Subscribe Bookmark RSS Feed

Analysis of seaonal data, help?

I have fish numbers ina lake rising to a peak and falling off over a number of weeks. I also have plankton, temp, pH, and other parameter data. Can JMP test for which parameters influence fish numbers the most, if so, how? Thanks.
1 REPLY
XanGregg

Staff

Joined:

Jun 23, 2011

Look into Time Series in JMP. Here is what I got from a statistician:

Time Series can model the phenomenon via transfer function modeling, in case the fish number is the output, and plankton, temp, pH are inputs. However, the answer to the question “whether JMP can test significance or not” is YES AND NO. The approach the time series platform uses is model selection; for a single model, the report gives simple T-tests for individual coefficients. Unless the model is equivalent to a regression model with auto-correlated errors, the report won’t tell much about “which factor influence fish numbers the most”. Therefore, in general, the work flow is to fit a batch of plausible models first, then select the best one or several among them, based upon which a conclusion might be drawn.