Skip to content

Science uses observation to test models.

January 17, 2012

Sometimes scientists can’t actively probe nature; they just have to take what’s on offer.

This is one of ScienceOrNot’s Hallmarks of science. See them all here.

In short…

Where conditions can’t be controlled, scientists can gather data by observing clues to past events or phenomena occurring within unfolding events.

Scientists engaged in prototypical historical work … are primarily concerned with evaluating hypotheses about particular past events. They cannot reproduce these events in a lab. They can, however, look for present-day traces of them, and search for a smoking gun that unambiguously sets apart one hypothesis as the best among the currently available explanations for the traces thus far observed.

Carol E. Cleland, American philosopher, 2002.

When do scientists use observation?

Scientists use observation all the time. While carrying out tests, they continually depend on observations to gather data. But when they cannot use controlled experiments or trials to generate data in their testing, they have to look for whatever clues turn up.

When a model cannot be tested by experiments or trials, scientists make ‘predictions’ that do not depend on controlling conditions. The evidence they need to find may already exist (as in the case of fossils) or it may be expected to appear in the natural course of events (as in the testing of climate models). The predictions from the model should give scientists clues about what to look for and where to find it. Data are then gathered by direct observation.

How do scientists test without experiments or trials?

In this type of testing, scientists are like detectives solving a crime. They often start with a number of models or hypotheses, and look for clues that support one of these and/or eliminate others.

The search for clues may involve techniques such as field expeditions, exploration of the universe with telescopes, epidemiological surveys, or the monitoring of environmental conditions over time. In all of these cases, scientists are unable to modify variables, and must make use of whatever evidence they can find. They are overjoyed if they locate a ‘smoking gun’ – a clue that offers conclusive evidence. However, the preferred level of support for a model is always that multiple independent lines of evidence converge to a single conclusion.

Why this process is needed

Pure observational testing is needed for scientific models that cannot be tested by controlled experiments. This might be because they refer to things that happened in the past, or things that are happening far away in the universe, or phenomena/processes that are too vast or complex to control easily. Others involve humans or other animals, whose well-being cannot be affected by experimentation.


  • In 1705, Edmond Halley published his prediction that a comet seen in 1682 would return in late December 1758. To make this prediction he used the laws of motion recently described by Isaac Newton, and an elliptical model for the comet’s orbit. Halley’s Comet reappeared in early 1759, and the model has been re-confirmed every 76 years or so, when the comet makes its appearance in the sky. Historical records of earlier visits have also been found.
  • A classic ecological field study on the energy flow in ecosystems was reported by John M Teal in 1962.
  • In 1964, Arno Penzias and Robert Wilson detected the cosmic microwave background radiation. Although they did not realize it at the time, their discovery gave critical support to the Big Bang Theory.
  • The link between smoking and lung cancer was established, after much controversy, by a succession of epidemiological studies in the latter half of the 20th century.
  • The meteor impact model for the extinction of dinosaurs was proposed by Luis and Walter Alvarez in 1980. The initial clue was a rock layer enriched with iridium. In 1987, Bohor, Modreski and Foord reported global fractures in quartz crystals, which was consistent with an impact. In 1991, scientists identified what could be a ‘smoking gun’ – the giant Chicxulub crater in the Gulf of Mexico.
  • The reliability of climate models can be judged by comparing their predictions with measured data.

Further reading

Can Science Tell Us What Happened in the Past? Historical vs. Observational Science by Fallacy Man at The Logic of Science.

In defence of observational science: randomised experiments aren’t the only way to the truth by Kathryn Snow at The Conversation.

Carol E. Cleland’s quote is from “Methodological and Epistemic Differences between Historical Science and Experimental Science” , Philosophy of Science 69 (2002): 487
Photo credit: X-ray (NASA/CXC/MIT/D.Dewey et al. & NASA/CXC/SAO/J.DePasquale); Optical (NASA/STScI). It shows the remnant of a supernova.

This is one of ScienceOrNot’s Hallmarks of science. See them all here.

Reviewed: 2013/09/09

Leave a Comment

Be part of ScienceOrNot? Write a comment, make a suggestion or add an example!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: