A scientific control is an experiment or observation designed to minimize the effects of variables other than the independent variable. This increases the reliability of the results, often through a comparison between control measurements and the other measurements. Scientific controls are a part of the scientific method.
In controlled experiments, the same experiment is done in at least two parallel experiments that different in only one way, with one experiment being the "control arm" and the other being the "experimental arm".
Controls eliminate alternate explanations of experimental results, especially experimental errors and experimenter bias. Many controls are specific to the type of experiment being performed, as in the molecular markers used in SDS-PAGE experiments, and may simply have the purpose of ensuring that the equipment is working properly. The selection and use of proper controls to ensure that experimental results are valid (for example, absence of confounding variables) can be very difficult. Control measurements may also be used for other purposes: for example, a measurement of a microphone's background noise in the absence of a signal allows the noise to be subtracted from later measurements of the signal, thus producing a processed signal of higher quality.
For example, if a researcher feeds an experimental artificial sweetener to sixty laboratory rats and observes that ten of them subsequently become sick, the underlying cause could be the sweetener itself or something unrelated. Other variables, which may not be readily obvious, may interfere with the experimental design. For instance, the artificial sweetener might be mixed with a dilutant and it might be the dilutant which causes the effect. To control for the effect of the dilutant, the same test is run twice; once with the artificial sweetener in the dilutant, and another done exactly the same way, but using the dilutant alone. Now the experiment is controlled for the dilutant and the experimenter can distinguish between sweetener, dilutant and non-treatment. Controls are most often necessary where a confounding factor cannot easily be separated from the primary treatments. For example, it may be necessary to use a tractor to spread fertilizer where there is no other practicable way to spread fertilizer. The simplest solution is to have a treatment where a tractor is driven over plots without spreading fertilizer and in that way the effects of tractor traffic are controlled.
The simplest types of control are negative and positive controls, and both are found in many different types of experiments. These two controls, when both are successful, are usually sufficient to eliminate most potential confounding variables: it means that the experiment produces a negative result when a negative result is expected, and a positive result when a positive result is expected.
Where there are only two possible outcomes, e.g. positive or negative, if the treatment group and the negative control both produce a negative result, it can be inferred that the treatment had no effect. If the treatment group and the negative control both produce a positive result, it can be inferred that a confounding variable is involved in the phenomenon under study, and the positive results are not solely due to the treatment.
In other examples, outcomes might be measured as lengths, times, percentages, and so forth. In the drug testing example, we could measure the percentage of patients cured. In this case, the treatment is inferred to have no effect when the treatment group and the negative control produce the same results. Some improvement is expected in the placebo group due to the placebo effect, and this result sets the baseline which the treatment must improve upon. Even if the treatment group shows improvement, it needs to be compared to the placebo group. If the groups show the same effect, then the treatment was not responsible for the improvement (because the same number of patients were cured in the absence of the treatment). The treatment is only effective if the treatment group shows more improvement than the placebo group.
Positive controls are often used to assess test validity. For example, to assess a new test's ability to detect a disease (its sensitivity), then we can compare it against a different test that is already known to work. The well-established test is the positive control, since we already know that the answer to the question (whether the test works) is yes.
Similarly, in an enzyme assay to measure the amount of an enzyme in a set of extracts, a positive control would be an assay containing a known quantity of the purified enzyme (while a negative control would contain no enzyme). The positive control should give a large amount of enzyme activity, while the negative control should give very low to no activity.
If the positive control does not produce the expected result, there may be something wrong with the experimental procedure, and the experiment is repeated. For difficult or complicated experiments, the result from the positive control can also help in comparison to previous experimental results. For example, if the well-established disease test was determined to have the same effectiveness as found by previous experimenters, this indicates that the experiment is being performed in the same way that the previous experimenters did.
When possible, multiple positive controls may be used -- if there is more than one disease test that is known to be effective, more than one might be tested. Multiple positive controls also allow finer comparisons of the results (calibration, or standardization) if the expected results from the positive controls have different sizes. For example, in the enzyme assay discussed above, a standard curve may be produced by making many different samples with different quantities of the enzyme.
In randomization, the groups that receive different experimental treatments are determined randomly. While this does not ensure that there are no differences between the groups, it ensures that the differences are distributed equally, thus correcting for systematic errors.
For example, in experiments where crop yield is affected (e.g. soil fertility), the experiment can be controlled by assigning the treatments to randomly selected plots of land. This mitigates the effect of variations in soil composition on the yield.
In blind experiments, at least some information is withheld from participants in the experiments (but not the experimenter). For example, to evaluate the success of a medical treatment, an outside expert might be asked to examine blood samples from each of the patients without knowing which patients received the treatment and which did not. If the expert's conclusions as to which samples represent the best outcome correlates with the patients who received the treatment, this allows the experimenter to have much higher confidence that the treatment is effective.
In double-blind experiments, at least some participants and some experimenters do not possess full information while the experiment is being carried out. Double-blind experiments are most often used in clinical trials of medical treatments, to verify that the supposed effects of the treatment are produced only by the treatment itself. Trials are typically randomized and double-blinded, with two (statistically) identical groups of patients being compared. The treatment group receives the treatment, and the control group receives a placebo. The placebo is the "first" blind, and controls for the patient expectations that come with taking a pill, which can have an effect on patient outcomes. The "second" blind, of the experimenters, controls for the effects on patient expectations due to unintentional differences in the experimenters' behavior. Since the experimenters do not know which patients are in which group, they cannot unconsciously influence the patients. After the experiment is over, they then "unblind" themselves and analyse the results.
In clinical trials involving a surgical procedure, a sham operated group is used to ensure that the data reflect the effects of the experiment itself, and are not a consequence of the surgery. In this case, double blinding is achieved by ensuring that the patient does not know whether their surgery was real or sham, and that the experimenters who evaluate patient outcomes are different from the surgeons and do not know which patients are in which group.