Center-point curvature contrast Created 2026-10-14 Updated 2026-11-05
For a two-factor factorial design with coded corners , observe each of the four corners once and add independent center runs at . Assume independent Gaussian errors with common variance . Write for the corner average, for the center average, and for the unbiased center-run sample variance. Then the center-point curvature contrast is
For the full quadratic regression mean , the contrast has expected value . The corner average cancels the linear and cross-product terms, while both squared coordinates equal one at every corner and zero at the center.
Under the null ,
The independent normal contrast and center residuals justify this exact F-distribution; a first-order mean model alone would not imply it without the error assumptions. The center observations supply pure error, while the extra corner-to-center comparison detects one aggregate curvature direction.[ref]
Zero contrast does not exclude curvature. The mean has a saddle shape but gives the same mean at all four corners and the center. A central composite design adds axial runs that can distinguish the separate squared terms.
Central composite design Created 2026-10-14 Updated 2026-11-05
A central composite design combines a factorial portion with axial points and center runs. For a full factorial design, quantitative factors and center runs, its run count is . The factorial portion supplies main effects and interaction terms, while the axial runs allow the individual squared terms of a full quadratic regression model to be estimated. Additional center replicates estimate pure error.
For two factors and , this gives runs. The four corners are , the four axial points are and , and the remaining five runs repeat . Choosing makes this full-factorial design rotatable. Choosing keeps the axial points on the faces and uses only three levels per factor, but does not give that rotatability in dimension at least two.
Circumscribed designs place the axial points outside the original factorial cube; inscribed designs rescale the cube inward to respect physical factor limits. A fractional factorial portion can reduce runs, but must retain sufficient resolution of a fractional factorial design to estimate the desired interactions: axial points have zero cross-products and cannot separate interactions that remain aliased in the factorial portion. The institutional handbook describes augmentation from a full or resolution V factorial portion.[ref]
For the general normal linear model, minimizing gives the least-squares normal equations . Full column rank makes invertible, so
The mean and covariance follow by substituting and using , . Normal errors also give the exact multivariate normal distribution of this estimator.
Use coded experimental variables
Their origin is the middle of the allowed rectangle, and the stated endpoints become . The four corners and center runs give mutually orthogonal columns , with
Writing when , the ordinary least squares estimates are
Their variances are , , , with zero covariances.
The repeated center runs estimate pure error without requiring the fitted mean surface to be correct:
They also permit a center-point curvature contrast. Let and . Under the first-order model, and . Under a quadratic surface, its expectation is , since the linear and terms average to zero. Thus
under the first-order normal model. With only one center run there is no pure-error degree of freedom. A zero center contrast does not exclude all curvature, because the two quadratic coefficients can cancel. The corners also estimate an interaction contrast; the full Lack-of-fit F-test compares the first-order fit with the five setting means, with two lack-of-fit and pure-error degrees of freedom.
If the first-order model is adequate and , steepest ascent in response surface methodology uses the unit coded direction . The Cauchy-Schwarz inequality proves that it maximizes the predicted directional derivative. Run sequential settings for increasing feasible , corresponding to physical increments from the center. Stop or reduce the steps when measured yield ceases to improve, and refit locally near promising settings. Respect the allowed rectangle; if a boundary is reached, investigate feasible boundary directions rather than extrapolating outside it. A near-zero gradient offers no first-order ascent direction and calls for further modeling.
If curvature is indicated, add the four axial settings , together with further center replication. This face-centered central composite design stays inside the allowed physical ranges and separates the two squared coordinates, which the corners alone cannot do. Fit the full quadratic regression surface
Write its fitted quadratic part as , where , and . Its gradient and Hessian are and . If is nonsingular, the stationary setting is
If is negative definite and is feasible, it is the fitted maximum. Otherwise compare feasible stationary settings, optima on each edge obtained from the restricted one-variable quadratic, and the corners. A singular or poorly determined calls for attention to ridge directions and more runs. Convert any chosen coded setting back using , , and confirm its yield experimentally. The fitted local surface guides the search; it does not establish an untested physical optimum.
