Science and

Scientific Reasoning

Critical Thinking
Some Common Myths
About Science
Science:
What it is and what it is not
 Science and Technology
• Science is not the same as technology
– The goal of science is to come to an understanding of
the world around us. The products of science are
theories about nature.
– The goal of technology is to help us accomplish
something. The products of technology are tools.
• Science and technology have a close relationship:
– Technology applies the knowledge given by science.
– Scientific research may be driven by technological
needs or desires.
• ‘Popular Science’ magazine should really be
called ‘Popular Technological Gadgets’
 Science and Scientific Theories
• Science is not a set of scientific theories
– Science is like philosophy (in fact, science grew out of
that branch of philosophy that studied nature and was
called “natural philosophy” until mid-19th century): it is
a method, study, investigation, or inquiry.
– Science can’t be wrong, as science makes no claims.
– Scientific theories are like philosophies: they are sets of
beliefs. Scientific theories can be wrong.
• Science is the process of figuring out which
theories are more likely than others.
– Who refutes scientific theories? Science itself!
 Science and Scientific
Experiments
• Science is not doing scientific experiments
using multi-million dollar equipment
– First of all, many scientific experiments can be
done without any equipment at all.
– Second, doing scientific experiments is only
that part of science where a scientific theory is
actually being tested. Scientists first have to
come up with a theory to be tested, have to
think of a useful test, and have to draw
conclusions from the results of that test.
• Many ‘scientists’ are really ‘technicians’.
 Science and Math
• Science is not a bunch of equations.
• To many people: science = numbers = truth
• Just because we are dealing with a number,
doesn’t mean that that number is accurate
(and remember: precision is not the same as
accuracy).
• And, again, science is so much more than
just numbers and equations: it is the process
that provided those numbers and equations!
 Science and Truth
• Finally, science does not have a stronghold
on the truth.
– In fact, it is quite the opposite. Scientists are
critical of every theory that they propose: that is
why they are constantly being tested. Indeed,
this inherent self-reflection, self-modification,
and self-correction is what really defines
science.
• Science is critical thinking!
 ‘Hard’ Science and
‘Soft’ Science
• Fields like physics and chemistry are often referred to as
the ‘hard’ sciences, where something like psychology or
sociology are ‘soft sciences’.
• Moreover, the ‘hard’ sciences are often seen as ‘more’
scientific than the ‘soft’ sciences
– Indeed, physics is often seen as the ‘epitome’ of science.
• Now, it is certainly true that the results from physics and
chemistry are more quantitative and allow for more
reliable explanations and predictions … but that doesn’t
make them more scientific!
– That merely reflects the ‘science = math’ and ‘science = truth’
mindset.
 The Scientific Method

