Hypotheses

Your favourite part of summer is experimenting in your garden; this year you've decided to try your hand at tomato plants. But no matter what you try, you can’t seem to get the tomatoes to ripen! You decide to add some fertilizer, guessing that it will help your tomatoes. You might not know it, but you just became one step closer to being a scientist.

A hypothesis (pluralized as hypotheses) is a proposed explanation for a specific observation.¹ Specifically, it attempts to describe the relationship between an independent variable (the variable that is manipulated) and a dependent variable (the variable that is being measured). In the gardening situation, the hypothesis that adding fertilizer will help the tomatoes ripen is proposing that a lack of fertilizer was impeding the process. The presence or absence of fertilizer is the independent variable and the tomato’s ripeness is the dependent variable. In order for a hypothesis to be scientific, it must be testable and thus falsifiable. Hypotheses are based on existing scientific data, personal experience, or intuition.

Independent versus dependent variables

Independent variable: The variable that is manipulated in an experiment, which is assumed to have a direct effect on the dependent variable.

Dependent variable: That variable that is being measured in an experiment, which will change depending on how the independent variable is manipulated.

Simple versus complex hypotheses

Simple hypothesis: Proposes an explanation for a relationship between one independent variable and one dependent variable.

Complex hypothesis: Proposes an explanation for the relationship between two or more independent variables and two or more dependent variables.

Directional versus non-directional hypotheses

Directional hypothesis: Proposes an explanation for a relationship between variables that can also predict its directionality.

Non-directional hypothesis: Proposes an explanation for a relationship between variables, for which there is no theory or prediction about the nature (direction) of that relationship.

Null versus alternative hypotheses

Null hypothesis: Proposes that there is no relationship between the independent and dependent variables.

Alternative hypothesis: Proposes that there is a cause-and-effect relationship between two or more variables, contrary to the null hypothesis.

Since hypotheses are a core component of experiments, the history of the hypothesis begins with Galileo Galilei, the Italian astronomer and physicist.² He has been described as the “father of experiments” for his mathematical reasoning that was used to overturn the religious dogma regarding Earth’s position in the universe in the 1500s. Galileo’s use of deductive reasoning categorized the hypothesis as a starting point based on unproven assumptions. Deductions were made from this premise, whose successes or failures were determined by subjective assessments on whether they were satisfactory explanations for the premise. While Galileo experienced “success” with his approach, the lack of a foundation for the hypothesis resulted in confusion for other physicists in the century.

After Galileo set the stage for hypotheses to be grounded in realism, philosopher Francis Bacon wrote Novum Organum in 1620, describing an approach to scientific methodology.² Bacon posited that deductive reasoning alone was not sufficient: since the premise is determined before the experiment, the reasoning could easily be manipulated to fit the premise. Rather, Bacon argued, scientific methodology needed to be founded in pure experimentation. Instead of deductive reasoning, scientists should use inductive reasoning: data is used to infer that the same result will be replicated under similar circumstances, and the specific case’s finding could be generalized to other similar cases.

Inductive reasoning can be applied to the case of gravity: it is observed that a mass falls toward earth at a certain rate, which it continues to do in predictable ways.² This means that after a rule is established for how quickly objects fall, it can be predicted that all falling objects will follow the rule. Using this example of gravity, we can bring physicist Isaac Newton into the history of the hypothesis. When Newtown published the 1713 edition of his seminal work that outlined his laws of motion, the Principia, he included the phrase “hypotheses non fingo” which translates to “I frame no hypotheses.” Later in 1721, Newton stated in Opticks - a book that analysed the fundamental nature of light - that “hypotheses are not to be regarded in experimental philosophy.” Clearly, Newton rejected the hypothesis.

Further rejection in the hypothesis’ development came from philosopher David Hume, who specifically rejected the notion of induction.² Hume proposed a “radical skepticism”: past experiences cannot be used as proof or predictions of future outcomes, because such claims would be based on the unprovable premise that something and its attributes will remain bound together. Nature’s laws are not stable, which is the fundamental assumption behind inductive reasoning. Such an assumption could only be justified through circular reasoning, believing that nature is uniform because it has been in the past.

