The problem:

I have developed a system of nonlinear differential equations that predict the time series neural response from the stimulus. I am looking for a goodness of fit measure between the actual observed time series of the nerve and my model’s prediction to quantify how well the model is working in some sense.

I would be grateful if anyone could point me in the right direction, thank you. ]]>

it's been awhile since I've done stats and everything has completely escaped me. I'm trying to conduct hypothesis testing for the question below. Everyone has been using the Chi-squared test and I feel that they are wrong. I think it should be a poisson distribution but the change in the control is throwing me off. Any help would be greatly appreciated! Thanks a million! Here's the question:

Suppose there has been the introduction of treatment to reduce the number of accidents at the Nathan road-waterloo road junction (treated site). Also suppose the junction of mongkok road and Shanghai street has been chosen as the control site. After in months, the accident frequencies of these two junctions have been tabulated below

Before After

Nathan 30 18

Mong Kok 15 12

Use an appropriate statistical test to test whether the treatment has been effective.

I'm assuming the change in Mong kok is just a natural change as it hasn't been treated.... so we're just weighing the treated site against the natural change

I did an initial hypothesis of

Ho:u>=uo

Ha:u<uo

Thanks in advance for any help. :D ]]>

I just stumbled across your site; I am learning statistics by myself for my thesis (Bear with me please and sorry if I posted this in the wrong place). I am working with geochemical data (which I have both raw and transformed data) for 6 different groups of rock. I've noticed that the lower 3 units show a distinct difference in mean, range, standard deviation, and median values of a specific recorded geochemical value, but is there a way to show that the mean (or any other parameter) is statistically different from the lower 3 to the upper 3? I've looked at the t-test (can you only compare two groups at a time?) and P-value calculations. I was told to do a chi-squared test, but I am confused as to what I would record for observed and expected values. If anyone could shed some light on this topic or provide a better method of approaching this, it would be gratefully appreciated. ]]>