scientific-method.md

Scientific Method

Science consists of making falsifiable predictions.

Broadly, the scientific method consists of a prior belief, a conjecture (hypothesis), and an experiment. Specifically:

1. Problem statement or observation. Explore. Do research.
2. Hypothesis. A falsifiable prediction. Use induction to predict a hypothetical situation. Make a generalization based on an observation. These can range from analogies, causal inferences, syllogisms.
3. Experiment. Design an experiment to test the hypothesis. Use deduction to ensure that conclusions are valid; both in terms of syntax and semantics. This may be done with either a formal or a natural language. A inherent challenge is to maximize the likelihood P(E | H).
4. Analyze. Observe the results and compare them to the predicted results. P(H | E)
5. Conclude. Draw conclusions without over-generalizing.
6. Report findings and share them freely. Facilitate other researchers to replicate the experiments.

There are parallels with Bayesian inference.

• The problem statement is the prior P(H): the probability that the hypothesis H is true.
• The hypothesis and the experiment are the likelihood P(X | H): the probability of observing the predicted result X, if the hypothesis would be true.
• The conclusion is the posterior P(H | X): the probability that the hypothesis is true, given the observation.

Challenges

• In practice, it my be impossible to proof that an hypothesis is true. Instead, a combination of experiments can be setup to disprove or reject any alternative hypothesis.
• The number of possible hypotheses is infinite. Science does not prescribe how to choose or select the right hypothesis. There is an art to thinking outside the box and coming up with novel hypotheses.