Similar to Hume, philosopher Karl Popper also rejected the inductive method of empirical sciences, but he took it one step further by providing a solution to inductive reasoning.² Instead of testing and verifying hypotheses by obtaining the same outcome, hypotheses could only be validated through the “falsifiability criterion”: scientists actively work to find exceptions to their hypothesis. If no contradictory evidence is found, then there is support for the hypothesis. Popper’s approach has been termed “critical rationalism” and is the methodology used by contemporary scientists.

Depending on the results of an experiment, a hypothesis is either rejected as false or accepted as true.¹ However, since all hypotheses are inherently falsifiable, they may be rejected at a later point if disconfirming evidence arises, regardless of whether they are accepted as true and supported by evidence at the time of experiment.

When a hypothesis is rejected, scientists will sometimes adapt the existing hypothesis to accommodate the new falsifying information, rather than disregarding the original idea.¹ This is why you may hear the scientific community say that hypotheses are never incorrect, and rather incomplete: hypotheses were rejected as false because a piece of the puzzle was missing from that specific hypotheses and experiment.

A common misconception is that hypotheses can be proven as accurate.⁶ While alternative hypotheses - relative to the null hypothesis - can be supported with evidence, this does not equate to proof. The very nature of falsifiability ensures this: there is always a chance for evidence that refutes the hypothesis to be available, yet unknown to the researchers at the time. This is how existing literature informs future research: if a hypothesis is rejected, future studies may explore why this is.⁷ If a hypothesis is supported, researchers may focus on more nuanced aspects of the original idea or look for disconfirming evidence, sometimes which can even be found accidentally.

Ultimately, we could not have scientific theories without scientific hypotheses.¹ This is an important distinction and can also be a point of misunderstanding.⁶ Hypotheses are a specific, tentative explanation for a phenomenon, and is a research tool that gathers data.¹ On the other hand, theories are broad, general explanations that include data from multiple scientific investigations. Thus, scientific theories are the result of numerous scientific hypotheses: it would be poor research practice to take one hypothesis as pure “proof” and to formulate a theory around it. Contrary to hypotheses, scientific models do need to be verifiable, with its success defined as its ability to predict specific outcomes.²

COVID-19 and the lab-leak hypothesis

As we pass the one-year mark of COVID-19 being categorized as a global pandemic, conversations surrounding its cause and virality have not subsided.⁸ Some discussions have centered around the lab-leak hypothesis, which proposes that the origin of the virus came from a laboratory accident. In May 2020, the World Health Organization labelled the lab-leak hypothesis as “extremely unlikely,” while the hypothesis of a zoonotic spillover - transmission from animals to people - was “likely to very likely.” ⁹ 

However, an open letter was published in the academic journal Science on May 14, 2021 by a group of 18 scientists.⁹ These scientists did not weigh in on evidence for either hypothesis, but voiced their concerns that the conclusion regarding the zoonotic spillover hypothesis - relative to the lab-leak hypothesis - was premature. They stated that both hypotheses must be taken seriously until there is sufficient data and that a proper investigation should focus on both hypotheses.

While there are other scientists who do not agree that the lab-leak hypothesis could be plausible, the open letter emphasizes the importance of hypotheses being falsifiable and testable, supported by hard data.⁸ While we can sometimes feel detached from scientific research and the academic community - due to difficult words and alien theories - research plays a large role in our daily lives. Certainly, this can be seen as the case with the COVID-19 pandemic.

Corporate sustainability and the Porter hypothesis

Corporate sustainability refers to an approach where organizations focus on ethical, social, economic, cultural and environmental dimensions of conducting business.¹⁰ One factor related to corporate sustainability is pollution prevention and control: although there is consensus that government legislation is required to regulate the environmental responsibilities of corporations, there is still debate on how these regulations should be formulated and how corporations can use said regulations to improve their own performance.

