How is it possible to get acurate results in psychological research?

Question

Psychology is often thought of as a "soft science." It seems that it is very difficult to control for all of the variables which exist within any given primary reserach scenario.  How can we trust the validity of this type of research?

Answer 1

We express our research with confidence intervals.  For example we know not everything is controllable 100% of the time without fail so we might stay something to the effect of "We are 95% confident that this person's real IQ is somewhere beween 85 and 93". When they might have tested at an 89 IQ on that given day.  We can lower our confidence (alpha level) to broaden our inclusive criteria but it lowers our confidence in our results.

Answer 2

The most accurate research in psychology is perform using individual sudies based on statistic methods and the use of the nul hypothesis.  The null hypothesis is actually a hypothesis posing the opposite of what you think is happening.  For example, if we think that only children above seven years of age can solve a Piagetian conservation task and children younger cannot, the null hypothesis would state that the proportion of children younger than seven that solve the problems and the number of children over seven years of age will be the same and there will be no difference.  If we then test both groups of children and compared the results statistically and we find that the probability of getting that difference was less than 0.05, we reject the null hypothesis and accept that children above seven years of age develop a cognitive change that allows them to solve the problem.

Rejecting the null hypothesis does not prove the alternative hypothesis.  The alternative hypothesis is acceptaed until other studies refine the explanation further.  This scientific process brings you progressively closer to the true explanation, it is hoped.

Not all psychological research topics allow for statistical ananlysis and analysis by null hypothesis.  Less rigorous methods might be used to help define what topics should be explored in a more scientific manner.