Warning: Your browser does not have JavaScript enabled. Please enable JavaScript and reload this page, or click here to view the older version of this page.

1. New: Feedback Form on our Webpage

We are always looking for ways to improve our services. With this in mind we have added a feedback form on our webpage. We would like to have you give us comments and suggestions. The form can be found at: http://www.osc.cornell.edu/feedbackform.htm

2. Careful Design of Data Collection Method
"To call in the statistician after the experiment is done may be
no more than asking him to perform a post-mortem examination:
he may be able to say what the experiment died of."
Sir Ronald Aylmer Fisher

As the summer data collection season approaches, we would like to take the opportunity to stress the importance of careful planning at the design stage. If you need assistance during the design phase, we will be more than happy to help you.

3. Software Availability

SPSS in 318 Philips Hall
There is now one more computer lab on campus open to general users that provides SPSS 13. The newest computer lab operated by Cornell Information Technologies is located on the third floor of Phillips Hall on the Engineering Quad in room 318. The lab is open 24 hours a day.

In order to enter the room, a user must bring their Cornell ID card as the door lock is controlled by a card-swipe system. This lock has been set up to allow entrance to all members of the Cornell community. Up to 19 copies of SPSS can be used at one time.

Minitab in Stimson Hall
The public computing labs in G-25 and G-27 Stimson Hall now provide Minitab in a Windows environment.

4. New Articles/Handouts

We are continually adding to our collection of useful articles. Recent useful additions include monographs on exploratory factor analysis and sample size.

A very useful monograph on how to write up results comes from Cornell's own Daryl Bem's "Writing the Empirical Journal Article" and can be found at: http://www.dbem.ws/WritingArticle.pdf

We also found useful the comparison of the interpretation of the fixed effects parameter between cluster-specific models and marginal models for binary data in the following article:

Pendergast JF, Gange SJ, Newton, MA, Lindstrom MJ, Palta M, Fisher MR, A survey of methods for analyzing clustered binary response data, International Statistical Review, 64(1), 1-30, 1996.

5. Off-campus Statistical Workshops: Summer '06

We have once again compiled a listing of statistical workshops happening this summer across the country. They cover many topics of interest to researchers. They cover topics as diverse as "Intro to Stata" and "Mixed Models for Categorical Data." Most are one day to one week long. Please take a look at: http://www.osc.cornell.edu/workshops/summer2006.htm

6. Noteworthy Upgrades in the New SPSS version 14 Available on Campus

The new SPSS version 14 is now available on campus. The noteworthy upgrades of this new version include the possibility to open several datasets at once and the ability to open and save data files in a Stata format.

7. Updated Complex Survey Capabilities in STATA version 9

Stata version 9 extended its capabilities for the analysis of multistage design survey data, i.e. cluster sampling with two or more levels of clustering. The variance estimators incorporate information about the sampling units at each sampling stage: primary, secondary, and lower stage sampling units.

Previous versions base the variance estimators on the primary sampling units (PSU) only, without requiring information about lower stage sampling units. That produced more conservative estimates, i.e. larger standard errors.

Also new are the addition of the jackknife and balanced repeated replication (BRR) variance estimation methods. The previous versions used "linearization" based methods for computing the standard errors.

8. New Useful Procedures in SAS

Researchers might find the following two new experimental procedures useful:

Proc GLMSELECT: The REG procedure supports a variety of model-selection methods but does not support a CLASS statement. The GLM procedure supports a CLASS statement but does not include model selection methods. Proc GLMSELECT fills this gap by offering a variety of model selection methods along with a class statement.

Proc QUANTREG: This new experimental procedure performs quantile regression. Ordinary least squares regression (OLS) models the relationship between one or more covariates X and the mean of the response variable Y given X=x. Quantile regression extends ordinary least squares regression to model the relationship between one or more covariates X and percentiles of the response Y given X=x. The main advantage of quantile regression over least squares regression is its flexibility for modeling data where these relationships differ.

Both these procedure are experimental and can be downloaded, along with their documentation from: http://www.sas.com/apps/demosdownloads/setupcat.jsp?cat=SAS%2FSTAT+Software

9. Problems using Output Delivery System (ODS Graphics) in SAS

When requesting graphs using ODS graphics commands in SAS (for example to examine residuals in Proc Mixed) do you obtain a little square with a red 'X' instead of a graph, accompanied by an error message along the lines of: "ERROR: Java class generated an exception"? If so, this might be easily corrected using the suggestions available on the following SAS web page: http://support.sas.com/techsup/unotes/SN/015/015237.html

Back to Stat Happenings Table of Contents

(This newsletter was distributed to faculty and graduate students in the Division of Nutritional Sciences and the College of Human Ecology, and faculty in the College of Agriculture and Life Sciences, by the Office of Statistical Consulting. Please forward it to any interested colleagues and research staff. Anyone not receiving this newsletter who would like to be added to the mailing list for future newsletters should contact statcons@cornell.edu.  Information about the Office of Statistical Consulting can be obtained at World Wide Web address http://www.osc.cornell.edu.)