Closure is sought in scientific experiments in order to assure the ability to make causal assignments.

Experiments change one variable and identify changes in limited number of other variables.

Closure in an experimental situation is gained by one of three ways:

Experimental closure

This is separating people from 'the street' context and putting them in contexts where responses to specific stimuli can be measured - in particular laboratory conditions. It us impossible to do perfectly, because there are always influences that affect people.

Theoretical closure

This uses approaches such as econometric models, where everything is reduce to equations or simplified models.

It is useful for exploration and modelling, but also is imperfect in an open-systems world and hence can be wildly wrong, as many economic predictions have proven.

Statistical closure

Where you cannot explore a complete population, then samples may be taken.

Statistics can be used to indicate the 'statistically significant'. Thus 'p<0.05' means 'less than a 5% chance of being wrong'. As a rule, experiments either go for a 5% or 1% boundary. If an experimenter can get under this hurdle, they can declare their results 'statistically significant.'

See also

Positivism

Original Location: http://changingminds.org/explanations/research/articles/experimental_closure.htm