1
Hi, fellas.
I measured the air temperature every hour of a day in two treatments: below the canopy of 10 plants and in 10 open areas at a standardized distance from each plant. Therefore, I obtained two regressions between the hours of the day (predictor) and the air temperature (response). The relations between the variables have bell shape, so I need to adjust polynomial models. Then I need to compare the slope of the lines. I can use an ANCOVA to evaluate polynomial regressions and select the best model, using p<0.05 and higher value of R²?
ancova_null_model=lm(Response~1*Category, date)
ancova1=lm(Answer~predictor*Category, date)
ancova2=lm(sponse~Poly(predictor, 2)*Category, date)
ancova3=lm(sponse~Poly(predictor, 3)*Category, date)
Anova(ancova_null_model, ancova1, ancova2, ancova3)
summary(ancova1)$r.squared
summary(ancova2)$r.squared
summary(ancova3)$r.squared
Or should I make two regressions and compare the inclinations using another analysis? In this case, how can I compare the polynomial regressions slopes in an analysis?
All contributions are welcome!
Unfortunately, hardly anyone will answer this question here, as she runs away topics that can be asked in the OS. What I suggest is asking this question on the list R-br, for statistical theory is not an off-topic subject there.
– Marcus Nunes
R is a language for specific purposes but it is a language and the question (badly formulated no doubt) is still about programming.
– user1084
Hi, Marcus. Thanks for the suggestion. I’ll send my questions to the R-br list. Hug!
– RRMoura