How to Test a Hypothesis: Bill Nye on Filters and Flags

Bill Nye is an award-winning entertainer and educator; former engineer at aviation giant Boeing; and CEO of the Planetary Society. He regularly speaks on climate change and other scientific matters on television shows, as well as at his alma mater and other universities. Wherever he goes, he brings a knack for engaging the population at large—adults and children alike.

Hailing from Washington, DC, Bill earned a bachelor’s degree in mechanical engineering from Cornell University in Ithaca, New York, where he studied under legendary astronomer Carl Sagan. By the time he launched his namesake television series, Bill Nye the Science Guy, in 1993, he had invented a hydraulic pressure resonance successor tube used in Boeing 747 airplanes to this day. He’d also performed as a stand-up comedian on Almost Live!, a sketch show based in Seattle, Washington.

In 2017, the streaming platform Netflix premiered Bill Nye Saves the World, a series aimed at fact-finding and solutions; it ran for three seasons and garnered three Emmy Award nominations. He has authored eight children’s books and three geared toward adults; led global marches and rallies in conjunction with Earth Day; and starred in a documentary, 2017’s Bill Nye: Science Guy, which chronicles his crusade against irrational thinking.

A hypothesis is an educated guess you can utilize to spearhead an experiment. In a statistical sense, a null hypothesis asserts you can expect a normal distribution with typical standard deviations from your data. On the other hand, an alternative hypothesis asserts you can prove something of unique statistical significance about a dataset previously unacknowledged or unknown to the scientific community.

Testing a hypothesis always requires you to use certain tools. Here are some of the key elements necessary for hypothesis testing:

• Alternative hypothesis: This type of hypothesis refers to what you formulate as a challenge to the established order of things. If you prove your alternative hypothesis has a high level of significance, it has the potentiality to overturn a null hypothesis.
• Null hypothesis: This element serves as a criterion or test statistic of sorts. A null hypothesis is what the scientific community already accepts as true about a given subject.
• Statistical tests: You’ll need to statistically evaluate your population parameters and datasets as you test your hypotheses. Statistical hypothesis testing of this ilk might entail chi-square tests, t-tests, and z-tests to root out standard errors and confirm the importance of your findings.
• Variables: You’ll use an independent variable as a control group and a dependent variable as a group you can experiment upon to confirm your hypothesis. The interaction of these two groups helps form the basis of your experiment.

Scientist and educator Bill Nye says testing a hypothesis can be “a great starting point for critical thinking.” Here are three filters and flags he suggests you use when evaluating your own:

• Ask who stands to benefit. When evaluating data, Bill thinks it’s essential to question the source. As an example, think of a company that makes a claim that paints their products in a positive light. “Don’t get me wrong, there are some advertising claims which are completely true,” Bill notes. “Just be sure that they are.” The same applies to any other institution as well.
• Assess your own knowledge. Rely on what you already know to assess whether a hypothesis is on track or off base. “If someone says, ‘The moon is made of cheese,’” Bill says, “you might know that the moon is made of rocks.” While you should be ready to learn new things when testing hypotheses, you can still rely on your learning and common sense to evaluate information.
• Evaluate your own desire. You have a vested interest in proving your alternative hypothesis to be true in relation to a null hypothesis. Bill calls this, “the last red flag, the most dangerous one!” In other words, prevent your desire to be correct from skewing your objective analysis of the data you have before you.

With enough curiosity and a large enough sample size of data, you can test just about any hypothesis. Follow these steps to do so:

• Ask a question. Look over a set of null hypotheses and think about whether you think any might be inaccurate. Formulate a research question to this effect. You’ll use this as a statistical inference and alternative hypothesis to begin your experiment.
• Assign variables. When you have a hypothesis you seek to test, assign both an independent variable (a group of data points you’ll hold constant) and a dependent variable (a group subject to change throughout your experiment). You’ll spend most of your experiment focusing on a sampling distribution tracking the interaction of these two datasets.
• Collect data. Statistical hypothesis testing requires you to gather sufficient evidence in the form of ample sample data. Start to collate plenty of information to pad out both your dependent and independent variables. Make sure you can state both what the null hypothesis assumes and what your alternative hypothesis seeks to prove. Find the sample mean in all the data you bring together.
• Evaluate in numerous ways. After you’ve gathered enough evidence, there are a multitude of ways to compute data as you seek to prove your hypothesis. Supplement your testing by performing confidence interval analysis. Scan your data for different types of errors. For example, try to avoid both type I errors (in which you reject a true null hypothesis) and type II errors (in which you fail to reject a false one).
• Submit your findings. Pass your research hypothesis, as well as your relevant experimental and statistical findings, along to other qualified professionals. Someone might discover something you passed over in the rejection region of the sample statistics you provided, or perhaps another person will further validate your conclusions with their own research.

Get the MasterClass Annual Membership for exclusive access to video lessons taught by science luminaries, including Bill Nye, Terence Tao, Neil deGrasse Tyson, Chris Hadfield, Jane Goodall, and more.