Intervention, locations, measurements, participants, variable control, and comparisons are important to consider when analyzing research design features. Quantitative research observes casual relationships such as the association between variables. Experimental study designs contain interventions, control groups, and randomized placements of participants in groups. Quasi-experimental designs also include an intervention but do not randomize in any other aspect. Descriptive designs do not have interventions or any form of distribution treatment, which is why they are usually classified as non-experimental. Typical research questions aim to analyze the differences between two groups or to find an association between certain variables. Randomized trials do not actually use random sampling, as the difference between the groups is technically only made apparent past the intervention. A control group does not receive any form of treatment, essentially being exposed to either a placebo, standard or usual care, lower doses, or an alternative treatment. Therapy and treatment research is best assessed through experimental designs.
In order to assess casualties and efficacy, randomized controlled trials, or RCTs, are the best option within intervention research. An intent-to-treat analysis is implemented for the purpose of eliminating biases due to participant attrition. Clinical trials, such as those used by the FDA prior to the selling of medical products or services, are intervention studies with RCT designs. A common cross-over design includes the two-period, two-treatment design in which participants are assigned to be provided with treatment for period one at random, while the control group will receive it in period two. A factorial design alters a number of independent variables, such as treatments. Nonequivalent control group design is much like the regular experimental design but does not assign participants at random. An interrupted time-series design provides a series of observations in which the dependent variable is measured prior to and after an intervention.