• # Question: how can you tell if the science bit can be wrong or right answer?

• Keywords:
Asked by khanm to Chris, Michael, Paddy, Phil on 22 Jun 2011.

Excellent question, and there are probably different answers from biology, chemistry and physics. In biology, science never claims to be 100% absolutely right. The process of evidence collection and results analysis is based on probability. If you ask a biologist whether the sun will rise tomorrow, he or she’d have to say “probably”. All the evidence we have from history and all the wonderful scientific discoveries we’ve made in the last couple of thousand years mean that we have no reason to believe it won’t, but that doesn’t mean we can say for sure that it will. For more complicated questions, we use a branch of mathematics called statistics. These are a set of rules (based loosely on probability) that tell us whether things are the same or different.

Let’s say you wanted to know whether you find more frogs in ponds north of Leeds or south of Leeds. You would count the frogs in 100 ponds north of Leeds and 100 ponds south of Leeds. How do decide whether or not the average number in the north is different from the average number in the south? Using statistics we can calcuate the probability that if I tell you the number of frogs in a pond (without telling you whether its from the north or south) that you can tell me whether its from the north or the south. If this probability is above a certain threshold (in biology this is usually 95% certainty) then we say that the number between north and south is “significantly” different (i.e. currently accepted scientific techniques say its different). And how far above the threshold of 95% our probability is, tells us about how likely we are to be right. If it’s 95.001%, then we’re more likely to be wrong, than if it is 99.999%.

An example for a reason why we don’t say we’re “right” when we get a significant answer, is that we have to be open to new evidence. The answer could change, for example, if we sample 1000 ponds in each zone, instead of 100. Then we would re-do the analysis and re-evaluate our answer. Most of the time what we have in biology is the best explanation given the data we have available. However, we’d ideally like more data to increase the confidence we have in our answer.

• Chris Jordan answered on 22 Jun 2011:

Paddy’s right. Statistics can be very difficult, but they are the foundation of what makes science. When you write a paper on the results of an experiment or an observation you have to tell people what the answer you got was, and how likely it is to be right – You would say the answer is 42 plus or minus 5 (and that way of writing tells people that the most likely answer is 42 and 95% of the time you got an answer between 37 and 47) A scientist is never 100% sure of anything which makes it very difficult for TV and newspapers and politicians who want scientists to say “this will NEVER happen” or “I’m sure this is correct”. We know that we can’t say that.

In my bit of astronomy we look at for example, all the fast pulsars and try and make some rules from what we measure that tell us something like ….. all fast pulsars are slowing down (that’s mostly true) and we can tell how old they are by how fast they slow down. That turns out to be true for only a group of them not for another group. So now we think we have two sorts of pulsars.
What we do then is to look for more pulsars and see how they fit into the rules that we are making.
If they don’t fit, we have to be open to saying that our rules are wrong, can we think of some other reason for what we see.
If you do science you can’t ignore an experiment that gives an unexpected result. You have to find out why. It could be your experiment was faulty. But it could be that your theory is wrong and you need to be prepared to change your mind!

• Michael Wharmby answered on 22 Jun 2011:

Awesome question!! In short, you can’t tell just by looking at the answer whether it’s right or not, you need to look at the evidence that lead the person reporting it to that conclusion.

Science is based on models and those models are developed from the evidence that we see around us. An example of a model would be predicting the flight of a tennis ball (it is Wimbledon after all!). There are a set of equations that govern the flight path of the ball and these equations have been worked out by thinking about how the object moves and observing it’s movement. These equations are our model. If however a piece of evidence comes along which challenges this model, we need to be open-minded enought to accept this evidence and change our model accordingly. If for example most times the ball flew 1m less than we expected, we’d need to change or add in equations (model). ‘Most times’ is important because we need this to be reproduceable – we apply (very basic) statistics to decide if we do need to change the model and under what circumstances.

A really nice example I’ve already used of models that are flawed and needed changing is the theory of the ancient Greeks about the orbit of the planets around the sun (https://titaniumj11.imascientist.org.uk/2011/06/michael-what-reasch-are-you-do-at-the-presant-and-can-science-be-wrong):
The ancient Greek astronomer Ptolemy was trying to understand the motion of the planets in the sky. The Greeks thought that the Earth was at the centre of the solar system and that everything rotated round it. Now if you watch the motion of the planets in the sky over a few days, you will see that as well as going forwards they ocassionally go backwards – so called retrograde motion. In order to explain this Ptolemy described the motion of the planets as orbiting the Earth, but then having an additional circular orbit on the Earth-centred orbit – an epicycle. Ptolemy’s model fitted all of the evidence he had about the motion of planets in the sky and allowed him to make (complex) predictions of their movements.

In the 17th century the astronomers Copernicus and Galileo began questioning Ptolemy’s system. They proposed a simpler system with no epicycles which put the sun at the centre of the solar system. This model fitted all of Ptolemy’s observations and also allowed predictions of the movements of the planets to be made more easily and with greater accuracy. This model is obviously what we now accept as the correct model.

The initial assumption of the Greeks was wrong – the Earth is not the centre of the solar system. It’s an important point that in science, if our data do not fit with our assumptions we must have another think about those assumptions.