|
|
|
Brief HistoryInterLab2.0 is the evolution of a statistical package originally developed for analytical methods standardisation. It is a computer programme designed to answer a request from ISO/TC 102/SC 2, which is the ISO (International Organisation for Standardisation) Technical Committee devoted to publishing standards for the commercial environment of the worldwide iron ore market. Sub-committee
2 is the division within ISO/TC 102, which is responsible for the development
and publishing of international standards for use as chemical analytical methods
in the iron ore trade. This subcommittee has been publishing methods as
international standards since the late sixties. During
the international meetings held in Ottawa in 1994, the committee decided to
review its statistical procedures, in order to contemplate three basic issues: a)
Incorporate some advancements included in the revision of ISO 5725, at
that time already available; b)
Include procedures for assessing trueness of a measurement method, in
accordance with modern statistical methodologies based in regression techniques; c)
Include procedures for evaluating if the precision attained by a working
group, developing an international method, is acceptable, by means of comparing
precision estimates with past results and general references. The
efforts for pursuing this objective, assigned to a Brazilian expert, produced
InterLab, the first version of the software, which was developed under MS DOS.
Later, in the year 2000, another version, InterLab 2000 for Windows was
prepared, as an upgrade to the Windows platform. These first developments were
made in-house, and they were mainly used by the statistical expert in ISO/TC
102/SC 2. During
the coming years, up to 2005, many revisions were made in the working document
used by that committee, document ISO/TC 102/TCR 5, which is a Technical
Committee Report, issued within TC 102 for its own use during the development
stages of its methods. This document is inspired in, and follows closely what is
found in ISO 5725-2. Due to the request of incorporating tests for Trueness,
also the methodology in ISO 5725-4 was incorporated. The idea now in the
committee, is that project leaders and working group conveners do their own
processing of laboratory or interlaboratory experiments, with the aid of the
statistical package, which now implements the procedures set out in TCR 5. Research
in interlaboratory experiments and statistical tools for them were undertaken on
request, and produced what is now the new revision of TCR 5, issued in the end
of 2005. This includes methods for comparing precision curves, assessment of
Trueness by regression analysis (using maximum likelihood in a functional
errors-in-variables model), and a new outlier test for using as an alternative
to Grubbs’ test. The implementation of all those new procedures, harmonised
with the ones already in use, is achieved with the new version of the software,
InterLab2.0 for Windows. As some changes in the document were
produced in the last quarter of 2005, they are scheduled to be incorporated in
the following months in the package. Some
of the new advantages of InterLab2.0 are: ü
As TC 102 started a proficiency test for iron ore laboratories,
calculation of Z scores, as recommended in ISO Guide 43-1, was implemented; ü
As an alternative to the Z scores based on the data from each round, a Z
score was introduced that considers the estimate of the target standard
deviation as the reproducibility standard deviation calculated with Horwitz’s
Function (see Horwitz, 1982, and Thompson & Lowthian, 1997); ü
The new environment is available through the Internet. Therefore, users
do not need to install any programme in their computers. Everything can be done
using an Internet connection, and Internet Explorer 6.0 or a newer version. With
this, communication between coordination and participants in an international
trial (for collaborative purposes, or for proficiency tests, for instance) is
greatly improved. Participants can enter their own data in the same database,
and after all have provided their inputs, the coordinator may grant them access
to the system, in order to see the results in real time (as a read-only access,
if necessary); ü
Graphics facilitate interpretation in every step of the evaluation, and
they can be exported in GIF format. These may be later incorporated in reports
as images, or in presentations; ü
Data can be easily copied/pasted to and from Excel spreadsheets; ü
Navigation is extremely easy between screens, and tools provided are
intuitive and commonplace in the Windows environment. |
Questions or comments about this website, please contact us at:
webmaster@kymi.com.br.
|