Experimental and longitudinal designs in research methodology

Experimental design employs comparison as its strategy for the given research. It uses two groups, which the researcher uses for comparison purposes. These include the experimental group and a control group. The two groups used in a study have similarities, but the experimental group uses the independent variable, whereas the researcher the control group is not assignment of subjects to either control or experimental group because it is central to chance. Nevertheless, the researcher assigns cases to the two groups randomly. In order to determine the influence of the independent variable, investigators will measure the dependent variable, designated as scores, two times from both groups (Frankfort-Nachmias and Nachmias, 2008).

In addition, researchers take a single measurement, the pretest, for all cases before introducing the independent variable in the experimental group. Moreover, they also take a second measurement, the posttest, for both cases after exposing the experimental group to the independent variable. Afterwards, the investigators determine the difference in the measurements obtained from the pretest and posttest; if the variation in the experimental group is larger when compared to the control group, the hypothesis developed is that there is a casual relation between the dependent variable to the independent variable (Walker, 2005).

Strengths

In the experimental design, it is possible to observe the effect of the independent variable on the dependent variable within a short time. In addition, research suggests that it is the strongest research design because it is possible to manipulate independent variables and offer substantial control over the intrinsic and extrinsic factors. Most importantly, the experiments will enable the investigators to control the independent variable to help them establish the potential pathway of causation.

Weaknesses

Although research suggests that it is the strongest design, it does not fall short of weaknesses. In experimental design, the external validity is weak because the design does not allow for replication in reality. Nevertheless, investigators must always rely on volunteer for their potential samples. This suggests that the sample may not be an investigator's choice, which further prevents them from generalization limiting the scope of their study.

Longitudinal/Observational design

Longitudinal research designs involve the collection of data and subsequent evaluation over time. Most importantly, this design is applicable when the objective of a given study is to measure social change. This is because the design allows for a diachronic evaluation of the incidence of conditions and cases. In addition, there are several types of data under this design, which include repeated cross sectional studies, prospective studies and retrospective studies. This term is broad suggesting that the techniques for evaluation of social change must vary in a broad manner. Nevertheless, this design has the capacity to provide comprehensive information in the context of individual behavior, but the use of the above data types pose substantial risk to the theory and methodological approach.

Strengths

Longitudinal designs allow for evaluation of data in the context of time. In addition, this design allows for measurement of variations or change in variable from one period to another, this means that it provides for the description of the various patterns of change over a given period, and can assist in locating the causes of the social issue. In so doing, this design can help in shaping or determining connections between phenomenal that are broadly separated in different times. Nevertheless, this design provides room for richer information concerning a given social phenomenal.

Weaknesses

Although longitudinal research design has the capacity to provide comprehensive data regarding individual behavior, the design poses theoretical and methodological challenges. In addition, longitudinal studies are expensive and have the capacity to consume much time. On the other hand, repeated cross-section designs are inappropriate in establishing the developmental patterns within given cohorts. In addition, retrospective inquiries regarding motivational or cognitive are problematic because participants find it hard to remember the timing of changes in these situations. Notably, retrospective studies rely on survivors suggesting that dead subjects will not be included, which can amount to bias.

Recommended Quantitative Research Design

In a quantitative research, the objective is to establish the relationship between an independent variable, and a dependent variable, within a given population. In addition, the widely employed designs in this approach are either descriptive, or experimental. In a descriptive study, it aims to determine the association between variables, whereas the experimental study aims to determine causality. In addition, a descriptive study will require a very large sample of over hundreds, or thousands, whereas the experimental one cans offer reliable results with only ten subjects (Walker, 2005). Research suggests that using a large sample is likely to generate case of bias; therefore, experimental design strands as the most appropriate method to employ in a research design. In addition, the experimental design employs randomization, an approach, which reduces the incidences of bias primarily because the subjects are assigned treatments randomly.

Pre-experimental Designs

These designs are not appropriate for experimental manipulations because they do not provide for randomization of cases to the experimental and control group. Nevertheless, most of the designs do not utilize a comparison group. Owing to prior research, the design represents the weakest research designs, primarily because it is not possible to control the sources of internal and external validity. On the other hand, the incidence of drawing inferences from these designs is very high, and they apply when pre-testing some hypothesis and for expletory studies. This suggests that it is applicable in exploratory studies, but research suggests it may lead to insights that could qualify for study as research hypothesis (Frankfort-Nachmias and Nachmias, 2008).

Quasi-experimental Designs

Quasi-experimental designs are weaker when it comes to internal validity when compared to experimental designs, suggesting that investigators will have to depend on data evaluation methods as their primary method of control (Frankfort-Nachmias and Nachmias, 2008). Nevertheless, although qausi-experimental designs allow investigators to choose samples randomly, they do not allow for random assignment of individual cases to the comparable groups. However, they do allow or involve the study of more than one sample over a given period. In addition, the designs allow for direct determination of comparative analysis such as mean, mode and median, but in some designs, it is impossible to determine the differences between the dependent and independent variable. This further complicates the move towards establishing a hypothesis (Frankfort-Nachmias and Nachmias, 2008).

Cross-sectional design

This design has associations to survey research, a technique of collecting information, where the investigators require a random sample to answer some questions concerning their backgrounds, experiences, history and attitudes. Although investigators who use this design aim at determining causal relations, many of them use it to establish a pattern of relation amid variables. A substantial limitation is that this design does not provide for manipulation of the independent variable, to assist in making before and after comparisons. In comparison with the experimental design, this one is not amendable. Therefore, it is difficult to integrate the research design the components of manipulation and control, which will help an investigator to determine causality. Although it is possible to improve the design in order to achieve comparisons, research states that it is not possible to conclude that two variable have a relationship, which makes it further complicated when developing a hypothesis (Frankfort-Nachmias and Nachmias, 2008).