This Demonstration considers the effect of interaction in a two-factor factorial experiment.
A regression model representation of a two-factor factorial experiment can be written as
represents the response,
factor 2, the
s are parameters with values to be determined and
is a random error term.
If you increase the value of the interaction coefficient
, the twisting of the response surface increases . This is most apparent in the curvature of the contour lines in the 3D and contour plots.
 D. C. Montgomery, Design and Analysis of Experiments
, 9th ed., Hoboken, NJ: John Wiley & Sons, 2017.