Design-Expert is a must for anyone wanting to improve a process or a product. With Design-Expert you can screen for vital factors, locate ideal process settings to achieve peak performance and discover your optimal product formulations.

Design-Expert Tutorial

Build and analyze a one-factor design factorial, "Bowling.

Design-Expert 9

Design-Expert offers an impressive array of design options. Version 8 includes dozens of new
features that increase the ease-of-use, functionality, power and appeal of an already great product.

Those of you who've used previous versions of Design-Expert software will be impressed with the many improvements in version 9. Here are the highlights:

Hard-to-change factors handled via split plots

  • Two-level, general and optimal factorial split-plot designs: Make it far easier as a practical matter to experiment when some factors cannot be easily randomized.
  • Half-normal selection of effects from split-plot experiments with test matrices that are balanced and orthogonal: The vital effects, both whole-plot (created for the hard-to-change factors) and sub-plot (factors that can be run in random order), become apparent at a glance!
  • Power calculated for split plots versus the alternative of complete randomization: See how accommodation of hard-to-change factors degrades the ability to detect certain effects.

Other new design capabilities

  • Definitive screening designs: If you want to cull out the vital few from many numeric process factors, this fractional three-level DOE choice resolves main effects clear of any two-factor interactions and squared terms (see screen shot of correlation matrix - more on that later).
  • On the Factorial tab select a simple-sample design for mean-model only: Take advantage of powerful features in Design-Expert software for data characterization, diagnostics and graphics - for example with raw outputs from a process being run at steady-state.

Much-improved capabilities to confirm or verify model predictions

  • Entry fields for confirmation data and calculation of mean results: Makes it really easy to see if follow-up runs fall within the sample-size-adjusted prediction intervals.
  • Enter verification runs embedded within blocks as controls or appended to your completed design: Lend veracity to your ultimate model by these internal checks.
  • Verification points displayed on model graphs and raw residual diagnostics: See how closely these agree to what's predicted by your model.

New and more-informative graphics

  • Adjustably-tuned LOESS fit line for Graph Columns: Draw a curve through a non-linear set of points as you see fit. *(Locally weighted scatterplot smoothing.)
  • Color-coded correlation grid for graph columns: Identify at a glance any factors that are not controlled independently of each other, that is, orthogonally; also useful for seeing how one response correlates to another.

Greater flexibility in data display and export

  • Journal feature to export data directly to Microsoft Word or Powerpoint: Fast and formatted for you to quickly generate a presentable report on your experimental results.
  • Improved copy/paste of Final Equation from the analysis of variance (ANOVA) report to Microsoft Excel: This not only saves tedious transcription of coefficients but it also sets up a calculator for you to 'plug and chug', that is, enter into the spreadsheet cells what values for the inputs you'd like to evaluate and see what the model predicts for your response.
  • New XML* script commands for exporting point predictions: Helpful for situations where one wants to automate the transfer of vital outputs from Design-Expert to other programs. *(Extensible Markup Language)

More powerful tools for modeling

  • All-hierarchical model (AHM) selection: Sort through all possible models up to the one you designed the experiment for, but all the while maintain hierarchy of terms so you do not end up with something ill-formulated.
  • Special quartic Scheffé polynomial included in automatic selection for mixture modelling: Sometimes this added degree (4th!) of non-linear blending helps to better shape the response surface - making it better for predictive purposes.

More choices when custom-designing your experiment

  • Enter a single factor constraint for response surface designs: Creates a 'hard' limit on inputs that cannot go beyond a certain point (such as zero time) physically or operationally.

More capability for numerical optimization

  • Include Cpk* as a goal: Meet quality goals explicitly. *(A process capability index widely used for Six Sigma and Design for Six Sigma programs.)

Enhanced design evaluation

  • One-sided option added to FDS* graph: Size your design properly for a verification experiment done to create a QBD** design space.
    *(Fraction of design space) **(Quality by Design - a protocol promoted by the US Food & Drug Administration (FDA).)

Many things made nicer, easier, more configurable and faster

  • Diagnostics report now can be sorted by any of the statistics listed: This enables a more informative ordering than by run number (the default).

