Treatment Structures in Designing Experiments

1. Placebo

Placebo means something to please. It is not rare that some patients experience relief from a sickness on using a medicine given by a doctor, even if the medicine is nothing more than an inert substance given to the patient, without his knowledge, in the form (shape and color) of a medicinal unit. I remember a doctor who practiced in a rural area. Some of her patients walked miles to come to her with complaints of various sicknesses: stom­achache, giddiness, weakness, headache, and so forth. They insisted on having an injection instead of tablets, even though they knew the doctor would charge them less if they accepted tablets. The doctor had tubes of distilled water waiting to be used for such patients, who paid her gladly after receiving the injec­tion. In either case, tablets or injection, patients were often given a placebo.

Imagine that a drug company comes up with a proprietary product to cure one such sickness, and it falls to the lot of an experimenter to justify or repudiate the claim of the company. If two patients, otherwise comparable, are experimented on, one is given the proprietary product, and the other is given a fake, a pla­cebo, that looks similar. If both of the patients are cured equally well, the proprietary product’s worth is called into question. If, instead, the one who used the proprietary product is cured and the other one is not, the product is worthy to that extent.

2. Standard Treatment

Suppose two plants of a kind, which seem equal in every way, are subjected to test. Let us say that one of these is given a plant food. To the extent that this plant gives a better yield, the plant food is considered beneficial.

3. “Subject-and-Control” Group Treatment

Instead of only two plants as mentioned in the previous section, two groups of plants, otherwise comparable, are subjected to test for the benefit of plant food. One group is subject to treatment, and the other is not. Yields of the two groups, analyzed on a sta-tistical basis, are then compared; to the extent that the group that received the treatment gives a better yield, the plant food is rated as beneficial.

4. Paired Comparison Treatment

A particular plant food is, let us say, claimed to benefit apple, pear, and orange trees in terms of yield. One member of each pair of comparable apple, pear, and orange trees is given the plant food, while the other member of each pair is denied it; both members of each pair are otherwise provided similar care. The yield for each kind of plant is recorded separately. This data pro­vides the basis to justify or repudiate the claim that the plant food can benefit these kinds of plants and further to quantify the benefit by statistical methods.

5. Varying the Amount of One of the Two Factors

In the simple case of one cause, one effect, if the cause is varied over a reasonable range, and the corresponding effects are recorded, we have the classic x-y relation that can be plotted as a graph. If, instead, two factors are known to be influential, and one is kept constant while the other is varied, we get a “surface” that can be plotted as a graph. Here, if X1 and X2 are the factors (causes), with X1 remaining constant and X2 varying, and y is the result (effect), we obtain a surface of width X1 and shaped accord­ing the X2-y relation.

Source: Srinagesh K (2005), The Principles of Experimental Research, Butterworth-Heinemann; 1st edition.