The Porter hypothesis, developed by economist Michael Porter in 1991, proposes that strict environmental regulations can result in a win-win scenario: efforts are made toward sustainable efficiency, while companies increase their overall performance and thus competitive edge.¹⁰ There is a wide range of literature that considers factors that would be important for the Porter hypothesis, two of which are flexible regulations (innovation friendly regulations that encourage businesses to develop new products and processes to meet regulatory requirements, rather than prescribed, specific processes) and innovative capabilities (underscored by corporations’ flexibility in meeting regulations).

In 2017, a group of researchers developed a framework to evaluate the design of environmental regulations and the ability of firms to achieve Porter’s win-win outcomes.¹⁰ After sampling case studies of major corporations, the researchers found that three assumptions related to the Porter hypothesis appear to be valid:

1. Inflexible regulations force firms to be reactive and adversely affect financial performance;
2. Flexible regulations help innovative firms meet regulations and improve performance; and,
3. Firms without innovative capabilities cannot improve their financial performance, even with flexible regulations.

Overall, the researchers provided valuable insights regarding the way that governments can design environmental regulations to encourage corporate sustainability, as well as how businesses can improve their private benefits of sustainability.¹⁰ For example, aligned with the Porter hypothesis, reducing consumption of energy and raw materials that reduce waste and pollution can be linked to better market performance.

Just-world hypothesis

Some people may not believe they play much of a role in the scientific community, and thus experience a disconnect when reading about hypotheses. However, we make hypotheses all the time, such as believing the world is fair and that the morality of our actions will determine our outcomes. As we navigate life, we may or may not experience disconfirming evidence to this just-world hypothesis.

TDL Brief: What’s next for polling?

Modern issues call for modern hypotheses. The fight for power between the Republican and Democratic parties in America is an influential one, and has seen an interesting pattern: there are few states where polls suggested that the Republican candidate would win, which were actually won by the Democratic candidate. This article explores some hypotheses for why this polling pattern occurs.

1. Rogers, K. (2018, September 5). Scientific hypothesis. Encyclopedia Britannica. https://www.britannica.com/science/scientific-hypothesis
2. Glass, D. J., & Hall, N. (2008). A brief history of the hypothesis. Cell, 134(3), 378-381.
3. Urbach, P. M. (2021, April 5). Francis Bacon. Encyclopedia Britannica. https://www.britannica.com/biography/Francis-Bacon-Viscount-Saint-Alban
4. Karl Popper. (2020, September 13). Encyclopedia Britannica. https://www.britannica.com/biography/Karl-Popper/additional-info#history
5. Santhanam, L. (2015, April 2). Babies resemble tiny scientists more than you might think. https://www.pbs.org/newshour/science/babies-resemble-tiny-scientists-might-think
6. Sotos, A. E. C., Vanhoof, S., Van den Noortgate, W., & Onghena, P. (2007). Students’ misconceptions of statistical inference: A review of the empirical evidence from research on statistics education. Educational Research Review, 2(2), 98-113.
7. Hevner, A. R. (2007). A three cycle view of design science research. Scandinavian Journal of Information Systems, 19(2), 87-92.
8. Danner, C. (2021, May 14). The COVID lab-leak hypothesis just got a big credibility boost. New York Magazine. https://nymag.com/intelligencer/2021/05/covid-lab-leak-hypothesis-just-got-a-big-credibility-boost.html
9. Bloom, J. D., Chan, Y. A., Baric, R. S., Bjorkman, P. J., Cobey, S., Deverman, B. E., … & Relman, D. A. (2021). Investigate the origins of COVID-19. Science, 372(6543), 694.
10. Ramanathan, R., He, Q., Black, A., Ghobadian, A., & Gallear, D. (2017). Environmental regulations, innovation and firm performance: A revisit of the Porter hypothesis. Journal of Cleaner Production, 155(2), 79-92.