Niceties that only statisticians might truly appreciate

  • Mean correction for transformation bias when responses displayed in original scale: All you need to know is that our statisticians figured out how to eliminate a tricky, little-known bias!
  • Propagation of error (POE) carried out to the second derivative: Makes POE more accurate - that's a good thing!
  • Allow averaging of categoric factors when viewing a graph: Convenient for getting the big picture of where to find robust operating settings.
  • Display confidence bands with or without POE added: Easier to match output with other programs that do not offer POE features like this.
  • Add unblocked results to evaluation of blocked experiments: Aids in comparing designs on the basis of matrix measures.

Good news for network administrators

New more flexible and easier-to-use license manager with greater power toserve enterprise users: For example, network 'seats' can be checked out to individual laptops and multiple opening of the program on a specific computer will only use one seat.


New graphics and improved interface

  • Half-normal selection of important effects on all factorial designs*: Simple and robust method for selecting important effectsformerly available only for two-level designs. For example, the screen shot to the right is from an experiment on 5 woods glued with 5 adhesives, using 2 applicators with 4 clamps at 2 pressures. The vital effects become apparent at a glance!
    *(Detailed in “Graphical Select-ion of Effects in General Factorials”—winner of the Shewell Award for best presentation at the 2007 Fall Technical Conference, co-sponsored by the American Society for Quality and the American Statistical Association.)

  • Smoother color gradations on 2D contours: More impressive for presentations to management, clients, or colleagues.

  • Rounded contour values: More presentable defaults requiring less ‘fiddling’ for reporting purposes.

  • Plant flags on 3D surfaces: Previously, you could only put flags on 2D contour plots. To the right we see a flag planted by numerical optimization on turbidity of a detergent formulation via mixture designa specialized application of response surface methods (RSM).

  • New and fully configurable mesh option that reflects smooth, lighted colors off your 3D surface: Dazzle your customers and colleagues while providing highly-informative graphics showing how responses will react to process changes. (Mesh can be turned off if you like.)

  • 3D graphs that you can spin with your mouse: When you see your cursor turn into a hand (I), simply grab and rotate! Double-click the graph to go back to the starting angle.

  • Push-button averaging on the factors tool: Provides far easier main effects plotting and makes interactions more meaningful. Previously, the only option to average factors came via a hidden drop-list.

  • More-interactive cube plots: Click on design points to see factor levels and response predictions on graph legends, as below.

  • Direct setting of discrete (fixed) numeric levels in response surface designs: Limit factor settings to reasonable levels but still produce continuous models.

  • Discrete factor levels adhered to in numeric optimization: Find the most desirable setting for factors that are not continuous, such as the number of passes through a spray coater.

  • Enter input variables vertically: When entering many levels, this may be more convenient than the horizontal layout.

  • Reference lines on plots: Horizontal, vertical, and free style-lines enhance plots.

  • Predicted vs. Actual graph availability in Model Graphs, not just in Diagnostics: This is useful when a response has been transformed because in Model Graphs mode, you can change the view back to the more relevant original scale.

  • Confidence, prediction, and tolerance intervals (CI, PI & TI) plotted with configurable colors in one-factor response plots: Convey prediction uncertainties via bands around the best fit. The screen shot at right shows actual run results represented as red circles. The solid line is the predicted value based on the polynomial model. The bands are the CI (narrowest), PI, and TI (widest).

  • Color-coded response surface graphs show where standard error increases: This makes it easier to understand why a predicted response will get you in trouble by extrapolating beyond actual experimentation regions. The example at right shows a flag set beyond the axial points of a central composite designmaking the prediction meaningless.

Better mixtures design and modeling tools

  • Partial quadratic mixture (PQM) analysis: Model non-linear blending behavior most effectively.

  • Design for linear plus squared terms in mixture models: Reduce the number of blends required for optimally-designed experiments that reveal non-linear blending.

  • Design for special and full quartic mixture models: Capture extremely non-linear relationships among all components.

  • Blocking expanded to simplex mixture designs: For example, blend your cakes and bake them in two oven batches.

  • Trace plot options show end points as actual values when building designs using U-pseudo coding: The upper (“U”) bounded approach is advantageous when inverting regions in certain constrained mixture situations. However, due to axis flipping, it’s easy to misinterpret trends when viewing a trace plot without this new feature.

  • Increased limit on components for screening and historical* designs. Design-Expert now handles up to 50 individual ingredientsup from 40 and 24, respectively.
    *(An example is happenstance data collected by assaying retained samples from a period of material production.)

