Scientific Method

The scientific method is the process by which scientists of all fields attempt to explain the phenomena in the world. It is how science is conducted--through experimentation. Generally, the scientific method refers to a set of steps whereby a scientist can form a conjecture (the hypothesis) for why something functions the way it does and then test their hypothesis. It is an empirical process; it uses real world data to prove the hypothesis. There is no exact set of $x$ number of steps to conduct scientific experiments, or even some exact $y$ number of experiments, but the general process involves making an observation, forming an hypothesis, forming a prediction from that hypothesis, and then experimental testing. The scientific method isn't limited to the physical or biological sciences, but also the social sciences, mathematics, computing and other fields where experimentation can be used to prove beliefs.

We could observe that whenever a fire is smothered, it goes out. For instance a small fire that is covered with a blanket is extinguished. We could hypothesize that the reason for this is that fire requires some gas in our air to form and remain a flame. We could then use a vacuum chamber to test this theory.

---

We would predict that outside of a vacuum, a fire could be lit but inside of a vacuum, with no air, that the fire would not ignite. If we were to test this theory, perhaps in multiple vacuums with multiple forms of tinder/fuel (wood, paper, petrol, etc.) and multiple means of ignition, we would notice that the fire never ignites.

If we wished, we could further refine our hypothesis, suggesting that fire can only ignite if there is sufficient oxygen in the air. This we'd also test in the vacuum chamber, by pulling out all the air, then adding in different gases. We would notice that the fire would only ignite in the presence of oxygen or an oxidizing agent.

It is possible that other, incorrect hypothesis could have been initially formed--such as smothering decreases the surface area the fire has, and could try making different sized fires--and been proven incorrect. Also, it is important to note that this single set of experiments is not enough to turn this hypothesis into a theorem. More experimentation and discovery would be necessary.

The scientific method also refers to the fact that science is ongoing. In some cases scientists continue to collect data to prove and disprove old theories. Or in other cases, scientists have hypothesis for why the universe behaves the way it does but are unable to gather sufficient data to prove their hypothesis. For instance, until recent discoveries at LIGO scientists could not confirm what happened when two black holes collided, although they believed (and it was confirmed in February 2016) that colliding black holes produced gravitational waves.

The scientific method is often presented as a set of steps, but not always with the same number or type of steps. However, philosophers of science generally agree that any presentation of the scientific method should have the following four steps:

1. Observe - Sometimes referred to as characterizing, defining, or measuring, experimenters first witness some aspect of the universe, for instance, an apple falling. These observations then form a question, such as "Why do objects fall to the earth?"
2. Hypothesize - Scientists then come up with a theory as to why this happens, for instance, the mass of the earth attracts the apple from the air to the ground.
3. Predict - Using the hypothesis, a scientist calculates what measurable data points they believe will result in a given experiment, for instance an apple at a height of $9.8$ meters should fall to the ground in $\sqrt{2}$ seconds, or should be at a velocity of $9.8\sqrt{2}$ m/s the moment before it hits the ground.
4. Experiment - A test is run to determine if the prediction was correct.

With the notion that repeating these steps is also important. If a prediction is proven to be incorrect then alternative predictions and tests are conducted. Maybe even a new hypothesis could be formulated. Even if the hypothesis and prediction are correct, additional predictions and tests need to be run to best support any theory.

While this process can be explained or categorized differently than this, all formulations of the scientific method have empirical observations, a testable hypothesis, and testing data to prove or disprove that hypothesis. Crucial to this, is that an experimenter searches for experiments that produce the most unlikely results and experiments that are least likely to be coincidental. Hypotheses that produce highly unlikely predictions, in situations where little else could explain the result, are more likely to be true. Bayes' theorem can be used to show which predictions are more or less unlikely given some evidence, i.e. which proven predictions are "stronger" than others. For instance, the theory of evolution has been supported by the consistency of DNA across species whose phenomenology are significantly different. Despite the diversity of plant and animal species on Earth, the majority of our DNA is the same, and only 20 amino acids are the building blocks for every known living organism. It would be highly unlikely that vastly different forms of life have the same building blocks after millions, if not billions, of years of external manipulation, if not for some common origin.

The word "theory" can lead to confusion about how true some scientific principle is. Under the scientific method scientists use the word "theory" even for key principles (like gravity) that have been rigorously proven by modern science. This is because the scientific community believes it is important that hypothesis be falsifiable. Falsifiability refers to the fact that theories have been tested in experiments where they could have failed but did not. So when scientists refer to a principle as a theory, for instance Einstein's theory of relativity, they're actually referring to a hypothesis that has undergone the scientific method, i.e. that has been tested and proven true.

