Quantitative Research Understanding Power and Effect Size
Power and Effect Size
A sample is the part of the population used for a study (Sink & Stroh, 2006). The choice of the sample is a critical step in research because all the findings are based on the selected sample. In the determination of statistical findings, hypotheses that are mutually exclusive are set and used to formulate a data analysis plan (Sink & Stroh, 2006). The hypothesis can be either a null or the alternative hypothesis. The effect size, which is an important determinant of statistical significance, is the difference between the true value and stated value (Sink & Stroh, 2006).
There is no specification of how the effect size should be. Statistical significance depends on the utility, impact, cost and benefits of the results (Sink & Stroh, 2006). The differences in effect size can result in important differences in outcome. For instance, in a study done by Segool and Carlson, a medium effect size of cognitive behavioral treatment (0.86) resulted in the large decrease in the social anxiety symptoms (Segool & Carlson, 2008).
Statistical power, on the other hand, is the probability of detecting a result when it is present (Sink & Stroh, 2006). Many factors affect statistical power, for example, sample size, the significance level, and the effect size. The power of the test is determined by how well a researcher rejects the null hypothesis that is false (Sink & Stroh, 2006). Therefore, in the study conducted by Segool and Carlson, since the p-value is less than 0.0001, it can be concluded that there is statistical significance between Selective Serotonin Reuptake Inhibitors (SSRI) treatment and a decrease in social anxiety symptoms (Segool & Carlson, 2008).
Quantitative Research Design
All research is based upon the findings, and the research methods employed to develop the knowledge for a given study. Study designs can involve qualitative, quantitative or mixed methods. However, one common feature across all research designs is data collection at one point of the research process (Crosswell, 2011). Qualitative research explores meaning, purpose, or realities that are capable of transforming the world. Quantitative methods, on the other hand, attempt to maximize objectivity, and generalize findings that are developed by predictions (Crosswell, 2011).
Surveys and experiments are examples of quantitative study designs. The design of the survey has a standard format. In it, the researcher comes up with conclusions about a population based on the sample results. Experiments also involve generalizations. However, in experiments the researcher tests the impact of a specific intervention on the outcome (Crosswell, 2011). All the other factors that might interfere with the final result are kept constant in order to develop a relationship with a specific intervention. One common use of experimental designs is testing the effectiveness of medications on individuals.
When choosing the sample, the researcher starts by identifying a suitable population. After determination of the population, the researcher should identify whether the sampling design is single or multistage (Crosswell, 2011). The selection process for individuals should then be identified. Selection can be either probabilistic or non-probabilistic. In probabilistic sampling, each person in the population has some chance of being selected. Simple random sampling, systematic random sampling, stratified sampling, and cluster sampling are the main types of probability sampling (Crosswell, 2011). In non-probability sampling, there is no equality in selection because the choice is dependent on the researcher, and some of the methods used include snowball sampling and quota sampling (Crosswell, 2011).
Surveys and experiments have specific similarities and differences. Both methods are examples quantitative study designs. In addition, sampling has to be done in both methods because of the impossibility of using the entire population. However, in an experiment, there are modifications that are made on subjects, and the use of a control experiment (Harwell, 2010). Surveys, on the other hand, do not make the modification of the respondents once they are selected for study. One advantage of an experiment is that a comparison can easily be done before coming up with conclusions (Segool & Carlson, 2008). However, the results in an experiment are subject to bias because the researcher can modify the participants to suit their desires. Surveys are advantages because most of them apply random sampling techniques that offer equal chances of selection. However, the sample size is limited, and the selected sample sometimes does not give the correct generalization for a population.
The task of selection of a study design and the participants is not entirely on the hands of the researcher. Eligible participants who have been chosen for any study should not be forced into study (Crosswell, 2011). Instead, they should sign a consent form to prove their willingness to take part in the study. The information collected during collected during any study should also be kept confidential. Some information collected, especially in experiments should not identify the participants. They should only be revealed in case the respondent agrees, or else the researcher can be sued because of a breach of contract.
Crosswell, J. W. (2011). Qualitative, Quantitative and Mixed Method Approaches. Lincoln: Sage Publications.
Harwell, M. R. (2010). Reseach Designs in Qualitative, Quantitaive and Mixed Methods. Minnesota: University of Minnesota.
Segool, N., & Carlson, J. (2008). Efficacy of Cognitive, behavioral and Pharmacological Treatment for Children with Social Anxiety. Depression and Anxiety, 25(7): 620-631.
Sink, C., & Stroh, H. (2006). Practical Significance: The Use of Effect Sizes in Scholl Counseling Research. Professional Scholl Counselling, 9(5): 401-411.