Participation in a science fair can be a positive and even life-changing experience for students. However, science fairs (and classroom labs) can also reinforce misunderstandings of how science works, particularly with respect to the notion of hypothesis formation.

Colleagues and I have investigated student projects at eight International Science and Engineering Fair (ISEF) competitions in recent years for a total of almost 2000 student projects. Students included a hypothesis in 78% of these projects but actually wrote only predictions 81.2% of the time. True hypotheses – potential generalizations or explanations – appeared in only 272 (18.8%) of the projects, as we showed in a 2015 article in The American Biology Teacher (77:17–23).

Thus, even though generating hypotheses is a fundamental component of most scientific investigations, students at the ISEF seem to misunderstand the difference between hypotheses and predictions, and this is likely true at local science competitions too.

This problem seems to start early. For example, in the first-year biology course I teach, nearly all my students fail to distinguish a scientific hypothesis from a prediction early in the year. When prompted to provide an example of a hypothesis, most students write predictive “If…, then…” statements in the form of “If I perform X experiment, then I will produce Y as an outcome.” There is no hypothesis present in statements of this form.

A hypothesis is not a prediction.

In science, a prediction is a forecast of what one is likely to observe or measure in the future given a set of assumptions, controlled conditions, and a data collection method. Well-formed scientific hypotheses make such predictions possible but are not themselves predictions.

There are two types of scientific hypotheses: generalizing hypotheses (“This pattern exists, and with data it might become a future law or relationship”) and explanatory hypotheses (“This mechanism causes this pattern and might become a future theory”).

For example, a student making field observations may propose the generalizing hypothesis (i.e., pattern) that prairie coneflowers with stem damage have more flower heads than those without such damage. With this hypothesis, students can design a method for collecting data to see if this suspected pattern is real. If this relationship is validated, the student may then propose a physiological explanation (mechanism) and suggest the related hypothesis (i.e., immature theory) that mechanical damage, from grazing, to the coneflower apical meristem releases apical dominance and promotes the growth of axillary buds and more flower heads.

However, in science a hypothesis is not always necessary.

Many science fairs and classroom teachers require students to include hypothesis statements in their projects regardless of the type of investigation. But not all scientific pursuits, and thus not all science fair projects, are or should be hypothesis driven. For example, a student might conduct a survey of arthropods in two different fields to explore the potential for patterns in community structure. This student could not possibly have a hypothesis in mind at the outset of the study, since there is nothing yet known to consider as a possible relationship or explanation. It is also important to remember that most non-science projects, such as those in mathematics, engineering, and computing, are focused not on hypothesis testing but instead on the study of measurement, discovering relationships, and designing technical solutions to problems.

Perhaps most science teachers are simply operating from their own prior experiences, drawing on some version of the discredited stepwise “scientific method.” Such models show a hypothesis as a necessary step of moving from research questions to data collection. Textbooks themselves also cause confusion. For example, in preparing for this commentary, I opened a recent physical science text and found that the “scientific method” section in the first chapter provides three different definitions of hypothesis. The authors note that a hypothesis is an educated guess and a hypothesis predicts what might happen. The notes for teachers offer even more potential confusion by stating that a hypothesis is a question.

Cognitive scientists have shown that even children with little or no training in the processes of science can propose functional hypotheses to explain a natural event and can also propose causal hypotheses to explain how one event in nature may affect another (this is a hypothesis with law-like potential). Thus, we may be unteaching a way of thinking that our students already engage in naturally.

So, let’s stop requiring inappropriate hypotheses in science fair and school science projects and instead teach students explicitly that hypotheses and predictions are distinctly different elements within the nature of science. Indeed, we should teach that science is a dynamic process, not a linear method.

For students who do have hypothesis-driven projects, we can help them identify their hypotheses and predictions by asking them questions such as these:

  • Do you feel like you are investigating a potential relationship or a possible mechanism? What is the relationship or mechanism?

  • What sort of data do you predict you could record that will show that the relationship or mechanism is valid?

The science fair and related challenge activities provide students the opportunity to hone inquiry skills that are essential in life and in careers and to see science in its most authentic light. Noticing and testing patterns and exploring the possible mechanisms driving those patterns is at the core of their journeys toward current and future scientific discovery. We can help by teaching students to recognize when, where, and how to use the logical strategies of hypothesis testing that will generate new knowledge and, ultimately, produce more scientifically literate citizens.