Hi all I am struggling to create a response surface for some data I collected. I have been asked to create a response surface to get a better understanding of how the two inputs of interests affect the output. I threw together a multiple linear regreession model since it was simple but my graphs do not feel right. for one thing the surface is planar. I know it was a linear regression but I feel that it is an over simplification. I guess I am looking for a more interesting surface and approach to creating this predicitve model, when you only have a bit of data. Any ideas? For the record I am an engineer with very limited exposure and understanding of stats so dumb it down for me. thanks in advance.