I am developing an artificial model to simulate the growth of two types of biological cells under different conditions.
The data I obtained from my model takes the form of two data-sets representing the number of cells in the culture across time in what we will call condition A and conditon B.
Ie:
A = {A1, A2, A3, A4, …, An}
B = {B1, B2, B3, B4, …, Bn}
For 1 <= t <= n
I now need to determine whether the two curves are statistically different, that is, whether growin the cells in conditions A/B produces any meaningful changes.
I am new to statistics, I have tried reading some papers on curve regression but got stuck.
Best Answer
I've been asking myself a similar question recently. I found this paper by Motulsky and Ransnas (1987) to be helpful.
Page 371 talks about fitting the same general model on two data sets as a way to determine whether the two data sets differ. They briefly discuss two ways to go about this.
One approach requires repeating your experiment several times, fitting the model to the two data sets for each time you do the experiment, and then comparing the parameter estimates between the two sets of models using paired t-tests.
The other approach is to fit the data together and then separate and then see if fitting the data separate is "worth it", so to speak, via an F-test.
This is at least my non-statistician's understanding of the paper. Hope that helps and good luck!
Similar Posts:
- Solved – Identifying the experimental unit
- Solved – What experimental design is this
- Solved – Pooling data from replicate experiments
- Solved – Biological experiment – statistical significance of 6 treatments vs control – ANOVA
- Solved – Averaging data then fitting vs fitting then averaging in non-linear regression