Page 15 - CAS- Undergraduate-Research-Manual

15

acceptance or rejection of hypothesis. The following are key factors that must be considered in
designing an experiment to help with making more confident conclusions:

Research design

Research design pertains to how data is collected from subjects in research. The Survey Research Design
is popular in social science research and allows for the use of many subjects. Data are commonly
collected via the use of questionnaires. For research conducted in a lab setting, the Experimental
Research Design is often used. It allows cause and effect studied to be studies with the flexibility of
manipulating the causes to produce different kinds of effects in controlled settings. Subjects are
assigned randomly to groups. The faculty mentor would provide additional guidance to students in the
use of specific designs.

Experimental unit

This is usually the lowest level at which observations are made
(e.g., one person, a plant, an animal, a product). These units are
placed in groups of like entities (a plot of plants, a collection of pots
th
in the greenhouse, a set of test tubes, a group of 5 graders). The
members in a group are treated alike in research.

Treatment

A treatment is what the research deliberately applies to
experimental units in research to elicit a measurable response. If
one is studying the effect of temperature on germination,
temperature is the treatment. But in experimental research, the
researcher may choose to study, in this example, different
temperatures (5, 10, 15, 20 o C). The different numbers represent
“levels of treatment.” A study may investigate different grade
levels at one time.

Control group

In comparative studies, one of the experimental units or groups receives no treatment or the standard
treatment. If the researcher is investigating the effect of nitrogen fertilizer on plant height, he may study
three different levels of the chemical (10, 20, 30 mg), which will be applied to three experimental
groups. In addition, an experimental group, identical to the three, will receive no nitrogen, just water
(i.e., 0 mg of nitrogen). This is called the control or control group, with which the others would be
compared to determine the effect of nitrogen fertilizer on plant height. In effect, the researcher is
investigating four treatments (including the control). In medical research, where a drug is being tested,
one group of subjects would receive a placebo (no drug). It is important to note that the control and
other levels of treatments are identical, except, the factor of interest (fertilizer, drug, etc.).

Randomization

Randomization is the process of assigning treatments to experimental units or groups without bias. In
the nitrogen study, each of the four experimental units has an equal chance of being assigned any of the
four treatments. Simply put, the treatments are assigned literally by the toss of a coin, if you will.