THE CONTINUUM OF RESEARCH DESIGNS
One way to conceptualize the differences between research types is on a continuum of control. Control refers to the extent to which factors responsible for a phenomenon are isolated. If you simply go out and measure behavior in the field, you don't really know what causes people to behave the way they do. For example, a researcher who is interested in studying public displays of affection might go to a mall and record the behavior of couples. But how do you know why couples are acting the way they are? Maybe they just came out of a romantic movie and are more likely to hold hands. Maybe one of them spilled a Slurpee, getting the other mad. Maybe they only acted affectionately when the researcher saw them. Maybe...maybe...maybe. The point here is that there is very little control over the situation in these kinds of settings. Let's take an example on the other end of the control continuum, consider a researcher who is interested in one's reaction time to emotional vs nonemotional words. The researcher places each individual subject in a small room with no decorations or windows with a computer on a desk. Red and green emotional or nonemotional words are presented individually on the computer screen and you press buttons to indicate whether the word is red or green. Clearly, this study has more control over the measure of reaction time than a researcher has over affectionate behavior in the mall.
The Continuum of Control
We can conceptualize types of research as a continuum from low control to high. Observational studies would be at the 'low' end of this continuum while experimental would be closer to the 'high' end. To be more specific, naturalistic observation would be at the lowest point of control while highly controlled laboratory experiments like the one described above would be at the highest end.
Correlational vs. Experimental Research
As you can see, correlational designs tend to be near the 'low control' end of the spectrum while experimental designs tend to have higher control. This is simply a result of the nature of experiments. Experiments must have at least one independent variable (i.e., something being manipulated) and at least one dependent variable (i.e., something measured). An experiment requires manipulation of an independent variable, thus there must be some control of the situation. It is NOT the case that all experiments have more control than correlational designs. Some correlational designs have a great deal of control while some experiments have very little. Because of this, perhaps a better way to conceptualize this continuum is through the concepts of external and internal validity.
The Continuum of Validity
There are a variety of types of validity. Validity refers to the extent to which something measures what the researcher is trying to measure. For example, something with low validity would be using jelly bean preferences as a measure of intellect. Presumably, jelly bean choice doesn't tell much about one's intelligence; thus, this measure would have low validity.
External Validity and Internal Validity
External validity is the extent to which the results of a research study can be generalized to different situations, different groups of people, different settings, different conditions, etc. Internal validity is basically the extent to which a study is free from flaws and that any differences in a measurement are due to an independent variable and nothing else. This relates to the idea of control: the more control one has over a situation, the higher the internal validity. That is, if you exercise high control over a situation, you have a pretty good idea that any differences in your dependent variable are due to your independent variable and nothing else. The less control you have, the more likely other factors influenced your dependent variable, thereby decreasing internal validity. Thus, the 'control continuum' above can be translated into a 'validity continuum.'
The Continuum of Research Types
Most research is conducted in the laboratory or in "the field." Field research is any research that occurs outside of a scientific laboratory. There is a clear relationship between the amount of control, the degree of internal validity, and laboratory research. Thus, the more one gets away from a laboratory setting and toward a field setting, the less the control, the less the internal validity, and the greater the external validity. We can therefore conceptualize this continuum in terms of location: