Hi Robert
Have you tried emailing Dan Johnson directly on the Rbrul list?
[log in to unmask]
Cheers,
Adam
--
Assoc. Prof. Adam Schembri, PhD
Director | National Institute for Deaf Studies and Sign Language
(www.latrobe.edu.au/nids)
Interim director | Centre for Research on Language Diversity
(www.latrobe.edu.au/rclt)
La Trobe University | Melbourne (Bundoora) | Victoria | 3086 | Australia
Tel : +61 3 9479 2887/6401 | Mob: +61 432 840 744
|http://www.adamschembri.net/webpage/Welcome.html
On 5/09/12 18:44 , "STRAW, Michelle (Dr)" <[log in to unmask]> wrote:
>Hi Rob
>
>I haven't used Rbrul, but it may be that you have to go back a step to the
>descriptive stats, means etc. and do some cross tabulations to draw out
>patterns between factors. Sometimes it is best to start looking at
>internal
>factors only to start with, then social to determine what factors should
>actually go into the model. It sounds like playing around with the data
>and I
>am sure someone else can direct you more formally on what might help.
>
>Regards
>
>Michelle Straw
>
>-----Original Message-----
>From: Variationist List [mailto:[log in to unmask]] On Behalf Of VAR-L
>automatic digest system
>Sent: Wednesday, September 05, 2012 12:03 AM
>To: [log in to unmask]
>Subject: VAR-L Digest - 1 Sep 2012 to 4 Sep 2012 (#2012-95)
>
>There is 1 message totaling 88 lines in this issue.
>
>Topics of the day:
>
> 1. Statistical analysis of TH-fronting in Rbrul?
>
>----------------------------------------------------------------------
>
>Date: Tue, 4 Sep 2012 16:40:12 +0100
>From: Robert Lawson <[log in to unmask]>
>Subject: Statistical analysis of TH-fronting in Rbrul?
>
>I'm currently writing a paper about TH-fronting in Glasgow, Scotland,
>focusing specifically on word initial position, and I'm having some
>difficulties with running the statistical side of my analysis which I'm
>hoping some knowledgeable souls might be able to help with.
>
>I'm using the R plug-in Rbrul (Johnson 2009) to find out which independent
>variables are most influential in determining rates of TH-fronting.
>
>Dependent variable - (TH)
>
>For those unfamiliar with Glaswegian English, I've split this variable up
>into two patterns.
>
>Pattern 1 words can take [th], [f] or [h]
>Pattern 2 words can take [th] or [h]
>
>I know that Rbrul can only deal with binomial variables. For pattern 1,
>this
>is fine, but for pattern 2, I decided to split it up into:
>
>Standard [th] vs. non-standard [f], [h]
>Innovative [f] vs. local [h]
>
>Independent variables -
>
>Preceding phonological class
>Following phonological class
>Number of syllables
>Word class
>Community of Practice membership
>Topic
>Speaker
>
>Taking 'speaker' and 'word' as random factors in Rbrul, almost every
>model I
>run returns the independent variables as non-significant.
>
>For example, here's the output for one of the runs on my Year 1 data:
>
>ONE-LEVEL ANALYSIS WITH word [random] and speaker [random] and cofp
>(0.0667)
>+ preceding_class (1) + following_class (1)
>
>My own reading of it here is that CofP is near significance (0.0667), but
>preceding and following class aren't at all (1.00)
>
>Another one from my Year 2 data:
>
>ONE-LEVEL ANALYSIS WITH speaker [random] and word [random] and cofp
>(0.231) +
>following_class (0.252) + preceding_class (0.885)
>
>Again, nothing is significant here or even approaching significance...
>
>One of the issues I've been having is that when I try to add more
>predictors
>to the model, I keep on getting an error message from Rbrul. I'm not sure
>if
>this is to do with the order I'm adding the predictors, or if there's
>another
>underlying issue I don't know about.
>
>In any event, if none of the predictors in the model are significant, does
>this mean that I've likely missed something in my coding to explain the
>analysis? Or is there something else going on here that I'm missing? Even
>if
>the model is non-significant, can I still report factor weights/log odds?
>
>Hopefully someone can shed some light on this! Happy to provide further
>information as required.
>
>Thanks,
>
>Rob Lawson
>
>########################################################################
>
>The Variationist List - discussion of everything related to variationist
>sociolinguistics.
>
>To send messages to the VAR-L list (subscribers only), write to:
>[log in to unmask]
>
>To unsubscribe from the VAR-L list, click the following link:
>https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=VAR-L&A=1
>
>------------------------------
>
>End of VAR-L Digest - 1 Sep 2012 to 4 Sep 2012 (#2012-95)
>*********************************************************
>
>########################################################################
>
>The Variationist List - discussion of everything related to variationist
>sociolinguistics.
>
>To send messages to the VAR-L list (subscribers only), write to:
>[log in to unmask]
>
>To unsubscribe from the VAR-L list, click the following link:
>https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=VAR-L&A=1
########################################################################
The Variationist List - discussion of everything related to variationist sociolinguistics.
To send messages to the VAR-L list (subscribers only), write to:
[log in to unmask]
To unsubscribe from the VAR-L list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=VAR-L&A=1
|