CETIS News and Recent Updates

Listed below are some of the significant changes and enhancements incorporated into the most recent versions of CETIS. It is highly recommended that you obtain and use the current version. Check back regularly for further details. The current changelog document can be viewed and downloaded for your convenience.

Version 2.1.0

  • CETIS is now offered as a 32-bit or 64-bit version. When ordering, specify the version you need.
  • The Graphics Server Active X component has been supplanted by the TeeChart AX component.
  • The TeeChart AX component offers improved and more capable graph exporting options.
  • Certain graphs now show 95% confidence regions displayed as shaded areas instead of dashed lines, yielding a more modern look.
  • The QC plot has been given more capability and a more modern look.

Version 2.0.1

  • Added new formatting of data analysis result forms, including tiling, cascading, shrinking or fit all.
  • The pseudo R2 value is now based on McFadden's R2.
  • Added more nonlinear models a PMSD can be calculated for.
  • A heterogeneity factor is now applied to beta-binomial nonlinear regressions as it is for GLM probit analysis.
  • Added hard restrictions on which nonlinear weighting can be applied to certain types of data. Poisson and Negative Binomial must be positive integers and the Gamma and Inverse Gaussian must be non-zero positive values.

Version 2.0.0

  • The file format has been changed to the accdb/accde format to improve password protection and encryption.
  • The interface has a new modern look with redesigned icons and a flat appearance. Changed the default screen font to Franklin Gothic Medium in an attempt to increase readability on screen.
  • A radically redesigned main CETIS form allows all information to be viewed simultaneously, improves record filtering and makes test/sample batching much easier.
  • While CETIS still makes use of the workgroup information file, it no longer needs to be referenced within the startup shortcut. A new Workgroup Administrator app has been designed to help with managing accounts.
  • Improved data import support.
  • Improved intensive database repair utility that fully recreates primary keys, indexes and relationships when found to be corrupted or missing.
  • Removed the historical restriction in how the back-end file is named.
  • The CUSUM plot is now a two-sided CUSUM plot as is widely described in the literature.
  • Made improvements to the graph axis configuration algorithms.
  • Added a new search tool on the main CETIS form to support the new embedded global test list.

Version 1.9.7

  • A new field named "MD5" was added to store the MD5 digest of the data set associated with each analysis. The MD5 hash will be printed on each analytical report. The last date of editing and the User ID associated with the edit is displayed on the reports. Any change to the data between analyses will be apparent from the change in the MD5 hash. The MD5 hash digest can be considered as a 128 bit digital fingerprint.
  • A MD5 digest of the data at the time a Summary report is printed is now shown on the reports. Any change to the data between printing of the Summary reports will be apparent from the change in the MD5 hash.
  • Added a new utility determining statistical differences between two point estimates (i.e., LCx, ECx). Both the Zajdlik two-sample Z test and the Robertson-Preisler ratio tests are implemented for both native and external data.
  • Major rewrite of queries to reduce the number of database connections during report generation. This was done to reduce or eliminate the "Out of database connections" error message.
  • Data on analytical reports now maintain their proper replicate number if other replicates are eliminated during analysis.
  • Multiple forms are now vertically expandable.
  • The user manual has been revised to reflect changes to the software.
  • All analysis result forms are optimized at run-time so no vertical scrolling is needed to view all the results.
  • A new CETIS to CEDEN (California Environmental Data Exchange Network) translator has been developed to work with v1.9.7.xx. For those that have to submitt these EDDs, the translator is a virtual lifeline.
  • Additional bug fixes, please download and view the changelog.

Version 1.9.6

  • Due to changes in Office Update algorithms and unauthorized registry key deletions, CETIS now stores the license leys in the file system.
  • CETIS now forces UNC addressing for linked tables due to Microsoft Windows bugs.
  • A control standard deviation plot was added to the QC Plotting utility.
  • Additional bug fixes, please download and view the changelog.

Version 1.9.5

  • Added Tarone Test and Rao-Scott adjusted Cochran-Armitage Test.
  • Switched to .Net 4.0 version of QLM support files.
  • New PMSD calculation for linear and nonlinear regression.
  • Automated saving of stand alone concentration-response plots.

