Please find information regarding a special one day event to be held on
the 19th June in Woking, Surrey, for advanced users of SPSS.
SPSS Power Users Day - June 19th
10 - 10:15 - Registration and Coffee
10.15 - 11.15 Scripting or Macros: which to use and when? Paul Callow,
University of Cambridge.
11:15 - 11:30 - Break
11:30 - 12:30 - Complex Sampling. Jon Petersen, SPSS International.
12:30 - 1:30 - Lunch
1:30 - 2:45 - Design of Experiments. Steve Groves, SPSS (UK) Ltd
3:00 - 4:00 - Amos. Jeremy Miles, Derby University
4:00 - Discussion and hands on examples
As well as these talks, computers will be available all day for
demonstrations and hands on usage. Fees are 45 pounds (plus VAT) per
delegate. To book your places, please call Lisa Brennan, SPSS Training
Coordinator, on 01483 719250. This event is organised by Assess, the
official SPSS user group. A synopsis of the event follows:
10.00 - 10:15 - Registration and Coffee
10.15 - 11.15 - Scripting or Macros: which to use and when?
Paul Callow, Computing Service, University of Cambridge
Repetitive tasks soon become tedious. Most SPSS users soon find
themselves wanting short cuts to perform essentially similar tasks with
as little effort as possible. Two particularly powerful mechanisms are
available to help them. One of these is the macro facility, introduced in
the days when user interfaces were quite primitive, and SPSS tended to be
run on mainframes. The other is scripting, which first appeared with
Release 7.5. Both in effect allow users to design their own SPSS
procedures.
SPSS macros are simply chunks of SPSS command language which can be
referred to by name. They are invoked by macro calls embedded in an SPSS
command program; optionally, extra information provided then can control
the way the SPSS Processor behaves. Despite the facility's age, it is
still useful and requires little knowledge beyond basic SPSS command
syntax.
Scripting exploits the object-oriented structure of modern SPSS user
interfaces and output, with the aid of the Sax BASIC language. As such,
it seems so far to have been used chiefly for tailoring output, though
its ability to incorporate SPSS command language makes it much more
widely applicable.
The talk will consider applications of both methods which exploit their
distinctive features, and discuss the extent to which they can be made to
work together.
11:15 - 11:30 - Break
11:30 - 12:30 - Complex Sampling
Jon Petersen, VP, International Sales & Marketing, SPSS International
For reasons of cost and feasibility it is often impossible to collect
survey data using simple random samples. In such cases complex sampling
techniques maybe used. A complex sample is one in which elements have
unequal probabilities of being selected or in which the chance of an
element's being selected is not independent of other elements' chances of
being selected. When you have survey data collected by a complex samples
scheme, then SPSS and all its conventional competitors do not produce
correct estimates of standard errors and variances of statistics. This
session will introduce you to WesVar Complex Samples 3.0, the only
well-documented, commercial PC-based program that exclusively features
the replication method for variance estimation with complex samples.
12:30 - 1:30 - Lunch
1:30 - 2:45 - Design of Experiments
Steve Groves, Services Manager, SPSS (UK) Ltd
Trial Run is a 'Design of Experiments' program recently released by SPSS.
In essence, the program enables analysts to study complex relationships
between variables with a view to discovering which factors affect a given
outcome and the interactions between them. Unlike traditional programs of
this nature, Trial Run guides the user to an appropriate experimental
design. It contains a range of General Linear Modelling algorithms and
supports 41 experimental designs. This demonstration will show you how to
successfully interrogate a data set with a view to analysing factors and
interactions affecting a particular outcome. Also included will be an
overview of how to produce surface and contour plots.
2:45 - 3:00 - Break
3:00 - 4:00 - Amos
Jeremy Miles, Applied Vision Research Unit, Derby University
Structural equation modelling (SEM) is a very general data analysis
technique, based on regression and factor analysis. In what might be
called "traditional" analysis, data are explored to examine the
relationships that exist between measured variables. Using some
techniques, for example factor analysis, latent variables are
hypothesised to exist, to account for the relationships between
variables. In SEM the process is reversed. A model is proposed and
subsequently tested to see if it is able to account for the data. A
chi-square distributed test statistic is used to test for significance of
difference between the data and the model. Until relatively recently SEM
programs were not especially user friendly, required large amounts of
computer time, knowledge of matrix notation and the Greek alphabet! In
the last few years, AMOS has changed all that and successfully carries
out SEM analysis. It has an intuitive user-interface, which involves
specifying models by drawing path diagrams. This presentation will show
some of the advantages of using a SEM framework to analyse data.
Statistical models and tests that are difficult, or impossible to carry
out using other statistical tests. The types of model that will be
examined include: comparing a one-factor and a two-factor solution in
factor analysis; comparing correlations or covariances both between and
within subjects; investigating the most appropriate model of causation of
a series of variables; comparing regression models across groups.
Thank you.
Chi Tang
Marketing Manager
SPSS (UK) Ltd.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|