Understanding How Science Works

and What It All Means
 The Scientific Method

Puzzling Explanation/
Phenomenon Hypothesis/ Creative
Theory

Evaluative
Experiment/ Prediction
Test

Real Speculative
 Puzzling Phenomenon
• Many people see science as something that
collects data and generalizes from that.
However, science doesn’t collect data for
the heck of it. Rather, science starts at the
moment that we observe something that
puzzles us: something for which we want an
explanation.
 Hypothesis
• The second step of the scientific process is to
come up with a possible explanation for the
puzzling phenomenon. This explanation is also
called a hypothesis or theory.
• Obviously, every proposed theory should explain
that what we want to have explained. However,
that is not enough to be considered a good theory.
Thus we come to the testing phase: prediction and
experiment.
 Prediction
• In order to test a theory, you need to make a
prediction based on the assumed truth of that
theory. Thus, in the third step, you consider what
would have to be the case if the theory were true.
• The prediction should be novel in that it should
predict something that we don’t already know.
• A prediction should also be as specific/precise as
possible.
Eliminating Alternative Theories:
The Surprise Principle
• The strength of any theory depends largely on how
many alternative theories there are: the fewer
alternative theories, the stronger the proposed
theory. A good prediction is therefore one that is
predicted by only few alternative theories, because
that way, if the prediction comes out true, we
would have strong evidence for the proposed
theory, Thus, the more ‘surprising’ it would be
that the prediction would come out true, the better.
Science can therefore be seen as a process of
eliminating bad theories.
 Experiment
• The experiment is where we see if the prediction
comes out true. That is, predictions often take the
form “If you were to do this, then that would
happen”. In the experiment, we do exactly ‘this’,
and see if ‘that’ indeed happens.
• Sometimes the ‘experiment’ consists of a single
and simple observation. Again, much of science
does not need multi-million $ equipment.
 Evidence for a Theory
• A theory is confirmed by an experiment if
the prediction comes out true.
– Confirmation doesn’t mean proof, because
other theories could have made the same
prediction. Thus, we call this evidence.
– Further testing is needed to choose between
competing explanations.
 Evidence against a Theory
• A theory is disconfirmed by an experiment if the
prediction comes out false.
– Disconfirmation does not necessarily mean that the
theory is false, as the prediction often relies on
auxiliary hypotheses. Example: Ships disappearing at
the horizon only disproves the Flat Earth Hypothesis if
light travels in a straight line.
– Disconfirmation of a theory can lead to either rejection
or modification of that theory. In fact, even if a theory
is false, it could still represent a useful approximation.
Example: Newton’s laws.
 Scientific Reasoning
• When doing science, we use all kinds of
reasoning. For example, when we make a
prediction, we probably use deductive reasoning.
And when we generalize from certain data, we use
inductive reasoning. However, the core reasoning
in science is called abductive reasoning (or
hypothetical reasoning): we reason what, given the
facts, would best explain those facts. Sometimes
scientific reasoning is therefore called “inference
to the best explanation”.
 Data and Theories
• Scientific theories are driven by data, not vice
versa. That is, science will adjust its theory
according to the available data, rather than trying
to find data that is compatible with a given theory.
To put it yet another way: the theory should fit the
data, rather than that the data should fit the theory.
Again, science tries to make a theory by trying to
break it, rather than by selectively picking and
choosing data to confirm whatever theory at hand.
 Scientific ‘Proof’
• Scientists are critical of their own theories,
because they know that every proposed theory can
be false: no amount of evidence can prove a
theory with 100 percent certainty.
• So, what then does it mean for a theory to be
“scientifically proven”? It means that there is so
much evidence for that theory that the truth of that
theory is put beyond “reasonable” doubt.
 Proof: from L. probare "to test”
• Using predictions, we put a theory to the test. And, we put the theory
to the test as hard as we can by considering the most unlikely
predictions. This critical process is what used to be called ‘proving a
theory’. That is, when you prove a theory, you are very critical of it,
and put some strain on it. The saying “the exception proves the rule”
uses this old meaning of the word ‘proof’.
• Nowadays, ‘proof’ means something that supports our theory, rather
than something that puts strain on it: almost the opposite! (although
you can think about it this way: by proving (testing) a theory, we may
end up giving support for it (if the theory passes the proof of course!)).
• This change in meaning is rather unfortunate, because while the goal
of science is to come up with true theories, its method is to put as
much strain on a theory as possible. So yes, science is proving theories,
but really in the old sense of the word!
 Provisional Acceptance
• Q: Given that science can’t conclusive prove anything,
why should I accept any kind of scientific theory?
• A: Am I certain to die if I jump off a cliff? No, but I
wouldn’t count on the possibility of not dying! The same
goes for scientific theories: as long as there is evidence for
them, you have reason to believe that they are true rather
than false. And remember, you should never blindly accept
any theory anyway. You should believe it to the degree
that there is evidence for it. Thus, it is ok to accept
scientific theories, as long as you do so provisionally, and
are ready to change it for a better theory if the evidence
points in a different direction.
• Note: Sometimes people say: “It’s just a theory”, implying
that that particular theory can be ignored, or at least its
importance downplayed. Again however, while no theory
is conclusive, this does not mean that we can ignore them.
 Criteria of Adequacy

What Makes for Good Scientific

Theories, Hypotheses, and
Explanations
 Criteria of Adequacy
• The purpose of theories, hypotheses, and
explanations in general, is to help us understand
something and (thus) to enable to help us make
reliable predictions about that something.
• The criteria of adequacy help us evaluate whether
a suggested explanation has the potential to fulfill
this purpose.
• An explanation that satisfies all criteria of
adequacy is called a proper explanation.
 Criteria of Adequacy and
Fallacies
• There is an interesting parallel between criteria of
adequacy and fallacies:
– A proper argument does not commit any fallacy, but an
argument that does not commit any fallacy may still be
an argument with a false conclusion.
– A proper explanation does not violate any criteria of
adequacy; but an explanation that does not violate any
criteria may still be wrong.
• Thus: Criteria of adequacy are the minimal
properties that a good explanation should have.
 Relevance
• At the very least, explanations should be a
possible explanation of that what it is that it
is supposed to explain.
• Explanandum: that which needs to be
explained
• Explanans: the suggested explanation
• The explanandum should follow from the
explanans.
 Non-Circularity
• Explanations should not be circular. A circular
explanation merely restates (using different
words) what needs to be explained.
• Circular explanations don’t increase our
understanding; they don’t provide us with any
insights into what we want to understand.
• Virtus Dormitiva: “Opium puts you to sleep, since
opium has sleep-inducing powers.”
 Testability / Refutability
• Theories should be testable …
– If a theory cannot be tested, then we can’t confirm it.
– So, we can’t give reasons to believe its truth either.
– Hence, we can’t justify that the explanation increases
our understanding or predictive powers.
– Thus, any untestable theory is a bad theory.
• Therefore, theories should be refutable!
– There must be a possible way of refuting the theory, i.e.
we should be able to disconfirm the hypothesis.
– For if we can’t disconfirm the hypothesis, then we can’t
possibly confirm the hypothesis either.
 Generality / Simplicity
• A sign of a good theory is that it is general:
– It can explain many things
– Even better: it can predict many things (avoids
after-the-fact story-telling and fact-fitting)
– It is not ‘ad hoc’ (“for this purpose only”)
• Generality goes with simplicity:
– Postulates few basic principles or entities
– Occam’s Razor: Do not postulate the existence
of extra laws or entities if they are not needed.
 Precision
• Theories should be precise:
– If a theory is not precise, then possibly no
predictions can be made.
– If a theory is not precise, then maybe only
vague predictions can be made, such that the
outcome of any experiment can always be seen
to be compatible with the prediction.
• In other words, precision helps us put
theories to the test.
 Conservatism
• New theories should be compatible with
existing well-established theories.
• Being conservative makes sense:
– Evidence that is found in favor of a well-
established theory that a new theory conflicts
with is likely to be evidence against that new
theory.
– Conservatism avoids postulating new entities
and principles, and thus it favors simplicity.
Errors in Science:
The Human Factor

