Researchers distinguish two kinds of noise in psychology experiments: "systematic error" vs. "random error."

To make progress in the study of consciousness, it is important to appreciate the nature of random error.

Random error usually cannot undermine the validity of a properly designed experiment.

When designing experiments, one is on the lookout for confounds. These are systematic, consistent differences between the experimental and the control conditions, differences that the experimenter is unaware of. They undermine validity. The only difference between the conditions should be the variable of interest.

Systematic vs. Random Errors

Imagine that caffeine is your variable of interest. You're interested in the effects of caffeine on reaction time to a beep. The experimental condition might have participants drink caffeinated coffee, and the control condition, with a different group of participants, might have them drink decaf. Apart from the presence of caffeine, everything else between the conditions should be the same, including the roast and temperature of the coffee. If, by accident, the temperature of the coffee was always higher in one condition than the other, then this would be a potential confound. Any effect of the temperature would be an unwanted effect; hence, it is called noise or “error.” Because the difference in temperature is constant between the conditions, this kind of error is known as “systematic error.” Thus, confounds are a kind of systematic error.

Then there is noise (or error) that experiments try to minimize but that is not systematic. These are variables that differ across participants but that do not differ systematically between the conditions. With a sufficient sample size per condition, these random effects should “wash out.” For example,........

© Psychology Today