To obtain the curvilinear regression I did a linear followed by a quadratic regression in SPSS but I cannot find a way to do the linear broken line analysis in SPSS. I want to be able to compare results obtained with curvilinear regression and broken line analyses. There must be a way to do this in SPSS or in an open source software but I only found publications using SAS, and I do not have this software.
Any idea how to do this?
It doesn't seem to work maybe because I have many values for each concentrations. Not sure what Iam doing wrong.
No point getting fixed on broken line. depending on your data configuration, you can write an exponential model which is implementatble in SPSS. then you would have at least some comparisons. Since your data fitted quadratic, it may likely fit some exponential models too. please check.
Thanks a lot, will try this. How will I get the requirement using the exponential model? Will it be the value closest to the asymptote of the plateau?
the classical approach is to use 95% of the asymptote. requirement is a tricky computation that is still a subject of intense conversation among nutritionists and physiologists. Since mathematics and biology could be a world apart in many regards. another dimension is the economics of the nutrient in question, as noted in a specialisation like mine. Even within the mathematics of it, different models may indicate different asymptote. Thus many (and rightly so) are of the opinion that whatever you obtain is only valid for your study condition. So in a sense the animal can never have a requirement, thus you need to qualify your figure(s) by stating the condition(s) of your study.
Thanks a lot for the info. Will use 95% asymptote and check with other computation to see if it differs a lot or if requirements are similar with different ways of calculation.
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Atmosphere