Experimental Design

ExperimentalDesign

Experimentaldesigns provide the best or rather most accurate findings whencompared to other research designs. They are made up at least twogroups i.e. a control/comparison group and a treatment/experimentalgroup. The experimental group entails participants whom theintervention is made. The results from this group are compared to thecontrol group. The control group on the other hand entails,participants who are untreated i.e. not intervened, but have similarto the experimental groups in all other aspects. Experimental designscan also be categorized depending on the mechanisms used i.e.randomized experiments [ CITATION Ken10 l 1033 ].

Inthat respect, they are quite versatile allowing the researcher toevaluate the program’s effect in a number of levels as well as thegeneral impact. Another advantage of this designs entails validity.The internal validity is quite high. This is because similar outcomesare attained whenever the program is assessed. It also requires fewerassumptions since the external influences tend to affect all groupsequally. They all receive the same intensity. Experimental designsare also relatively simple to assess as compared to other researchdesigns [ CITATION Gro09 l 1033 ].

Experimentaldesigns tend to pose ethical issues depending on the kind ofintervention. For instance, a program to supplement vitamin C tochildren so as to improve children’s health. It is unethical towithhold the vitamins from the control group. Experimental designsare also time consuming. It may take a long time to evaluate clinicaltreatment of certain groups [ CITATION Gro09 l 1033 ].

Inmatched pair design, I would group participants into pairs inaccordance to some blocking variable. It generally involvescomparing two treatments by choosing pairs of the participants thatare closely matched. Then assigning random treatment to oneparticipant and applying another treatment to the other participant.Or even applying same treatment on a pair. Then randomizing order oftreatment [ CITATION Kos07 l 1033 ].

Pair

Treatment

Placebo

Vaccine

1

1

1

2

1

1

3

1

1

………

…………

………..

499

1

1

500

1

1

Theabove table illustrates a matched pair kind for Acme experiments.1000 participants can be divided into 500 matched pairs. Then eachpair is categorized on age and gender. For instance, pair 1 may havetwo women both aged 21. Pair 2 having two women aged 22 etc. [ CITATION Pur10 l 1033 ].

Onemajor advantage of this design is that it clearly controls twoprobable lurking variables i.e. gender and age. This design does nothave issues resulting from carry-over effects hence counter-balancingis not necessary.

However,the design requires lots of effort to match the subjects.Additionally, matching requirements might disregard potentialsubjects for not obtaining sufficient suitable matches.

References

Bordens, K. S. &amp Abbott, B.B. (2014). Research Design and Methods:A Process Approach. New York: McGraw-Hill.

Grosskurth H, M. F. (2009). Experimental Study Designs. Experimental Study Designs.

Imai, K. (2007). Use of matching methods for casual inference in experimental and observational studies. Use of matching methods for casual inference in experimental and observational studies, 1-11.

Purdue Unversity Fort Wayne. (2010). Using Between-subjects and within-subjects experiments designs. Using Between-subjects and within-subjects experiments designs, 1-22.