Question

Please help! In cross-predictions for biology, how do I account the difference between expected values and the experimental values? For example, (purple smooth = 75, white smooth = 28, purple wrinkled = 24, white wrinkled = 8).

Answer 1

Before you start accounting for variation, you want to do a Chi Square goodness of fit test to see whether the difference between the observed and expected values are actually statistically significant.

If the differences are not statistically significant, then the variation is due to random chance events among the population in question. Such as particular gametes happening to be fertilized by particular other gametes, at an individual level.

If the differences are statistically significant, then you start wondering about whether the genetic relationships you have assumed among the genes and alleles are actually true. Or whether you need to come up with some alternative scientific explanation for what you're seeing.

And to muddy that nice discussion, what you use as a threshold to determine statistical significance - i.e., the amount of variation that could be due to simple biologic randomness and stochastic effects rather than underlying principles - is an arbitrary choice you make.

Is that more or less helpful?