Version 1.9.1

  • CETIS now implements the ribbon UI for Access 2010 and higher.
  • Improved drop-down filter implementation.
  • Added a new special version of the CETIS Main form intended just for adding new tests.
  • Improved export to Excel.

Version 1.9.0

  • CETIS now is installed with a downloadable installer.
  • The license is now registered online.
  • Just about every aspect of the program has been improved including the user interface, queries and reports.
  • The negative binomial and beta binomial distributions have been added to the nonlinear regression module.
  • An improved test cloning utility that significantly decreases the time to setup a test.

Version 1.8.7

  • A test type can now have multiple default endpoints. Previous versions allowed only a single default endpoint. If more than one endpoint is designated as a default endpoint, they will all be initially selected in the Available Endpoints listbox. This new feature can be used effectively with the Calculate All button.
  • Various non-critical software bugs have been addressed.

Version 1.8.6

  • The Analysis Options form has been revised and operationally streamlined.
  • Various non-critical software bugs have been addressed.
  • CETIS now stores licenses in new locations that can be more easily managed by admin.

Version 1.8.5

  • Test Types and selected endpoints can now be cloned and then revised.
  • Improved editing of test type fields and field properties.
  • Revised and reorganized menu system.

Version 1.8.4

  • CETIS now has a database diagnostic tool and repair utility.
  • Improved Access 2010 and Win7 compatibility.
  • Improved printing resources and reduced the chance of partial report printouts.
  • Exact permutation method for the Mann-Whitney, Wilcoxon and Fligner-Wolfe Tests.
  • Added the Cochran-Armitage Linear Trend Test including Monte-Carlo methods.
  • Improved nonlinear regression starting parameter estimator.

Version 1.8.1

  • Added the Hochberg-GT2, Dunnett T3, Gabriel, Tamhane-Dunnett Step Down and the Games & Howell multiple comparison methods.
  • Increased number of data entry fields to 50.
  • CETIS is now compatible with 64 bit Windows® 7.
  • Changed Shirley's test to the Williams modified Shirley test. This improves statistical power by re-ranking at each step level.

Version 1.8.0

  • CETIS now has full batch processing, meaning all endpoints for a given test can be analyzed with a single click.
  • CETIS is now Section 508 compliant.
  • Added the new U.S. EPA Test for Significant Toxicity (TST) Welch's t Test described in document EPA 833-R-1-003. The TST method is a bioequivalence approach that controls the false negative rate.
  • CETIS now provides a dynamic printer device selector.
  • Added the capability of creating and saving QC plot templates.
  • Modified the reports for multiple sample designs to present the data more effectively.
  • Added the Anderson-Darling A2 test to the comparisons, linear regression and nonlinear regression modules.
  • Added the Likelihood Ratio test to the linear and nonlinear regression modules and added the Pearson Chi-Sq test to the nonlinear regression module.
  • Added the ANOVA based Lack of Fit test to the linear regression module.
  • Fine-tuned the algorithm used to automatically set the nonlinear regression starting parameters.
  • Added the Bayesian Information Criteria (BIC) to those parameters calculated with the linear and nonlinear regression modules.
  • Added the nonlinear Gustafson and Holden model, also known as the First Order Multi-Compartmental model (FOMC), the Double First Order in Parallel (DFOP) model, Simple First Order (SFO) model, and the Hockey Stick model as described in the FOCUS (2006) document. These models are standard models for chemical degradation and kinetics studies.
  • Improved queries used to generate reports, resulting in much faster report generation for large databases.

