Home / DOE / Create Designs / Design Wizard

Design Wizard¶

QXL DOE New > Design Wizard

The wizard asks one question per page and picks the design type for you from the answers. It builds the same designs the Create Design menu builds, and writes the same DOE Design worksheet.

Moving around¶

A strip of step buttons runs across the top of the dialog, and clicking one jumps straight back to that page. The steps are Levels, Number of Factors, Categorical, Designs, Center points, Factors, Replicates, Outputs and Runs. They are revealed one at a time as you advance, so the strip shows how far you have got as well as where to go back to. Back steps back one page and is disabled on the first.

Finish appears four pages before the end, from the Factors page onward, and taking it early fills in the pages you skipped without asking:

Skipped	What is used
Replicates	for a 2-level factorial, the replicate count the software suggests for that run count; for anything else, 1
Blocking	one block, blocked on replicate
Folding	no fold
Outputs	a single output named Output 1
Randomize runs	the current Run order for new design setting from the Options dialog

The questions, in order¶

The first page is How many levels? and it has three answers: Two levels, Three levels and Mixed. That answer decides the whole route through the rest of the wizard.

	Two levels	Three levels	Mixed
1	How many levels?	How many levels?	How many levels?
2	How many factors do you have?	How many factors do you have?	Enter Factor Names and Coding
3	Choose a n-factors design	Do you have any Qualitative or Categorical Factors (Inputs)?	Enter number of replicates and blocks (optional)
4	Enter Factor Names and Coding	Choose a n-factors design	Define outputs (optional)
5	Enter number of replicates and blocks (optional)	Enter Factor Names and Coding	Do you want to randomize runs?
6	Define outputs (optional)	Enter number of replicates and blocks (optional)
7	Do you want to randomize runs?	Set number of center points and alpha value, only when you answered No to the qualitative question
8		Define outputs (optional)
9		Do you want to randomize runs?

How many factors do you have? carries one control, Select number of factors:, and the range it accepts depends on the route: 1 to 19 on the Two levels route and 2 to 23 on Three levels. Mixed has no such page, taking you straight to the factor grid. A value outside the range is refused with Number of factors must be a number between n and m.

Two of those pages need explaining, because what they do is not obvious from the wording.

Do you have any Qualitative or Categorical Factors (Inputs)? is asked on the three-level route only, and answering Yes removes the center points page. The page explains the term on screen: categorical or qualitative factors (inputs) are those that take on values that are names or labels, such as Gender (Male, Female) or Color (Red, Green, Blue).

Set number of center points and alpha value appears only on the three-level route and only when you answered No to the qualitative question. When the design has no alpha to set, the same page appears with the shorter heading Set number of center points.

Choosing the design¶

Choose a n-factors design lists the designs that fit the number of factors you gave, with a comment about each one beside the list.

With one factor the wizard stops and hands over. The page reads One factor, and finishing closes the wizard and opens the Statistical Tools hypothesis test wizard instead, because a one-factor experiment is a hypothesis test rather than a design of experiments. No DOE Design sheet is written.

This is reachable only on the Two levels route, whose factor count starts at 1. The Three levels route starts at 2, and Mixed has no factor-count page at all, so neither can arrive here.

The last two pages¶

Define outputs (optional) is the same page the builder uses, and the word optional is accurate: you can finish without naming an output and add it later with Modify Design > Add/Remove, Change Name/Type. See Define outputs.

Do you want to randomize runs? is Yes or No, with Use this setting always beside it. Ticking that box writes your answer into the Run order for new design setting in the Options dialog, so every later design follows it without asking. The page also states why the question is there: randomization is a way to reduce the effect of external noise on a design.

What you get¶

Finish writes a new worksheet called DOE Design, or DOE Design (2), DOE Design (3) and so on when a sheet of that name already exists: one row per run, one column per factor, the columns for you to enter the responses into, and a description of the design in text. Nothing is calculated until you enter response data and run the regression.

How many columns an output gets depends on the replicate format, and it is rarely one:

Format	Columns per output
Stacked	one data column, plus the weights or events and sample size columns if you ticked them
Table, quantitative output	one column per replicate, plus Y-bar, plus S when there is more than one replicate. The per-replicate columns are headed Y1, Y2 and so on, or block1, block2 and so on when the design is blocked with replicate into more than one block. Y-bar and S arrive carrying `=AVERAGE` and `=STDEV` formulas, so they fill themselves in as you type. With a single replicate it is two columns
Table, binary or nominal output	one column per replicate, and never fewer than two