Science, exact and otherwise

Science, good science, always starts with an observation. If you cross a tall plant and a short plant, a monk once noticed, you get all tall plants. But if you cross those, you start to see short ones again.

Then comes a model. Observed traits are due to interaction of the distinct units of inheritance. Let's call them genes, shall we? Ok. A sexually reproducing organism gets one copy of each gene from mom and one from dad. Different copies of the same gene (let's call them alleles) can encode for differences in the same trait, like short and tall, green and yellow, round and wrinkled (that's pea shapes, you perverts, pea shapes, p-e-a). In each pair one trait can dominate the other, in a good way, of course, if one allele encoding each is present, and then the dominant trait will be observed. But if an offspring of these mixed organisms (called heterozygotes) gets two alleles of the gene encoding for the recessive trait, that trait will be seen again.

Models have to have predictive powers, i.e. you must be able to articulate a hypothesis consistent with your model of what will happen in a situation that was not part of the data set based on which you put together your model. Hypotheses never travel alone, by the way. Where there is hypothesis, there is always a null hypothesis, in short a NOT! to whatever your proposed hypothesis is. So if your proposed hypothesis is that if you cross two types of plants with two differential traits, say tall green and short yellow, you will get only one kind of a plant out, but if you cross these new plants among themselves, the progeny will distribute in a roughly 9:3:3:1 proportion of different possible combinations of observable traits, then your null hypothesis would be that the distribution of traits will not follow this pattern.

Then you conduct experiments to test your hypotheses. Sometimes there is more than then one working and one null hypotheses, and then you should take care to design your experiment to give you as unambiguous an answer to your question as possible. If the outcome of your experiment is consistent with your model, you say yay, do a jiggly dance in the middle of your lab or office, and design a new experiment to test another part of your model. If the results are inconsistent, you adjust your model, and take it from the top. You keep doing that until you feel all aspects of the model are sound. Then you spend weeks or more writing the whole thing up, and you send it to a journal and await the fate of your paper. Easy, right?

Well, that's until you attempt to think in this way about, oh let's say for the sake of argument, your own procreative history. You might suppose, based on the old semen analysis from before you had your daughter that your male factor issues are not necessarily new, although they might have gotten worse lately. You may then observe that the time you conceived your son it was after a twelve day business trip. You may throw into the mix, for good measure, the hypothesis articulated by the urologist that there may be anti-sperm antibodies circulating in the system due to an early adolescence surgical event (the answer to the question of whether or not the antibodies are present is in possession of the clinic, since JD did take the test, but since they are not talking to us, we do not as of yet know what the answer is). And then you may add the part from your time well spent in the exclusive company of PubMed that indicated that said antibodies are most strongly implicated in making it more difficult for the sperm to penetrate the mucous.

So then, in your spare time, you might devise the model that postulates that due to the male factor issues involved, your chances of conception would be increased if you did almost everything in the almost exact opposite way of what is recommended for your average no-known-issues TTC couple. The model would say to save up the goodies as opposed to regularly refresh the collection. It would also suggest that the best time to hit that would be in the exact one day before the egg white cervical mucous makes its glorious ovulation-heralding entrance (or is it exit?).

So imagine then that having constructed this hypothetical model, you start to suspect that your ovulation, having gone AWOL and caused you to seek the services of an RE in the first place, has now returned, slinked back, as it were, under the cover of other stressful life events. Imagine further, if you will, that life conspires in such a way that you end up performing an experiment to the exact specifications of your hypothetical model. And confirming, a few days later, that ovulation did indeed occur at the exact right moment from the point of view of said model.

Right, so... yeah... this is exactly where we are today. On CD24, 11dpo. I bought, yesterday, the C.ostco-sized pack of sticks, and even opened the pack already to facilitate easier access for tomorrow morning's ritual. But here's the thing-- doing actual science with a sample size of one and a trial number of one? Not exactly possible. Either outcome would be consistent with the model. And given that my boobs are not sore and my progesterone level on Monday, 8dpo, was measly 11.8, I think I know what that outcome will be. The year 2007 has one last eff-you in store for me-- CD1 on it's last day. But hey, at least I will be able to send this bastard off on it's way in appropriate fashion-- with copious amounts of alcohol.

P.S. The sticks I bought, 1st res.ponse, tell you to test on the day of the missed period. However, their sensitivity is supposedly the same as all the other tests, including the brand that tells you that you can test up to three days before. Any personal experiences with the different brands out there? Am I completely wasting my money by testing tomorrow, three days before the expected arrival? It's just that we are going to be at the ski house starting tomorrow night, and I would rather have the option to drink if I needed it, if you know what I mean...