Version 1.7.0

  • Certain terms used in previous versions of CETIS have been substituted for more meaningful terms to help the new user grasp the basics of using CETIS. The terms 'link' or 'test link' are no longer part of the CETIS lexicon except where it applies to linked tables.
  • The task of linking sample with a test run has been altered. Instead of linking, the term now is referred to as sample batching.
  • A single sample can be added to multiple test batches from the Samples page of the main CETIS form. Also multiple samples can be added to a single test batch from the Test Batch page of the main CETIS form. Either method accomplishes the same goal, that is, adding a sample to a test batch. A user-selectable 'Span' variable allows you to specify the maximum time period in days between the sampling event and the testing event, thus keeping the number of records displayed in the form or listbox to a minimum.
  • The Mann-Kendall Trend Test and Grubbs' Outlier Test can now be run with any module. In the case of the Linear Regression and the Nonlinear Regression modules, the Mann-Kendall and Grubbs' tests are directed against the residuals. In the case of the Comparisons, Linear Interpolation and Spearman-Karber modules, the tests are directed against the empirical dataset. For the Linear Interpolation and Spearman-Karber modules, the Angular (Corrected) transformation is used to transform the proportions derived from quantal tests prior to running Grubbs' test. Non-quantal data is left untransformed. Grubbs' test can only be run when the data are shown to be normally distributed.
  • The record locking status is now memorized when closing either the main CETIS form or the Test Data form.
  • The value in a date/time field can now be easily changed by using the arrow keys in combination with the SHIFT key. The SHIFT-UP arrow and SHIFT-DOWN arrow will increment and decrement, respectively, one hour to the time. The SHIFT-RIGHT arrow and SHIFT-LEFT arrow will increment and decrement, respectively, one minute to the time.
  • Added a new command, 'Paste From Excel,' accessible by right-clicking in a cell on the Test Data form. This new command will automatically determine the dimensions of the selected array held in the clipboard and will select the same sized array on the CETIS Test Data form. Due to an Access imposed limitation, previous versions required the user to manually select the correct block of cells, otherwise the paste command would fail.
  • Revised the 'Stored Comments' form that is accessible from each of the comment fields on the main CETIS form. The revised form allows the editing of long comment text and allows users to save long text for reuse as a template. To save text, enter a new 'Code' up to 10 characters long and click on the 'Save Template' button. The new entry will be seen in the drop-down list after saving the record.
  • When revising a multiple sample design such as a sediment test, you can reorder the ordering of the samples. This may be useful if after you have set up a test and entered data you decide that the ordering was incorrect. You can open the "Test Design" form, change the order in the 'Test Group' field and click on the 'Update' button.
  • Changed the behavior of the nonlinear regression 'Test Fit' graphing form. Instead of being a modal dialogbox, it is now a popup so it will remain on the screen. The user can de-select the Autofit option after getting an initial fit and then adjust the parameters accordingly and refresh the graph by clicking on the 'Test Fit...' button again. This should make it easier to analyze problematic data sets where the intitial parameters are not close enough to achieve convergence.
  • Added the capability of changing nominal concentrations to verified concentrations from the 'Test Data' form. This feature is available by clicking on the third command button on the bottom left of the form.

Version 1.6.6

  • The decision trees have been expanded to include logic for point estimates as well as comparisons. Different analysis templates can be assigned to both quantal and quantitative data. You can select up to four different analyses to run in series. The primary method will run first, followed by the secondary, tertiary and finally quaternary method. You can specify to stop after the first acceptable fit or run and display all analyses. By selecting different decision trees, you can affect how the data will be calculated and which templates will be loaded prior to the calculation.
  • Decision tree templates and analysis option templates can now be imported from special 'template' files. Tidepool Scientific may also provide you, on occasion, with a file to load into your existing database.
  • Added the Mann-Kendall trend test for control data. When there are less than 10 replicates, CETIS performs an iterative permutation calculation to provide an exact probability, even when ties exist. When there are 10 or more replicates, the user can select either a Monte-Carlo based probability or a normal approximation using Z-scores. The normal approximation method has a correction for ties and Monte Carlo inherently addresses ties.

"Decision flowcharts applied to my toxicity data can be complicated and overwhelming. How will CETIS help insure I am analyzing my data correctly?"

CETIS implements decision-making logic that make most, if not all, decisions for you such as selecting the correct method and verifying that all assumptions are met. You can override any decision and manually select a method of your choice.

Read More