Somewhere in the dusty archives of the internet, there is reference to the “Committee on Boring Questions” or “The Ministry of Boredom”… or something like that. Their mission statement is “To answer seemingly boring questions”… or words to that effect. You can probably tell I have not been able to re-enter the murky depths of the internet to find these people again. And that’s a real shame, because I really liked their mission statement
The operative word in their mission statement (as I remembered it) is “seemingly”, because their implicit goal is to show that real questions are not boring, no matter how boring they superficially seem. One question I remember they tackled was the issue of whether milk tasted the same from different supermarkets. That is actually an interesting question; to answer it, they had to unpack an as-close-to-objective-as-possible way to measure taste. And if the answer was no they have to unpack why not. In turn, that shows us a little of the dairy industry.
If anyone thinks they know what I’m talking about and can link me to it, I would be much obliged.
I have a question, under a similar mission statement, I’d like to investigate. The implicit goals are to illuminate parts of the scientific method and perhaps to go some way towards explaining why some scientific experiments just haven’t reliably taken place. The question is:
Does eating a teaspoon of local raw honey every day help stop hayfever?
Many people swear by this. But there aren’t any experiments big enough to reliably say it works. There are plenty of anecdotal accounts of it working, but the plural of “anecdote” isn’t “evidence”, is it? The truth is that anecdotes can be evidence, but you have to be clever about how to use anecdotes.
Think about how you would measure the effects of honey on the symptoms of hayfever? Would you even measure the symptoms? You could go straight to the antigen levels in test-subjects blood, analysing the biochemistry of hayfever directly. That is a lot more work, and relies on expensive medical equipment. The options for data collection seem to be to record the symptoms through some sort of pollen-diary (which can be done remotely, making the investigation less invasive into the subject’s life), or to do blood tests (but that is invasive and limits you to a local sample).
If you use this year as your baseline year, next year will show an improvement no matter what; this year (in the UK) has had the worst pollen count in 50 years. The climate is a compounding variable. The other compounding variable is the real exposure to pollen; what if people don’t go outside? Perhaps you won’t choose to do a baseline test. Instead you will simply have a control group (no treatment) and a test group (honey all summer, winter, autumn and spring, ready for next year). But then you have two complications: ethics and exclusion.
The ethical dilemma is simple: can you demand some people not treat their hayfever symptoms? There is a related practical issue: if people have promised they won’t take any treatment, do you trust them to follow that when they are blinded, swollen, itchy, bunged up and in pain?
The exclusion criteria is more difficult. Before reading on, can you guess what kinds of people would need to be excluded from the trials? Hayfever will have to be your only known allergy, else your data can be skewed by dust allergies, pollen-induced asthma and other allergies. They would make me and my girlfriend viable test-subjects.
There is a way around this: do a retrospective study comparing those who do have locally produced, raw honey and those who do not, as a part of their daily routine. But then you have a self-selection issue: the only people who take honey for that purpose are people who think it works, for whatever reason. It might be that only people who have outgrown their hayfever take the honey, because they have convinced themselves honey is the reason they don’t suffer anymore. The other problem is that you don’t have any data on how severe the hayfever was before they self-treated with local raw honey, nor do you control any other treatments they take.
If I were to do this (and sell my finding to the “Committee of Boring Questions”) I think I would focus more on inclusion criteria. My study would be done on people with severe hayfever allergies. The definition of severe, in this case, would be that symptoms still persist when treated with 2 varieties of double-dosed antihistamines, nasal steroidal and nasal barrier sprays and steroidal eye drops. (I would also like to define “pathological hayfever” as when the symptoms that persist through the treatment I just described in such a way that significantly interferes with daily function. I have severe hayfever, my girlfriend has pathological hayfever. I made these definitions up).
The benefit of this kind of inclusion criteria is that subjects are permitted to to go about their normal hayfever treatment, and there will still be measurable symptoms; there is not the same ethical concern. There can then be a control group that goes about their normal treatment and a test group that tops-up treatment year-round with honey.
How many people will this study need to be reliable? 20 people in each group is a reasonably priced university study, but damned if the NHS would use it to support recommending honey. 5,000 people in each group is a bit more like it, but collecting all that data is suddenly a team effort. The question now is what team on the face of the planet cares enough to do it, has the resources, and doesn’t have a vested interest?