Science is a method to reduce insecurity, by getting to know more about the world (and Universe) around you. Sometimes the knowledge that was gathered in a scientific manner is itself called science.
Science describes the world with theories. In daily life, something is a theory if there is little or no evidence for it. In science, a theory is an explanation for how something works, whether there is little, or a lot, or no evidence at all for it. You can make predictions with a theory. If a theory is good, then you can make good predictions with it.
Suppose you have a theory about the weather. How can you convince even scientists that the theory is correct? That's simple, if the theory is in fact correct. Use the theory to make a large number of weather predictions that are clear and precise, about things that can be measured accurately (such as the temperature and pressure and the amount of rain that falls in a certain area on a certain day). Write down the theory and the predictions and publish them well in advance where many people can read them, in a publication with the date on it. Invite everybody to look for themselves if the predictions come true. If the predictions do not come true, then your theory is apparently not good after all, and then you must reject it. If the predictions do come true, then the theory seems to be correct, and then scientists should become interested in it (if they weren't already).
In this way you cannot cheat. Anyone can see that you made the predictions beforehand and not afterwards. Anyone can see for themselves if the predictions come true, and you have no influence over all of those people. Anyone can use your theory to make new predictions and see if they come true. And if someone claims that he and not you invented that wonderful theory, then you can point at the publication with your name and the date on it that many people had already read before the predictions came true.
In science, essentially the same method is used to determine which theories describe the world the best.
To be more certain about some subject means that you can better predict what will happen to it. The better your predictions are, the more certainty you have.
Measurements of something that you want to be able to predict (such as the maximum outside temperature) always show a certain spread around an average value. Once you have determined what the average value and the spread are, you can predict that future measurements will find values equal to the average, with a possible error that is less than the spread that you measured earlier. That prediction has a good chance of coming true, but is not very useful.
I predict for every day of the coming year that the maximum temperature where you are will be between −200 and +100 degrees Celsius (Centigrade) and that between 0 millimeters and 20 meters (0 and 720 inches) of rain will fall. I have great confidence that 100 percent of my predictions will come true. Yet, this perfect prediction score will not get me a job with the Meteorological Office or the National Weather Service, because my predictions do not exclude anything for the future that has been seen in the past. You cannot figure out from such a prediction whether you'll need your winter coat tomorrow, or your umbrella, or whether you'll need to water your plants a bit more.
A prediction is only useful if it excludes a large share of the natural spread (for the conditions and period for which the prediction is made).
To be able to predict well you must learn which clues are the most important. You can find those only by looking at many examples of the subject (i.e., by taking many measurements) and seeking patterns in the behavior.
If you recognize a pattern in the behavior of the subject, then you can assume that that pattern describes the most important part of your subject. A scientist calls such a pattern a model or theory for the subject.
You can tell how good a theory is by checking how many of its predictions come true, and how many do not.
Usually you can see more than one pattern, and can invent more than one theory, which often each predict a different outcome. You can tell which theory is wrong by checking which theory's predictions do not come true.
Not all predictions are the same. You can find out more quickly whether a theory is wrong if you test a clear and precise prediction than if you test a vague or more permissive prediction, so a clear and precise prediction is more valuable.
To prove that a theory is wrong (so it does not provide a good description of the subject) you need find only a single case for which the theory provides wrong predictions. To prove that a theory is correct you must show that it is correct in all cases. It is easier to prove a theory wrong than it is to prove a theory right. And that a prediction comes true does not necessarily mean that the theory is correct.
A theory cannot cover everything in the Universe to the smallest detail. The things or details that are not in the theory but that are in the Universe can make a prediction fail even though the theory is good (except for the missing details). So, a failing prediction does not necessarily mean that the theory is wrong.
If a theory cannot provide predictions of which it can be checked whether they come true or not, then that theory is not interesting to science. Scientists say that a theory must be falsifiable.
If the theory is clear and precise, then the predictions can be, too. Scientists often describe a theory with mathematical formulas, because those are clear and precise.
A measurement is never exact. A scientist must estimate for each measurement how uncertain tbe result of that measurement was. From this, one can calculate what the chance is that the measurement had an error of a certain size. A scientific measurement is often written as a value plus or minus an uncertainty, for example 6.3 ± 0.2 for a value that is probably between 6.1 and 6.5.
A theory is often based on measured things, which are not precisely known. Predictions made from such a theory are then also not exact.
If you compare a measurement with a prediction from a theory, then there is a chance, because of the uncertainty in the measurement and in the prediction, that the theory is judged wrongly, that it is rejected when it was in fact correct, or accepted when it was in fact wrong.
One can calculate how probable it is that the theory is wrong, given the measurements and their uncertainties. In practice, one often works with a threshold of 5 percent. If the chance that a theory is wrong is more than 5 percent, then the theory is rejected.
If you want more certainty about the theory, then you can just make more predictions and take more measurements. If the theory is wrong, then the chance that you will discover this increases when you take more measurements and compare them with the predictions from the theory. Conversely, your trust in a theory increases with every good prediction that it provides.
A good theory is a summary of knowledge. For example, you can calculate the positions of the planets with great accuracy for thousands of years into the past and the future, with just the right theory and a couple of initial values (the masses, and the positions and velocities at one moment). The theory and the initial values together are a summary of all of those positions together.
In this way, science has provided a network of theories about the different parts of the world. All of those theories must fit together.
You can sometimes find patterns even in a collection of good theories, and from those patterns you can make new theories that summarize even more knowledge. There is, for example, the Law of Conservation of Energy, which states that energy (defined in a certain specific way) cannot just disappear and cannot appear out of nothing. This law (which is what scientists call a theory for which there is so much evidence that none doubt it) plays a part in very many situations, and severely restricts what is possible.
The restrictions that a scientific law imposes are of a different kind than the restrictions that a judicial law imposes. A judicial law says what (according to the law makers) is not allowed, but a scientific law says what is not possible. You can change or break a judicial law, but you cannot change or break a scientific law (if it is correct).
languages: [en] [nl]
Last updated: 2016−02−07