Given the emphasis on STEM education these days, it's fantastic to see a middle school student getting hands on experience doing experimentation. However, because I'm sure this student is also reading the comments, including your rebuttal remarks about pesticides, I think some "factual clarification" is in order. You purport that "larvicides" do not affect adults, having argued that insecticides "fall into two separate classes, either killing insects in their larval or their adult stages." This statement is incorrect. Both larvae and adults are susceptible to toxicity by the currently registered insecticides, although the potency towards larvae and adults might be different on a body weight basis. Also, potency may be different owing to pharmacokinetic considerations. Nevertheless, both immature and sexually mature life stages are susceptible at some dose. Even your own statement gives evidence to this fact. You speak about acetylcholinesterase inhibiting insecticides, which all have a common biochemical mode of action if they are organophosphorus esters, so why would the larvae not be susceptible to the insecticide (or vice versa, why would the adult not be) given the commonality of neurophysiological mechanisms among all animals.

The potential confounding by insecticides is certainly something to think about, but one way to begin probing that possibility is to check with the USDA Pesticide Data Program databases to garner an idea of what kind of residues (which will be mostly insecticides) are found on various commodities. You could also check with the FDA Pesticide Surveillance Program. These databases are easily accessible on the internet.

It is no longer true that OP insecticides (AChE inhibitors) are the most used insecticides. Only a couple of products in this group are still used. The chemical classes used today have changed fairly rapidly and now include neonicotinoids, pyrethroids, spinosyns, and several other groups. If you do probe the USDA Pesticide Data Program database, you'll have a pretty good idea of what residues are being found. Fungicide residues are also found but not too the same degree as insecticide residues. Although these insecticides do have biochemical mechanisms on physiological systems that are commonly shared by the invertebrate and vertebrate nervous system, the difference in potency at the receptor or enzyme target sites is quite large (often orders of magnitude), making the compounds much more insect specific than vertebrate effective.

Regarding the results, I'm not so impressed by statistical significance, because other than the effect of soybean cultivars on survival, the other survivorship curves are visually quite similar. Furthermore, sometimes "organic" food was favorable and sometimes not. However, if you really wanted to follow up on one food, the soybean response was quite distinctive. Soybean production, which is dominated by cropping systems in the Corn Belt, only infrequently use insecticides. Weeds are the biggest pest, and the present systems rely on glyphosate herbicide. That chemical shouldn't be toxic to Drosophila given a mode of action inhibiting aromatic amino acid synthesis, but the latest USDA Pesticide Data Program report shows that glyphosate residues are prevalent at ppm levels in harvested seeds (compare that to low ppb levels for insecticides).

The research presented has perhaps generated a hypothesis, and other comments about cultivar differences might be germane for interpretation with respect to antioxidant levels. However, plenty of literature has been published that does not support the blanket statement that certified organic produce has greater antioxidant levels. These studies are plagued by problems with cultivar differences and field location differences. Even the same cultivar grown in different soil types will have variable levels of antioxidants.

The conclusion seems to be that the effect observed is real, meaning it must have a physiological cause related to the treatment effect. Before going off on a wild goose chase, why not repeat the experiment to see if the result is consistent and repeatable?