2001

Anchorage @ the Rabbit Creek Rifle Range, 28 - 29 June

The last week of June saw a convergence of Alaska's statisticians in Anchorage. The first half of the week was spent attending a short course on Bayes/Empirical Bayes/WinBUGS software by Dr. Brad Carlin (sponsored through the courtesy of the Alaska Dept. of Fish and Game Sportfish Division). The course was a great success - many attendees were seen wandering somewhat dazed and lost in thought after each day's presentations. From the questions asked, it was clear that the material struck a chord. Look for more intensive discussions of Bayesian topics in the near future (so how do I get WinBUGS to calculate....). The enthusiastic response was due, in no small part, to the fantastic presentation by Dr. Carlin. Now that he's had a chance to wander a bit in central Alaska perhaps we can expect a return visit in the future?

The last part of the week was spent attending the annual chapter conference. The focus was a great short course on Mixed Effects Models by Dr. Jose Pinheiro. While it made for an intensive week of methodology, it was a nice counterpoise to the Bayes class as attendees to both courses were able to see two alternative methods to multi-level modeling. In addition the short course, three members gave presentations on current work (see below) each leading to a good exchanges of ideas. The highlight, however, had to be the social bbq on Thursday evening - Greg Rosencratz and Brian Tarras did an outstanding job cooking the halibut and salmon (thanks for the donations, Greg), giving the Pinheiro family a taste of Alaska. The were already discussing a return visit in the winter (when the bears are asleep) before Nicky brought out the fresh rhubarb/blueberry crisp!

Member Presentations:

Mark Udevitz (USGS/BRD) ' Preliminary evaluation of aerial survey methods for Dall sheep', co-authored with Brad Shults, Chris Kleckner, and Layne G. Adams

Shareef Siddeek (ADFG, Com. Fish) 'Estimating Natural Mortality of King Crabs from tag recapture data '

Joel Reynolds (ADFG, Gene Conservation Lab) "Reducing Bias in Mixed Stock Analyses"

Original Announcement:

The two-day meeting consists of morning short-course lectures followed, on the first day, by member presentations and, on the second day, by member presentations and the chapter business meeting. There will be a social mixer the first evening. Anyone interested in statistical methods and applications is encouraged to register and attend. It is a great way to overcome the geographic barriers of the state and meet other professionals, discuss your work, exchange ideas, and make friends.

Registration & Course fees:
Sponsored attendees - $125
Nonsponsored attendees - $25
Students - free.

If you would like an invoice before the conference please send Bob Sutherland your name, address and fax number and he will fax you an invoice. Otherwise all attendees will be invoiced after the conference.

Location:

Rabbit Creek Rifle range is on the new Seward Highway south of Anchorge heading towards Girdwood. As you are traveling south, you'll pass the Rabbit Creek Road Exit and about a mile later at the bottom of the hill on the right will be the Range. The Range is clearly marked with a highway sign. Potters Marsh is on your left. If you have passed the Marsh, the Chugach State Park Headquarters or the Weigh Station you have gone too far. Here [link no longer works... removed] is a map to the meeting site. The Red Star laid on top of the 1 at the bottom right hand side of the map is the Rabbit Creek Rifle Range. Call or email Bob Sutherland (267-2524) if you would like more specific directions.

Schedule:

Thursday - 06/28/01

8:15-8:45 - Participant Introduction

8:45-10:00 - short course - Introduction
- motivating examples of linear and non-linear mixed-effects models
- a sample of S-PLUS functions and methods for fitting and analyzing mixed-effects models.

10:00-10:15 - break

10:15-11:15 - Linear Mixed-Effects Model
- assumptions, notation, strengths and weaknesses
- estimation methods: ML and REML
- inference: confidence intervals, hypothesis tests, and predictions
- multilevel LME models

11:15-11:30 - break

11:30-12:30 - Fitting LME models
- model building with the lme function in S-PLUS
- assessing the adequacy of an LME fit
- fitting LME models to multilevel data

12:30-2:00 - Lunch

2:00-5:00 - Member Presentations - to register for a presentation contact Bob Sutherland.

6:00- Salmon Barbecue at Bob Sutherland's. In-town attendees - please bring salads; out of town guests can bring Beer! and chips. We can set up badminton or croquet for those who really want to tie one on. Directions to The Barbecue will be available at the meeting.

Friday - 06/29/01

8:30-8:45 - Coffee

8:45-10:00 - short course cont. - Extending the basic LME model
- modeling the random effects covariance structure
- variance functions to model unequal within-group variances
- serial and spatial within-group correlation structures

10:00-10:15 - break

10:15:-11:15 - Nonlinear Mixed-Effects Models
- brief review of nonlinear regression for independent observations
- obtaining starting estimates: self-starting regression functions
- nonlinear mixed-effects (NLME) models for grouped data

11:15-11:30 - break

11:30-12:30 - Fitting NLME models
- the nlme function in S-PLUS
- assessing the adequacy of an NLME fit
- confidence intervals, hypothesis tests, and predictions
- extending the basic NLME model: variance functions and correlation structures
- incorporating covariates in an NLME model

12:30-2:00 Lunch

2:00-5:00 Business meeting

Saturday
Day hike to Crow Creek Pass (from Girdwood) . It is a 3 1/2 mile hike mainly on an old wagon trail. There is a short steep portion so hiking boots are recommended as well as rain gear, water and a lunch. The scenery in the pass is spectacular.