For instance, scientists sometimes refer to evolution as the "theory of evolution," which has contributed to the erroneous belief that the modern scientific theory of evolution is false. Really what the "theory of evolution" refers to is the ample research, testing, and empirical evidence that all consistently prove evolution to be true.

That isn't to say that theories can't be later disproven. Part of the advantage to the scientific method is that no theory is ever considered an unbreakable rule. Some theories seem correct given experiments that are run at the time they're created, but are proven wrong as new methods of experimentation are conducted. For instance, Einstein himself believed that the universe was static, not growing or contracting. That was later proven to be false and replaced with a theory that the universe was expanding (the Friedmann-LeMaitre model of an expanding universe, which Einstein himself accepted), but that its rate of expansion was slowing down. This was, in turn, also proven incorrect. The rate of the universe's expansion is speeding up.^[1] Generally though, theories are modified over time, they are shown to be true under certain conditions, or partly true, and the strength of a theory may also be related to how long it has held up, without modification, to scrutiny.

Peer review:
In modern science, experimenters present both their findings and their methodology for review by their peers, other talented scientists and experimenters. This is done before a work is published, but also publication itself is considered a way of inviting peer review. By sharing and disseminating work widely, the greatest number of others can review the work and offer criticism as needed.

Reproducibility:
Related to peer review, is the notion that the results from experiments should be possible to reproduce. If one scientist conducts some experiment, others should be able to conduct the same experiment on their own and achieve the same results. Reproducible experiments strengthen theories.

Double-Blind Testing:
Primarily used in medical, psychological, and behavioral economic testing, double-blind testing refers to having a test and control group, and running the experiment such that the person conducting the experiment does not know which is which. For instance, in testing the efficacy of a new drug, a pharmaceutical company may have a medical practitioner administer the new drug to one third of the test population, an existing known drug to another third, and a placebo, meaning something that isn't a drug but seems like it, to the remaining third of the test population, but without the nurse knowing which drug is which. The practitioner would then, still blind, track the progress of the entire testing population, gathering data about each test subject.

Double-blind studies are done to avoid biases that manipulate data, like controlling for the placebo effect where just giving a patient a drug that they perceive will be a cure can be causally linked to a decrease in symptoms. This positive causal effect occurs even with the drug that shouldn't affect the patient in anyway, when it is a sugar pill, or water, so long as the patient believes they are receiving a cure. Also double-blind studies help prevent observation bias, where the administrator of the drug may expect the population who received the new drug to outperform others, and so many inadvertently rate their progress better than other test groups.

The theory of the scientific method has evolved over time, with modern historians pointing to Aristotle as an originator, and many looking to Thomas Kuhn's seminal work "The Structure of Scientific Revolutions" as a key influence on current conceptions of the method.

Aristotle classified reasoning into three types:

• Abductive - Also known as guessing, abductive reasoning supposes that the most likely inference is correct. While this isn't rigorous, a well-informed individual is likely to make good guesses, and many significant theories of science have developed first from a guess.
• Deductive - Deductive reasoning uses premises to reach conclusions. One of the most famous examples being "All men are mortal. Socrates is a man. Therefore, Socrates is mortal."
• Inductive - Inductive reasoning is the one preferred by scientists, and can be considered an early version of the scientific method. Namely, inductive reasoning uses empirical observations to make inferences, and accounts for probability in those inferences. A theory reached by induction is said to be more or less likely to be true, stronger or weaker.

The philosophy of science refers to the logic and thinking behind the scientific method. It questions what makes something scientifically valid. For instance, the scientific method assumes that reality is objective, and that explanations exist for all phenomena humans can observe.

Thomas Kuhn's book is foundational to the philosophy of science and the way sociologists and historians look at science through the ages. In it, he popularized the term "paradigm shift" and promoted a historical understanding of scientific discovery not as a linear accumulation of understanding, but as a set of scientific revolutions that "shift" humanity's understanding. Further, paradigm shifts open up whole fields (for instance quantum mechanics, behavioral economics or genetics) with new approaches to understand the universe. Also what scientists consider true is not purely objective, but based on the consensus of the scientific community.

1. Nobelprize.org, . The Nobel Prize in Physics 2011 Saul Perlmutter, Brian P. Schmidt, Adam G. Riess. Retrieved October 24th 2016, from http://www.nobelprize.org/nobel_prizes/physics/laureates/2011/