I need help with discussion board

If the tentative assumption of a first-order model is reasonable, there is little point in
using eleven different experimental levels. Of course, we need at least two levels to
estimate the parameters in the model, and at least one more to detect curvature in
the true model, if curvature exists. By taking repeat observations at some or all levels,
we can obtain a pure error estimate of (1″2 to use in checking lack of fit. Thus for an
experiment of about the same size, one possibility would be to choose three widely
spaced levels-the extremes of the X-range and the center, for example-and to take
four observations at each of these levels. This would lead to an analysis of variance
table of the form below. If the tentative assumption of a first-order model is reasonable, there is little point in
using eleven different experimental levels. Of course, we need at least two levels to
estimate the parameters in the model, and at least one more to detect curvature in
the true model, if curvature exists. By taking repeat observations at some or all levels,
we can obtain a pure error estimate of (1″2 to use in checking lack of fit. Thus for an
experiment of about the same size, one possibility would be to choose three widely
spaced levels-the extremes of the X-range and the center, for example-and to take
four observations at each of these levels. This would lead to an analysis of variance
table of the form below. If the tentative assumption of a first-order model is reasonable, there is little point in
using eleven different experimental levels. Of course, we need at least two levels to
estimate the parameters in the model, and at least one more to detect curvature in
the true model, if curvature exists. By taking repeat observations at some or all levels,
we can obtain a pure error estimate of (1″2 to use in checking lack of fit. Thus for an
experiment of about the same size, one possibility would be to choose three widely
spaced levels-the extremes of the X-range and the center, for example-and to take
four observations at each of these levels. This would lead to an analysis of variance
table of the form below.