More choices when custom-designing your experiment

  • D-, IV-, and A-optimal design selection: New and expanded criteria when crafting experiments to models of choice within realistic constraints.

  • Constraints calculator: Simplifies derivation of constraint inequalities. At right, food scientists cooking starch must bake it longer at low temperatures. With program Help guidance, the design space’s lower left corner can be excluded using a multilinear constraint equation generated from a few user inputs. An optimal design is then fitted to this region.

  • Tolerance-interval-based design sizing: Enhances your fraction of design space (fds) plots to assess whether your planned experiment is large enough, given the underlying variability (noise), to establish tolerances within the acceptable range.

Additional statistics adn more concise reporting of vital results

  • Improved curvature testing for factorials with center points: All design points are now fitted to the polynomial model used for predictions. This provides a more realistic impact of significant non-linear response behavior. Diagnostics can be done for the model adjusted for curvature or, via a view option, unadjusted. Models without a term for curvature (unadjusted) are used for model graph and point predictions.

  • Coefficients summary: After modeling your response(s), see a concise table of coefficients that’s color-coded by relative significance.

  • Condensed “Fit Summary” table: See vital details on model choices before delving into all the particulars.

  • Tolerance interval (TI) estimates on point prediction: This is important for verification studies to ensure your process stays within manufacturing specifications.

Increased visibility and versatility of tools and features

  • Many new, high-visibility tools: Options previously available via hidden View menu options are now easily seen and capitalized upon.

  • Design layout column widths now adjust automatically by double-clicking column-header boundaries: Multiple columns adjust simultaneously!

  • Attach row comments by right-clicking on row headers: A handy way to record important observations, as shown below.

  • Topic Help, Tutorials, and Sample Files now also reside in the main Help menu: Follow these alternate paths for getting timely program advice.

  • Screen Tips is now a main menu item (“Tips”): Great visibility and easy access to very useful just-in-time advice, shown below.

  • Response surface method (RSM) models can be fitted with factors in their actual levels: This enables no-intercept model functionality.

Enhanced design evaluation

  • Several new matrix measures are now provided: Most notable is the G-efficiency. (This criterion, expressed on a 0 to 100 percent scale with higher being better, leads to designs that generate more consistent variance of your predicted response. However, like any other single measure, it may not accurately reflect the overall effectiveness of a particular matrix. That’s why Design-Expert provides an array of matrix statistics and graphics for overall design evaluation.)

  • Fraction of paired design space (FPDS): This resourceful tool lets you assess the power of RSM or mixture designs to detect specified signals (response differences judged important) in the presence of noise (system-standard deviation).

  • New, powerful tools for multiple response optimization: Options include standard error models. All else equal, choose system settings in regions predicted to exhibit the highest precision.

Many things made nicer, easier faster throughout the program

  • One-click updates: Check for free releases with one press and download them directly.

  • Better defaults and tick marks: Nicely rounded values provide presentable graphs straight away.

  • Zoom up graphs with your mouse wheel (a right-click resets to original size): Quickly zero in on regions of interest.

  • Hold down your left mouse button to drag graphs into various positions (a right-click resets original placement): It’s a fast way to situate the region of interest where you want it in the coordinate space. Components G and H in the mixture trace plot at right are constrained to very tight ranges relative to other ingredients. They are hardly visible without first zooming and then dragging the intersection (the overall centroid of the formulation space) to the middle.

  • Separate preference tabs for X-Y versus surface graphs: DX8 delivers plotting and graphing simplicity.

  • Reduced graph-updating flicker: Now it’s less distracting when you redraw responses at varying input-variable levels.

  • Categoric factors (established via general factorials, for example) are now convertible to discrete numerics: This lets you apply response surface methodologies while adhering to processes that run most conveniently only at specific settings.

  • Color-by-point-type added to graph columns: Very useful addition to scatter-plots, such as this one below for a central composite design (CCD).

  • Upgraded MFC (Microsoft Foundation Class) common controls: This new application framework provides an improved look and feel.

  • XML utility offers new script feature that lists all possible commands. You can parse files with extensions other than .xml. It also provides new import/export/reset-preference commands: More power to operate Design-Expert programmatically.

Back to Top