The reach of science #3

Posted on January 20, 2011


Jenny: ” I get it. If you’re trying to map a large unknown space it takes a lot of observations to do it, and the map is usually obsolete by the time you finish if you ever do.  Particularly if that space is filled with humans who are moving around and changing, getting fatter or thinner, older, poorer or richer, healthier or sicker, some taking different pills, some on a good diet, some not, some with allergies some without. So if you want to test a new vaccine and you test it on a sample of several hundred who happen to be available, you never know how  a different sample  of people will react – people with  different diets, allergies, on different medications – whatever. And you don’t know what the long-term effects will be on that sample unless you follow them up for years….Notice the number of medications that are pulled off the market because of dangerous long term side effects. I see why science is a work in progress.”

Peter: ” OK, so you say science has trouble getting a trusted map of a large unknown space over time. They have to rely on small samples and anyway the territory may be changing, so even if you got a lot of sample this year things can change and this years map becomes obsolete. A vaccine that works pretty well this year may not work in the future – people change, flu bugs change.  So when does science work best?”

Professor Wiggly: “Science works best under two conditions: 1) when the things being observed or measured are stable, and 2) when the size of the  territory or space-time frame being mapped is small (eg  when all you want to know about is this sample of 20 people on this monday morning VS this population of  200 million people over 10 years. Science does best when attempting to map spaces composed of stable components like rocks, because unlike people  rocks don’t change rapidly. So when you observe or measure them today and a week or a month or even a year later you  usually get the same measurement. Whereas unlike rocks, many aspects of people change, although some remain relatively stable. For instance their height and  eye color remains relatively stable so scientific maps describing the distribution of  heights of a given sample or population would be more stable (reliable) that maps of their weight, or political preferences, or  marital status. So science works best mapping the characteristics of  stable components. It also work best mapping the qualities of components in small spaces over a short time. For example scientists obviously have an easier time mapping the marital status of all the people living in Podunk Penn, in October than the marital status of all the people living in the United States over a year – even though we usually want to know about populations rather than small samples.”

Peter: ” Now I get it – it’s obvious. Scientists do better providing trusted maps of a territory when observing stuff that doesn’t change, for instance when studying thing like rocks rather than my girl friend’s mood. Also, they do better studying small samples for a brief time, like my girl friend’s mood at 7.30 am rather than trying to track it over day, a month or a year – forget it.”

Jenny: ” Now I understand why scientists need so much money – in order to study large samples repeatedly so they can accurately map changes in large spaces – like various characteristics (health, income, happiness, etc.) of  the population of the United States over many years. That’s why scientists take so long to create  some maps and why good researchers hedge  their bets by warning us that ‘this map is probably accurate’. Maybe scientists should start stamping a ‘best before date’ on their scientific laws? ‘After that date use this map at your own risk – don’t trust the bridges’.”

Posted in: Sciencing