Over the past several millennia, science, as a systemized method of learning, has led us to a rich and wonderful understanding about the world around and within us. In its earliest form it was simply the gathering of facts and knowledge about the natural world—when to expect the annual, life-giving flood of silt from the Nile, what is the buoyant force of an object in water (Archimedes’ principle), what herbs to use in treating a fever or headache. Later, as the Middle Ages progressed into the Renaissance, science became a legitimate discipline in its own right with the seminal experiments and observations of such notables as Galileo, Copernicus, Vesalius, and Harvey. Science developed rapidly as a method to search, using observation, experimentation, and calculation, for patterns and laws in nature.

In more recent times, science has become more than a storage system for facts and figures. It has become a way of knowing, a way of looking critically at the universe as well as the molecules inside our bodies and predicting how they will behave. This accumulation of information and the ability to construct predictive models has allowed us to apply this knowledge in very practical and useful ways—to explain the most confounding mechanisms of the universe, to cure devastating diseases, and to build materials which alter our lives in dramatic fashion.

The Scientific Method

The process of thinking scientifically has culminated in what has come to be known as the Scientific Method (Figure 1). The starting point for any scientific investigation must be observation. We are all observers of what happens in and around us, and therefore, in a sense, we are all scientists. However, proper science, which goes beyond mere observation and discovery, teaches us that a methodical, systematic approach to observation and conclusion is necessary in order to get accurate explanations for the phenomena around us.

Figure 1. Infographic of the Scientific Method.

Credit: https://newmr.org/blog/what-is-the-scientific-method-and-how-does-it-relate-to-insights-and-market-research/

Asking an appropriate and carefully prescribed question, known as a central question, is the first step in attempting to explain phenomena. The question, Where are our oceans’ fish?, will probably yield less effective explanations than the question, What are the main contributors to loss of fish in the ocean today? Further, asking several specific questions can often yield more useful information than one generalized question. How did the leopard get its spots?, will probably prove less fruitful than, Does a leopard’s spotted pattern help to camouflage it from its prey?, or Can natural selection confer an advantage to a leopard with a certain camouflage pattern of spots?

At this point in the scientific method, a scientist can propose an educated guess or reasonable explanation for a phenomenon. A hypothesis can be developed that falls out naturally from the central question and, if designed properly, is testable. For instance, a hypothesis might be stated as:

Increasing levels of mercury in ocean ecosystems has led to dramatic declines in fish

stocks.

In this example, the researcher can set about measuring or collecting data on mercury at various locations and determine if there is a correlation with declining fish stocks. It may be difficult to determine a cause and effect relationship but the initial study can lead to further hypotheses and testing to gain even more detailed data on why or how fish stocks are declining. An alternative hypothesis might be:

Overfishing and violation of international treaties on fish quotas has led to declines in

fish stocks around the world.

Historically, there may be no clearer example of the process and the progress in use of the scientific method and hypothesis-based testing than that of the effects of cigarette smoking on health. After initial observations and scientific testing by independent researchers, cigarette packs, in 1965, were required by federal law to display the warning: Caution: Cigarette Smoking May Be Hazardous to Your Health, reflecting the correlation that had been discovered between smoking and health. After several decades of further testing, the labels reflect the process in scientific research related to cigarette smoking. As a result, today’s cigarette packs display more forceful and powerful cause and effect labels: Cigarettes cause lung cancer and tobacco smoke causes fatal lung disease in nonsmokers.

Scientists often design what is known as a null hypothesis for an experiment, more in keeping with statistical probability. In this way we can eliminate a hypothesis by “falsifying” it rather than trying to prove that it is “true.” In the example above, we can never be 100% sure that cigarettes cause lung cancer. By stating the hypothesis as a negative, I’m 95% sure that cigarettes are NOT a cause of lung cancer, we have a very precise yardstick for rejecting our hypothesis. We use much the same logic in our courtrooms where the accused is innocent until proven guilty beyond all reasonable doubt (and still, juries sometimes make mistakes)!

From here, the researcher can post a series of predictions that will help to further frame and outline the experimental design. Often, predictions follow an if/then format. Consider the following example. Contrary to expert advice, homeowners often operate under the misguided prediction, if the recommended dose of a particular pesticide kills a certain pest, then a high dose will be even more effective. If an experiment was conducted from this idea the data may not support this hypothesis/prediction and then the homeowner would reject the hypothesis and would need to modify the hypothesis and prediction.

In the case of cigarette smoking, the researcher may formulate a prediction such as, if people are exposed to increasing levels of cigarette smoke, then increased incidence of lung cancer in these people will be observed. After experimentation, we should be able to definitively confirm or reject our hypothesis and the prediction that followed from it.

Once the preliminary steps of question, hypothesis, and prediction have been carefully elucidated, an experimental test that properly addresses the hypothesis can be designed. As any academic researcher can attest, this is often a laborious process. Trying to get the experimental protocol just right so that there is enough detail for others to repeat the experiment, verifying that experimental variables are well-defined, and ensuring that other variables are controlled are all essential for good experimental design. If the experiment is conducted several times or several different researchers conduct the same experiment (such as the lab groups in our lab), it is imperative that each follows the protocol exactly in order to properly address these variables and obtain valid data. Therefore, repetition, or repeating an experiment many times, is also an important aspect of a credible experimental protocol.

In any experiment, the independent variable is the variable we manipulate or change. This is the variable we already know BEFORE we start the experiment. For example, we can establish the age of all participants in a study before we test them for the presence of oral Streptococcus bacteria. The dependent variable is the variable that was measured or recorded as a result of the independent variable being investigated. We do not know the values or data for the dependent variable until we have conducted the experiment. For instance, we would record data regarding which of the participants (age) tested positive for oral Streptococcus bacteria. Lastly, a control variable is one that is held constant or does not change during the course of an experiment. For instance, did the Streptococcus study control for gender, diet, frequency of teeth-brushing, etc.?

Results and interpretations can be presented once the experimental data has been collected. Data can be summarized and systematically organized to determine if they support or refute the hypothesis. Calculations and statistical analysis may be necessary to aid in this determination. Generally, the independent variable is presented on the horizontal, or x-axis, of the figure; the dependent variable is depicted on the vertical, or y-axis. With manipulation of the data, major trends and patterns often become evident and aid in confirming or rejecting the hypothesis. Often, the use of tables and figures allow a clearer examination of the data as well as providing the reader with an effective visual presentation of what might otherwise be obscure or ambiguous.

Choosing the appropriate type of graph or figure for visual presentation of the data is also crucial. Should you use a bar graph or a line graph? Line graphs are appropriate when the independent variable consists of continuous data, when intermediate values make sense. For instance, plant growth over time, or a change in percentage of smokers over time, would generate intermediate values. Bar graphs are appropriate when the independent variable consists of categorical data. This information fits into discrete categories and intermediate values do not make sense. For instance, a study of food preferences among different species of squirrels generates categorical data.

Scientists and researchers know the story does not end here. Like the mythical snake swallowing its tail, the process of science is a cyclical one in which knowledge and conclusions gained from one experiment can be applied in the design and development of related experiments with their new hypotheses and predictions.

Please proceed to the Scientific Method Worksheet