And How Science

Has Evolved
to Correct For This
 Science is Fallible
• 1. Science is conducted by humans
• 2. Humans are fallible
• 3. Therefore, science is fallible

• Note: the fact that science cannot prove theories

with 100% certainty is not part of its fallibility:
that merely shows the inherent limitations of any
method that tries to gain knowledge. Even the
Perfect Rationalizer would not be able to decide
between two competing theories, if both fit the
available data equally well.
 What can go wrong?
• Mechanical Errors
• Physiological Errors
• Psychological Errors
– Reasoning Errors
– Perceptual Errors
– Interpretational Errors
– Theoretical Immunity Errors
– Other Psychological Errors
• Sociological Errors
 Mechanical Errors
• Anything that goes wrong in the mechanical
execution of the scientific method.
– Miscalculation
– Misreading
– Misrecording
– Error in experiment (dirty beaker, wrong voltage, etc.)
• These errors are rather innocent. Repetition of
experiment will usually correct for these kinds of
errors. Better technical equipment helps too.
 Physiological Errors
• Errors due to physiological factors:
– Tired
– Distracted
– Drugged
– Sick
– Etc.
• Science can try and correct for these errors by:
– Repetition
– Mechanization
 Reasoning Errors I
• Hasty Generalization
– Relying on too little data
– Relying on biased data
• Scientists try to correct for this:
– Gather as much data as possible
– Gather data under controlled settings
– Do not accept anecdotal evidence and evidence from
personal experience
 Reasoning Errors II
• Errors in causal reasoning:
– Confusing correlations with causations: A correlation
between A and B can mean any of the following:
• A causes B (what we want to conclude)
• B causes A (reverse causation)
• C causes A and B (common cause)
• Nothing at all (it was just coincidence)
– Post Hoc reasoning (‘after that, hence because of that’)
– Bad sense of probabilities
• To correct for this, scientists use statistical
methods and controlled experiments. Also, causal
claims will be made with much caution.
 Perceptual Errors
• Our perceptual apparatus is not like a camera:
– Selective Perception (Selective Attention): We notice
certain things, but don’t notice others
– Constructive Perception: Our brain fills in missing
details and sometimes changes details to make the
perception fit our expectations
• Memory is selective and constructive as well!
• To correct for this, scientists use mechanical
instruments to measure and record
– Obviously, this is also because sometimes the human
brain is simply too limited to make the necessary
observation or to remember all the data
 Interpretational Errors
• Science is not conducted in a vacuum. We are
doing it with certain preconceptions in place that
effect our interpretations, and even our
observations:
– Setting of the sun
– Slide in biology class
– Sensitivity of equipment determines what is observed
• This is very hard to correct for:
– Have other scientists scrutinize work
– Always keep an open mind and try and come up with
alternative theories and views
 Theoretical Immunity Errors
• All humans are reluctant to be self-critical
• Scientists are humans
• Therefore, scientists are reluctant to be critical of
their own views and theories.

• Through rationalization, self-deception,

confirmation bias, and other means, scientists can
hang on to ‘their’ pet theory.
• Combined with the fact that we have basic
conceptual limitations, theories often change
through paradigm shifts: conceptual revolutions
 Other Sources of Psychological
Errors
• Herd Instinct
• Wishful Thinking
• Striving for Certainty
• Striving for Simplicity
 Sociological Errors
• There are many outside, sociological forces
acting on scientists as well:
– With a new theory can come fame and fortune
– Scientists need to produce to remain at position
• These forces can effect the reliability of the
science in all of the aforementioned ways:
– Trying to find something, and actually ‘finding’
it, without it being there.
 The Evolution of Science
• A final myth about science: science and the
scientific method have been the same for several
hundreds, if not a couple of thousand, years.
• In fact, it is exactly by becoming aware of the
preceding pitfalls of doing science (especially
cognitive biases) that science has evolved:
– white labcoats, notebooks, dedicated and calibrated
equipment, peer review, standard operating procedures,
repetition, statistical analysis, blind and double-blind
controlled experiments: these are all things that only
gradually have become part of science as we know it.