For the initial normal linear model, the design matrix has rows . The eight factorial corners have balanced signs and orthogonal coordinate columns; the six centre points contribute only to the intercept. Consequently
Using the least-squares normal equations gives the least-squares estimator
It is a linear transformation of independent normal responses, so its sampling distribution is
The four estimated coefficients are independent, since they are jointly normal with diagonal covariance matrix.
Adding the axial settings produces a central composite design: eight factorial corners, six axial points and six replicated centre points. The axial distance is the rounded value of . With that exact distance it is a rotatable design for a full quadratic model. Indeed, sign symmetry annihilates odd moments, all coordinates have the same second moment, and the fourth moments satisfy
Rotational symmetry requires the first to be three times the second, giving . Using the printed rounded distance makes rotatability approximate, rather than mathematically exact.
The full quadratic regression mean has ten coefficients:
The expanded design matrix has rank ten. To verify this rather than just count parameters, suppose such a quadratic vanishes at every setting. Its centre value sets its constant term to zero; subtracting the values at opposite axial points sets every linear coefficient to zero, and adding them then sets every squared-coordinate coefficient to zero. On the factorial corners, orthogonality of the three pair-product columns forces their coefficients to zero as well. Thus all ten columns are independent, leaving residual statistical degrees of freedom.
There are distinct settings: eight corners, six axial points and the centre. The six centre replicates give the only pure error, with degrees and sum of squares . The saturated setting-mean model has parameters, so the lack of fit beyond the quadratic model has degrees. For any setting with replicates, the identity
splits the residual sum of squares into pure error and lack of fit; the cross term vanishes because the deviations from the setting mean sum to zero. Summing gives
Under a correct quadratic mean with independent errors of constant normal variance, the pure-error and lack-of-fit projections are orthogonal, hence their sums of squares are independent with respective distributions and . The Lack-of-fit F-test therefore uses
This is below the given and critical values and . There is no significant lack of fit at either level. That is failure to reject this quadratic mean, not proof that it is exact or that a maximum has already been located. The pure-error denominator also assumes the error variance at the centre represents the common variance throughout the experiment.
A natural next fit is quadratic regression, which remains a normal linear model in its unknown regression coefficients:
The new design matrix has rows and must have matrix rank three; three distinct predictor values suffice. The curved regression residual pattern suggests trying a positive quadratic term, but its sign and adequacy should be checked after fitting. Inspect the new regression residuals to see whether the systematic curvature has disappeared.
Residual curvature diagnostic Created 2026-10-07 Updated 2026-11-05
A systematic curved pattern in regression residuals plotted against fitted values suggests a misspecified conditional mean. Positive residuals at both ends and negative residuals in the middle motivate trying quadratic regression. The diagnostic does not by itself identify the error probability distribution or prove heteroscedasticity.
Response surface methodology Created 2026-10-14 Updated 2026-11-05
Response surface methodology fits local regression approximations to a quantitative response as experimental settings change. A first-order model estimates a local gradient; augmenting the design permits quadratic regression, curvature assessment and local optimization. Extrapolating a fitted surface outside the experimentally supported region requires caution.
Rotatable design Created 2026-10-14 Updated 2026-11-05
A rotatable design for a specified regression model has variance of a fitted regression mean depending on only through its Euclidean distance from the design center. It concerns prediction precision, not rotational symmetry of the mean response itself.
For a full quadratic regression model in variables, consider all factorial corners, the axial points , and center runs giving a nonsingular information matrix of an experimental design. Sign symmetry cancels odd moments and permutation symmetry makes the second coordinate moments equal. Fourth-order rotational balance requires, for distinct ,
The factorial corners contribute to both sums, axial runs contribute only to the first, and center runs contribute zero. Thus , giving
For , ; for , . Center replication can estimate pure error but does not change this fourth-moment equation. If a fractional factorial portion is used, its own moments and estimability must be checked; simply substituting the original factor count into the full-factorial formula is not justified. In one dimension rotational invariance only means reflection symmetry, so the distinct-coordinate fourth-moment equation does not impose the displayed choice.