 |
Design Express is an easy to use PC Windows package that integrates with Compusense five to deliver statistically valid designs that are specifically tailored for the consumer or sensory scientist.
Have you ever tried to design a fully-balanced round-robin experiment?
Are your designs balanced for carry-over effects and order effects?
Have you ever tried to balance order of trial in an incomplete block design from a textbook?
Design Express is the answer to all of these problems and more.
Design Express has been designed by statisticians for use by non-statisticians. At the touch of a button the software chooses from a wide range of designs for single and multi-session experiments.
|
|
Design Express is easy to use:
- PC Windows format
- Easy specification of requirements
- 'Blinding' codes generated randomly in a variety of formats or can use existing samples codes
|
- Designs can be copied directly into Excel, with its associated functionality
- Designs can be saved for import to sensory data collection packages
- Design properties can be readily inspected: order, concurrence, carry over
|
|
Design Express main features:
- Complete block designs for single or multi-session experiments
- Incomplete presentation designs for single or multi-session experiments
- Designs that have balance for presentation order effects
- Designs that have balance for carry-over effects
|
- The software draws from a comprehensive library of over 400 balanced incomplete block designs
- Optimal search algorithms for best possible balance when library designs are not available
- Allows designs with up to 51 samples and 435 consumers
|
|
A unique feature of Design Express is the ability to construct appropriate designs when a complete set of samples is presented to all consumers over a number of sessions.
For example in the view of the "IBD Selector" screen, the user has specified that there are 12 samples and that a design is required for somewhere between 122 and 152 consumers. The list of designs on the left indicates that a range of resolvable balanced incomplete block designs exist all for 132 consumers.
In particular the desired multi-session experiment may be achieved in four different ways, either presenting
- 2 samples in each of 6 sessions (~6)
- 3 samples in each of 4 sessions
- 4 samples in each of 3 sessions or finally,
- 6 samples in each of 2 sessions.
The user has highlighted the "present 4" samples design, has checked "Expand" to spread the full set of 12 samples over three sessions and finally has requested it to have order and carry-over balance.
Note that if 132 consumers were not suitable for some reason a much wider selection of designs are available using the optimal search method, these designs are listed on the right side of the screen.
For more information, please visit the Qi Statistics web site
|
