The right approach :
The first part of the question is quite simple as it asks to differentiate. But I am afraid that several students would write the merits and limitations of these designs. Actually, we should distinguish in terms of how these designs work in fetching valid data. The second part, is about the possibility of using quasi experimental research designs for valid the psychological researches
Model answer outline :
A research design is a blueprint of conducting a scientific research. This is why it precedes data collection process. A scientific research, in the ultimate analysis, is all about controlling variance and research design spells out clearly how the required things will be controlled. Even in descriptive researches like naturalistic observation, we don’t allow several factors to intrude the natural setting.
Experimental research designs and quasi experimental research designs differ in three important ways :
1) scope of control : In experimental research designs, maximum controls are exercised. Be it internal ( like controlling other possible causes except the IV or the hypothesized cause in a linear experimental design) or extraneous like temperature, noise, etc. Everything is brought under control.
Individual differences are controlled through “within or between group sampling” designs. Randomization is an important technique which is possible in experimental research but is not possible in quasi or expost facto research design.
2. Scope of the Manipulation of the hypothesized cause… (IV)
In an experimental research design, we actively and repeatedly manipulate the hypothesized cause of the phenomenon to see its effect on the dependent variable. The scope of active manipulation is not available in quasi experimental research designs. Rather to manipulate, we select the people who have been, naturally exposed to the the IV. This is why it is called nature’s experiment.
3. Scope of direct observation of the effect
We can see how reinforcements influence performance in an experimental design but cannot observe the impact directly in a quasi experimental research design because we move from the effect ( that has already taken place) to the probable cause. In quasi experimental research designs, rather to observe, we calculate the strength of correlation between IV and DV. The correlation serves the purpose of observation.
Ex post facto research designs are indeed more frequent in the social science as human subjects /participants can not be controlled, manipulated and observed as we do with chemicals in chemistry, for instance, due to practical and ethical reasons.
But moving from effect to cause, indeed, is a more difficult process. For example, Diana Baumrind, correlated the different behavioural characteristics of school children like social, adaptive, aggressive to the parenting styles they had experienced. On the basis of strong correlation with authoritative parenting style, she inferred that such parenting model makes children socially adaptive.
But such inferences might have confounding effect, might undermine individual differences, and more limitations of this kind.
Moreover, Internal validity of such findings can not be as good as we have in the experimental design. Statistical scrutiny of data, however, reveals that strong correlation is always found in a cause – effect relationship. And it can be inferred from the both sides( from effect to cause too).