Model fitting nlparci - interpreting confidence intervals/getting p values.

Hi,

I have some code to fit data with lsqcurvefit. I then calculate confidence intervals for the parameters using nlparci. The data that I have refers to two different conditions, so I have fitted the data using the following format:

k1 = exp(log(k1)+log(ratio)*logical(is_condition_2?)) 

So that k1 = k1 or ratio*k1 if the data is from the second condition.

I have 95% confidence intervals for the ratios from nlparci, but I would like to find out if the ratios are significantly different from 1 (i.e., can I say that the parameter change is significant between the conditions). Is this just a case of seeing if the 95% CI includes 1, or is there a different test required?

Many thanks.

Answer 1

[Neha]

Hi,

To determine if the ratios are significantly different from 1, you can indeed check if the 95% confidence intervals include 1. If the confidence interval includes 1, it suggests that the parameter change is not statistically significant between the conditions and then you can reject the null hypothesis that the ratio is equal to 1 at the 0.05 significance level. On the other hand, if the confidence interval does not include 1, it indicates that the parameter change is statistically significant. You can also refer to the following MATLAB Answer to understand how to interpret nlparci results:

https://www.mathworks.com/matlabcentral/answers/155649-how-to-interpret-nlparci

Hope this helps!