an indicator variable for a combination of a category of variable1 and a category of variable2). In your model the combination of variable1=1 and variable2=1 corresponds to the constant term. The combination of variable1=1 and variable2=2 corresponds to the "main effect" for variable2=2 (and similarly for variable2=3 and variable2=4). The combination of variable1=2 and variable2=1 corresponds to the "main effect" for variable1=2 (and similarly for variable1=3 and variable1=4). The coefficients for the other 9 combinations of a category of variable1 and a category of variable2 appear under variable1#variable2 in the sketch of output that you included. I hope this is helpful. David Hoaglin On Mon, Mar 4, 2013 at 11:30 AM, K Jensen <k.x.jensen@gmail.com> wrote: > Sorry to try your patience but can I test my understanding here? > > The results for 1-2 (say) are calculated from the main effect betas > for variable1=1 and variable2=2? > > Thankyou > Karin > > On 4 March 2013 15:37, Maarten Buis <maartenlbuis@gmail.com> wrote: >> On Mon, Mar 4, 2013 at 3:48 PM, K Jensen wrote: >>> I am fitting two four-point scales as predictors in a logistic >>> regression and am interested in looking at the interaction between the >>> two. >>> >>> If I fit variable1##variable2 I get odds ratios for the following: >>> variable1 >>> 2 >>> 3 >>> 4 >>> variable2 >>> 2 >>> 3 >>> 4 >>> variable1# >>> variable2 >>> 2 2 >>> 2 3 >>> 2 4 >>> 3 2 >>> 3 3 >>> 3 4 >>> 4 2 >>> 4 3 >>> 4 4 >>> >>> I obviously don't understand this because it doesn't include cells >>> 1-2, 1-3, 1-4, 2-1, 3-1 or 4-1. Can someone explain this? >> >> For both variables the category 1 is the reference category and is >> thus, together with its interaction terms, excluded. More on this, and >> alternatives, can be seen here: M.L. Buis (2012) "Stata tip 106: With >> or without reference", The Stata Journal, 12(1), pp. 162-164. >> <http://www.maartenbuis.nl/publications/ref_cat.html>. >> >> -- Maarten * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/

**Follow-Ups**:**Re: st: Question about interactions***From:*K Jensen <k.x.jensen@gmail.com>

**References**:**st: Question about interactions***From:*K Jensen <k.x.jensen@gmail.com>

**Re: st: Question about interactions***From:*Maarten Buis <maartenlbuis@gmail.com>

**Re: st: Question about interactions***From:*K Jensen <k.x.jensen@gmail.com>

- Prev by Date:
**Re: st: Question about interactions** - Next by Date:
**Re: st: unconventional lag length in VAR model?** - Previous by thread:
**Re: st: Question about interactions** - Next by thread:
**Re: st: Question about interactions** - Index(es):