Binary response with temporal explanatory variables
Problem description
240 plots (with a 160 soybean plants) in a field where we measure a continuous response for each of 8 wavelengths repeated over 12 days that are close but not quite equally spaced. In addition, we measure a binary SDS diseased/not diseased status. Both the wavelength measurements and diseased status are measured at the plot level.
Client wants to do a functional analysis based on a suggestion from their advisor.
Scientific question
Want to classify disease status based on wavelength information.
Questions
Will disease status of plots affect nearby plots?
Advice
Use final disease status as the response. For each timepoint and wavelength, perform a t-test with the two groups determined by disease status. See how the significance changes temporally.
blog comments powered by Disqus