This paper critiques an observational study conducted on cyberbullying, which engaged in longitudinal analysis of adolescent attitudes and behaviors regarding the practice. It then discusses more general issues involved in constructing experimental and observational research studies, such as the value of research in business and the need for appropriate sampling techniques.
Cyberbullying
Independent and dependent variables used in the study
In an experiment, the independent variables are the variables manipulated by the experimenter, while the dependent variables are the aspects of the experiment that are 'dependent' or affected by the independent variables. In Hinduja and Patchin's 2011 research for the Cyberbullying Research Center, the independent variables under study were the rates of cyberbullying, while the dependent variables were the characteristics associated with teens who engaged in bullying.
The study found that cell phones were the preferred method to use in cyberbullying and girls were found to be far more likely to engage in the practice. There was no distinction in terms of the racial profile of cyberbullying victims or perpetrators -- all were equally apt to engage in the crime. Other characteristics correlated with cyberbullying were a negative school atmosphere. Regarding the victims, victims were more likely to have low self-esteem and to have suicidal ideation than those students who had not been bullied.
Sampling used to gather subjects: Reliability and validity of the study
The study involved the use of longitudinal data collected from multiple sources of similar demographic groups. A February 2010 study involved a random sample of 4441 youth between the ages of 10 and 18 from a large school district in the southern United States. 37 schools were involved in the sample. The June of 2009 study surveyed a random sample of approximately 900 youth between the ages of 11 and 18 from a moderately-sized school district in the southern United States. The sample consisted of 8 different schools. The June 2007 study surveyed a random sample of approximately 2000 youth between the ages of 11 and 16 from a large school district in the southern United States from 30 different middle schools (Hinduja and Patchin 2011).
Other than the regional specificity, the study sampling was designed to be sufficiently broad to be representative of a large, general population of students. The use of different school districts and different sampling sizes was designed to make use of the 'law of large numbers' regarding the study population, to increase validity -- namely that random samplings of large numbers of the target population are more likely to yield representative and accurate results than a small sampling size. The size of the sampling, and the fact it was taken over time, supports the validity of the study.
Reliability is demonstrated that, through three different samplings of slightly different populations within the same age range and demographic, similar results were yielded regarding the study.
How an experimental research methodology can be used to solve a management problem
Managers must often grapple with problems related to human behavior, much like the cyberbullying of adolescents being studied in the survey. Managers can survey employees of sufficiently large numbers to determine if employees are satisfied with compensation or diversity training programs or other components of their work and engage in longitudinal surveys of those attitudes, much like the designer of the cyberbullying study.
To create a truly experimental, rather than purely observational study, however, managers would have to introduce a 'change' between the control and experimental subject groups under study. For example, when measuring the effectiveness of a new sensitivity training program, the managers might survey employee attitudes and knowledge of harassment policies before the program, after the program, and then six months later, to see if the training had a measurable impact upon attitudes. The 'before' group would serve as the 'control' group, compared with the attitudes of the experimental group 'after' they had been subjected to training. In this experimental study, the conditions of the 'control' and 'experimental' group would be able to be controlled by the managers. For example, the managers would be reasonably confident that the subjects would not be attending other sensitivity training elsewhere that could affect the results.
How structured observational methods can be applied to management decisions
It should be emphasized, however, that not all studies need to be experimental in nature. Some studies can be purely observational in the sense that the conditions surrounding the experiment are not controlled. For example, managers could study employee attitudes regarding race and gender at different intervals, just to gauge the overall attitudes and tolerance of the workforce. Observational studies are often embarked upon as preliminary studies to provide the research foundation for a longer, experimental study or to design something like a rewards program or incentive plan.
Observational methods in general can also be used to set the foundation for a longer, larger study. Observing how individuals interact with one another during meetings and casually in the workplace, and notes upon performance reviews regarding positive or negative workplace attitudes can also highlight whether the company is functioning well, in terms of its interpersonal relationships
How the scientific method can be applied to the decision-making process
Business decisions can be affected by emotions, such as when there is change resistance and employees (or managers) refuse to change because of a desire to keep things 'the way we have always done them.' This is not necessarily a bad thing, because customers make emotional decisions, and ultimately managers are dealing with human beings. However, scientific research provides data-driven, hard evidence that can temper emotions and enable managers to make sound decisions based in logic and real world circumstances.
Describe the various sampling techniques presented in this course
With random selection, "you must set up some process or procedure that assures that the different units in your population have equal probabilities of being chosen" (Trochim, 2006, Probability sampling). In contrast, with nonprobability samples, such as accident or convenience sampling, there is no such a guarantee (Trochim, William, 2006, Nonprobability sampling).
Evaluate the appropriateness of various sampling techniques for a research application
You’re 80% through this paper. Sign up to read the full paper.
Sign Up Now — Instant Access Already a member? Log inAlways verify citation format against your institution’s current style guide requirements.