SCIN 137 AMU week 4 lesson lab Orographic Lift and Lapse Rate Lab.Introduction to Meteorology American Military university
Topics to be covered include:
- Introduction to experimental design
- Variables and factors to consider in experimental design
- Principles of treatments, randomization, replication, and sample size
- Case studies to analyze elements of experimental design
In the last lesson, we covered the characteristics of an effective hypothesis and practiced analyzing scenarios and a research abstract to determine null and alternate hypotheses. Now we will build off that knowledge and learn how a hypothesis guides the formation of an experiment. A hypothesis includes independent and dependent variables, which in turn helps researchers determine how to set up an experiment and which statistical analyses are most appropriate to test the relationship between the two types variables. Researchers must also take care to control any additional variables that could confound true relationships and avoid bias through randomization and replication of the experiment to generate sufficient data.
The Study was Not Replicable
Additional research labs attempted to verify Wakefield’s results, in keeping with the scientific method, by replicating his study, but a decade’s worth of studies has failed to uncover a connection between vaccinations and autism (Rao & Andarte, 2011).
Replication, repeatability and sample size all lend credibility to a stud. How do scientists replicate experiments? Replication is sometimes performed by the original scientist but is more often repeated in other labs. The expectation is that if the result is describing a natural phenomenon in the universe it should be observation in other laboratories in other locations (Understanding Science, 2017).
The Study Demonstrated Conflicts of Interest
Scientists help ensure the trustworthiness of their research by avoiding conflicts of interest or freely disclosing where they receive their funding. This safeguard minimizes the effects of bias on the results of a study. In Wakefield’s case, he did not disclose that lawyers were paying him as a medical expert to advise parents who were suing over alleged vaccine injury. Furthermore, some of the children of these parents were also subjects in his research (Novella, 2010). He also did not disclose that he had filed a patent for a vaccine that would compete with the MMR vaccine (Barrett, 2010). Disclosing conflicts of interest are essential to building credible science.
The Study had a Small Sample Size
Wakefield’s study only looked at 12 children, which is an extremely small sample size for a medically relevant study (Rao & Andrade, 2011). As we will learn, sample size plays a major role in statistical power, with more samples providing more data and, therefore, more trustworthy results. With a sample size of only 12, the data from even one child could potentially cause conclusions about underlying patterns or relationships between the independent variable and the dependent variable that are not accurate. Preliminary studies especially in human research may be small; however, if results support the initial hypothesis they are followed by large sample size experiments to verify results with greater statistical confidence.
As a point of reference, a clinical trial that led to the Food and Drug Administration (FDA) approving Kyleena, a five-year intrauterine device for contraception, included 1,452 women (Bayer, 2016). Sample sizes of this magnitude allow for greater confidence in statistical analyses and help ensure that any random variation in the data will not mask significant patterns, such as the device’s ability to prevent pregnancy. Large sample sizes also help reassure women who use IUDs that they will effectively prevent pregnancy.
The Study Acquired Samples Unethically and, Quite Possibly, Illegally
Any research that involves live animals or humans, must receive approval from appropriate legal boards to ensure that the study is conducted ethically and safely. Proposals for human research is usually evaluated by an Institutional Review Board (IRB). Wakefield obtained blood samples for his study from children attending his son’s birthday party, paid them for their discomfort, and joked about it during a lecture (Barrett, 2010). He also performed procedures such as colonoscopies and lumbar punctures on children with autism without prior approval from a research review board (Barrett, 2010). These practices are counter to accepted ethical standards that medical doctors and scientists are subject and agree to within their institutions. If research will involve the use of vertebrate animals, an Institutional Animal Care and Use Committee (IACUC), ensures that researchers treat laboratory animals humanely.
The Authors Committed Deliberate Fraud to Support Their Hypothesis
You can get the recommended vaccinations for all age groups at:For Everyone: Easy-to-read Schedules
Finally, the authors stated the data was conclusive to support their hypothesis but it was not, which helps explain in part why follow-up studies could not replicate their findings (Rao & Andrade, 2011). The lab that they used to search for measles in the intestines of the study subjects used flawed techniques to show false positives (Novella, 2010). Furthermore, the British Medical Journal outright stated that the authors falsified facts and picked and chose data to support their hypothesis, most likely for financial gain (Rao & Andrade, 2011).
Since Wakefield’s study was debunked, vaccinations rates have risen back to 91 percent in the United Kingdom, but the anti-vaccination movement still persists (Editorial Board, 2012). Researchers blame measles outbreaks in the United Kingdom in 2008 and 2009, as well as outbreaks in the United States and Canada, on lower vaccination rates possibly related to this study (Rao & Andrade, 2011). Parents may still fear that their children will develop autism simply due to lack of educational outreach, or they advocate for parental choice over which vaccines their children receive or a delayed vaccination schedule (Sohn, 2017). Studies have shown spikes in levels of vaccine preventable diseases in schools globally (Sohn, 2017). Overall, the anti-vaccination movement demonstrates the severe repercussions that can result from poor experimental design and why researchers must take great care to ensure the trustworthiness of their results.
Designing an Effective, Reliable Experiment
Experimental design begins as soon as researchers state their hypothesis. In some cases, hypothesis formulation and experimental design go hand in hand. A hypothesis essentially defines the independent and dependent variables and states what the researchers will be testing. Formal predications are written under hypotheses for the specific scientific experiment which identifies the specific test population, location, and other specific variables. A hypothesis and experimental design will help researchers identify which statistical analyses are most appropriate and how to conduct an experiment that will provide the necessary data to test the hypothesis.
So how do we design a scientifically valid, reliable experiment? As we saw earlier, experimental design represents a complex process with numerous factors to consider to avoid a debacle like that one that spawned the anti-vaccination debate. Let’s consider each of these factors in turn.
Relating the Hypothesis to the Experiment
First, the hypothesis should relate directly to the experiment. The hypothesis outlines the independent and dependent variables, and the experiment should test the relationship between those variables. Ideally, the hypothesis should make experimental design clearer.
Let’s consider an example of a plant growth experiment with independent, dependent and controlled variables. The hypothesis might be: plant growth height is decreased with increased salt concentration. The independent variable is salt concentration and the dependent variable would be plant growth height. This experiment might have two treatments a high salt concentration and low salt concentration. There would also be a control group that received water only. Other variables that would need to be accounted for, so that the correct relationship between growth height and salt is measured, include seed type, amount of sun, humidity levels, type of soil. Each of these variables should be the same between the two treatment and control variables. There would be multiple plants in each treatment. The dependent variable would be measured which is plant height for each individual plant. The prediction is that plants that received no salt will be tallest followed by the low concentration plants and then high concentration plants.
It is common to revise or even expand a hypothesis during the course of an experiment. A researcher could design an experiment to test a hypothesis but encounter difficulties with lab techniques or field research. Or perhaps the researcher could analyze preliminary data and revise and retest the hypothesis based on those findings. The scientific method is a dynamic process, and this concept applies to the interactions between hypotheses and experiments.