Science and Engineering in Derby and Derbyshire
(Picture-“5” by Svenwerk )
Scientific research is a diverse world. There are people trying to solve climate change, people working on an alphabet of diseases, and people working on technology that most of us can barely imagine. As different as all those disciplines are though, the research will all have some things in common.
Generally when people say ‘science’ they are referring to those three subjects which are bunched together as ‘science’ in most schools: Biology, Chemistry and Physics. Subject wise this isn’t all that far from the world of research, as most areas are a subset of one or two of those three areas. So for example from a single subject Zoology and Botany are types of Biology, whilst Engineering and Astrophysics are subsets of Physics. Whilst Forensic Science is a mix of Chemistry and Biology (and oftentimes Physics too), and Geology is a mix of Chemistry and Physics. Being related to the ‘science’ subjects isn’t the whole story though.
The Science Council defines science as:
“the pursuit of knowledge and understanding of the natural and social world following a systematic methodology based on evidence.”
In other words, how you do it is just as important as what you do.
There are some core components to any kind of science from Fluid Dynamic Engineering to Occupational Psychology…
If you look in a dictionary will likely find two definitions of hypothesis. The definition that scientific research is interested in is:
“a tentative assumption made in order to draw out and test its logical or empirical consequences”.
This is essentially a prediction of what you think is going to happen. Research will usually work with two hypotheses. The first hypothesis is called the Null Hypothesis. This is a slightly strange prediction, because it is forecasting that nothing will happen, or that if anything does happen it was just down to chance. This is because scientists are cautious people. They’ll get excited when they’ve actually opened the birthday present, and not before. So research is planned from the assumption that nothing is going to happen.
The second hypothesis, is called the Alternate Hypothesis. This is the one that all those cautious researchers secretly hope will win out. This hypothesis predicts that something will happen, and that it isn’t because of chance. The hypotheses are important to give a framework for what the researchers are investigating. It keeps them on track, and stops them being distracted by new shiny things that they find along the way.
Measurability, Reliability and Repeatability
It’s all very well and good coming up with a prediction, but you need to be able to test this prediction. For something to be testable, there needs to be something observable and measurable. Observation doesn’t just mean something you see. Nor does it have to be something only human eyes, ears and noses can observe. In fact, since we humans are a pretty unreliable bunch, most modern research makes use of computing power to observe, and computers can see very tiny things happening.
For results of research to be taken seriously that testing needs to be done in a rigourous and repeatable manner. They have to write up their experiment in a way that means other people can repeat the experiment to check the original results. It is important that experiments are both accurate and precise. In everyday language we often use these terms interchangably, but in scientific research they have very specific meanings.
Accuracy is about measurement and how ‘true’ to reality something is, whilst precision is about reliability. So for example, an accurate singer can aim to hit a C5 note and acheive it, whilst an innacurrate singer aims for C5 and could hit, well frankly anything if The X Factor is anything to go by. A precise singer will be consistent, they could be consistently bad, but they’re consistent. Whilst their inconsistent colleague may sing lovely clear notes, but never the same one twice in a row.
Once the experiment has been carried out the next stage is to establish if the results actually mean something. This is the part that makes people who don’t much like maths start to cry a little bit. The thing that the statistics are designed to show is whether or not something happens by chance. Much of statistics makes use of something called the Normal Distribution, also known as a bell curve. For example if you measured the heights of all the people in your street you would probably find that though there would be some very short people, and some very tall people, most people were somewhere in the middle. If you measure enough people to make a graph of it you would likely find that it forms the shape of a bell, a big lump in the middle with two tails either side. The tail parts of that bell are the bit that are often the most interesting. Researchers want to know if results that don’t land in the big middle bit, do so just because of chance, or because there’s something special about them.
Control Groups, Blinding and Placebos
One of the problems of research is that you need to know whether your intervention is making the difference to your sample, or whether what happens would have done so anyway. The control group is a second sample that does not recieve the intervention that you are testing. So if you wanted to test whether listening to a podcast about algebra whilst you sleep, helps people perform better on a maths test, you would need another group of people who don’t doze to differentials.
However, you might find that your two groups talk to each other, especially if they come from the same population such as a school. Researchers can get people to promise not to reveal things, but humans are humans, and if group two discover they are not expected to perform as well, they may not concentrate on that test as much. So to prevent this kind of problem placebos and blinding are used.
A placebo is something used in place of the effective intervention, so that your control group are also recieving ‘something’. Giving the placebo group something though is not enough to ensure that they are responding in as natural way as possible, which is where blinding comes in.
It sounds like some horribly violent activity, but fear not, there are no pokers in eyes. Blinding simply means not letting your partcipants know whether they are recieving the intervention that you want to know about. So researchers would likely give both sets of students a podcast to listen to, but you wouldn’t tell them that you expected those listening to the algebra show to perform better. Much research, especially medical research, will add another layer to blinding called double blinding. This means that though the research team back in the laboratory know that participant 657 is taking the placebo medicine, 657 and the doctor of 657, who is distributing the pills to her patients, don’t know. This is so that the people who interact with the participants do not affect the outcome by dealing with them differently.
Newspaper headlines love to say that scientists have proven something: “Boffins prove crisps make you fat!”, “Computer games proven to make you a homicidal maniac!”
Scientific research is very rarely about proof, it is about supporting theories. Most researchers would not be daft enough to say that they had proven anything, partly because it sounds terribly show-offy, but mostly because they know that someone will come along in six months time and disprove. The cautious nature of science is to say that there are not absolute proofs, but that some theories have more evidence than others. In the case of many things, usually the sort of things that people are taught at school, all of the evidence so far supports that theory. At some point though, as techniques get better, there might be evidence that doesn’t support those theories.
The important thing is that science is about following a process, a system that tries to ensure that any results are the best they possibly can be.