Scientific evidence

Scientific evidence consists of empirical information used to support or counter a scientific theory or hypothesis. Scientists also apply this evidence to solve practical problems or to develop new experimental tools and techniques. For information to be considered scientific evidence, it must be empirical and interpretable through the scientific method. The strength of this evidence is not a fixed value. Instead, it varies based on the specific field of study. Generally, the strength of evidence is determined by the results of statistical analysis and the quality of scientific controls used during testing.

How a person interprets evidence is heavily influenced by their background beliefs. These are the assumptions or prior knowledge a person holds about the world. These beliefs act as a filter for how observations are utilized. For example, the apparent lack of motion of the Earth once served as evidence for a geocentric cosmology, where the Earth was the center of the universe. However, once enough evidence was presented for heliocentric cosmology, that same observation was discounted as evidence for a stationary Earth. This shows that the observer, rather than the observation itself, creates the causal link between a fact and a hypothesis.

One formal way to understand this influence is through Bayesian inference. In this mathematical approach, beliefs are expressed as percentages to show a person's level of confidence. A researcher starts with an initial probability, known as a "prior." They then update this probability using Bayes' theorem after observing new evidence. Because two independent observers might have different priors based on their previous observations, they may rationally arrive at different conclusions from the same event. This demonstrates how personal history shapes scientific reasoning.

Philosophers have long debated the logic of how evidence relates to theories. In 1959, Karl Popper introduced influential ideas regarding the scientific method. He noted an asymmetry between proving a theory right and proving it wrong. Evidence can prove a theory wrong by establishing facts that are inconsistent with it. However, evidence cannot prove a theory is absolutely correct. This is because new, undiscovered evidence might eventually emerge that contradicts the theory. Because of this, many scientists argue that true, infallible "scientific proof" does not actually exist.

Throughout the 20th century, different frameworks emerged to categorize evidence. In the 1950s, Rudolf Carnap suggested three distinct categories. These were classificatory evidence, which confirms a hypothesis; comparative evidence, which supports one hypothesis over another; and quantitative evidence, which measures the degree of support. Later, in 1990, biologist William Bechtel identified four specific factors used to settle scientific controversies. These factors include the clarity of the data, replication by other scientists, consistency with results from alternative methods, and consistency with plausible mechanical theories.

In 2001, philosopher Peter Achinstein published "The Book of Evidence," which organized evidence into four concepts. He identified epistemic-situation evidence, subjective evidence, veridical evidence, and potential evidence. Veridical evidence provides a good reason to believe a hypothesis is true. Achinstein argued that while scientists primarily seek veridical evidence, they also rely on other forms of probability. This work helped distinguish between the different ways researchers weigh the importance of their findings.

Evidence is not only used to test existing ideas; it is also used to create them. This process is sometimes called abduction. A famous example is the Geiger-Marsden experiment involving alpha particles and gold foil. The scientist Ernest Rutherford used the resulting data to calculate the mass and size of an atomic nucleus. He did not just test an old idea; he used the evidence to develop an entirely new atomic model. This shows that the relationship between evidence and science is an iterative process of competition between plausible rival hypotheses.