POSSIBLE SOLUTION(S)

We enter our data and plot it.

The conjectured model!!!

We formulate a Least Squares function.

We find the best fitting (in least squares sense) parameters a and b.

We build our best fitting model with parameters a and b.

And plot this best fitting function

over the data to see how good the fit looks to us.

Now as to parameters for comparison, one could use the "steepness" in the center of the fitted plot, i.e. find where the first derivative is maximum and compare maximum first derivative values. This idea comes from the fact that the perfect knowledge model has "infinite" steepness at the 100 g mass.

Now compare this maximum value of u'[t] to other comparable values for other populations.

Of course, one could use the sum of the squares obtained as the first output in the FindMinimum command to compare groups, but perhaps it would be better to compare the sum of squares of the differences between the group's response and the perfect response.

For that purpose we need the "perfect response" function.

We note that the SumOfSquares between the perfect response and the guessing response would be

Every group should do better (less) than this!!!