Replicability is a high valued aspect of any research. A research that is not replicable is useless and if results of a research are not replicable it makes the results redundant.
Suppose you have a billiards table with one white ball on it. You grab a stick and hit the ball with a certain speed at a certain place. The ball will start rolling and will end up somewhere else on the billiards table. You mark this spot. Then you replace the ball on its original space and hit the ball with exactly the same speed on exactly the same place. Again you mark the place on the billiards table where the ball stopped rolling. You can repeat this over and over again. In theory – and yes, this is theory – the ball always ends up in the same place.
This research can be repeated all over the world. Maybe due to differences in gravity and or humidity, minor changes in the results will occur. So, this research is fairly well replicable. Now you can pronounce a theory that explains what is happening. This theory is pretty useful because it is independent of time, place (on earth) and the experimenter (person).
Now look at this experiment. Suppose you have an empty room with just you and another person. You ask the person to stand still with his back to you. You grab a stick and hit the person firmly between his shoulders. You watch the reaction and write it down in your notebook. Then you ask the person to stand still again on the same place and you hit him again firmly between the shoulders. Well, if you are lucky you get the change to repeat this experiment a second time, but the reaction will be quite different (I suppose). This research is not replicable over and over again.
Although both experiments are essentially the same (one experimenter, one stick, one empty but limited space and one object), the results are quite different. The results of experiments in physics and chemistry are more replicable than in the human or social sciences. That is because the object of study is quite different, a person is not the same as a ball. This does not mean that the knowledge gained from research in the social sciences and humanities is useless because human behaviour cannot be predicted. Yet it does provide useful knowledge. To develop a well-fitting theory of human behaviour in certain circumstances, a great deal of experimentation must be done. And whether the theory will hold in any situation must always be investigated.
To qualify a research it should give warrants about the methodological (= internal) validity. This is done by describing fully how the research was performed on four aspects:
- What was the design of the research?
- Who (or what) participated in the research?
- What was measured and how was that done?
- Which analyses were conducted?
A second question that has to be answered is if the results will hold in another context. This is called the question about external validity. All these aspect are described elsewhere in this dictionary.