Transcript scientific method

Research Methods in Science
UC LEADS
Summer 2003
Lecture 1
Research Methods in Science:
Outline of talk
• Overview of general principles of the
scientific method
• Philosophy of science
– examine objections
•
•
•
•
Bayesian and frequentist approach
Humanistic side of science
Ethics in science (case studies)
Scientific writing
What do you think the scientific
method is?
Elementary Scientific Method
•
•
•
•
Hypothesis formation
Hypothesis testing
Deductive and inductive logic
Controlled experiments, replication, and
repeatability
• Interaction between data and theory
• Limits to science’s domain
The Scientific Method
(“mission statement”)
• The scientific method is the process by
which scientists, collectively and over time,
endeavor to construct an accurate (that is,
reliable, consistent and non-arbitrary)
representation of the world
General principles that pervade
all of the sciences
•
•
•
•
Hypothesis generation and testing
Deductive and inductive logic
Parsimony
Science’s presuppositions, domains, and
limits
Hypothesis generation and testing
• Formulation of a hypothesis to explain a
phenomena
• “Educated guess”
• A hypothesis must be falsifiable
A hypothesis must be falsifiable
• The Loch Ness Monster is alive and well
• The Loch Ness Monster does not exist
• There is life on Mars
• There is no life on Mars
• DNA is the genetic material of all life
• DNA is not the genetic material
Hypothesis Generation and
Testing
Based on my (or someone else’s)
observations, I predict that:
•H0: no differences
•HA: significant difference
}
Treatments, controls,
independent &
dependent variables,
etc.
}
Let’s do an
Experimental Test!!
Treatments, controls,
independent &
dependent variables,
etc.
Experimental Tests: What are the
main features?
• Clear hypothesis
• Identify independent and dependent
variables
• Assign controls
• Repeatable, hence verifiable results
• Used to support or refute claims
General principles that pervade
all of the sciences
• Hypothesis generation and testing
• Deductive and inductive logic
Deductive and Inductive Logic
(distinction #1)
• The conclusion of a deductive argument is
already contained implicitly in its premises
• The conclusion of an inductive argument
goes beyond the information in its premises
Deductive and Inductive Logic
(distinction #2)
• Given the truth of all of its premises, the
truth of an inductive argument’s conclusion
follows with at most high probability
• Deduction argues from a given model’s
general principles to specific cases of
expected data
Deductive and Inductive Logic
(distinction #3)
• Deduction argues from a given model’s
general principles to specific cases of
expected data
• Induction argues in the opposite direction,
from actual data to an inferred model
Deductive and Inductive Logic
• One is based on statistics (inductive)
• The other is based on probability
Deductive and Inductive Logic
(telling the difference)
Given: A “fair coin” is one that gives tails
with probability 0.5 and head 0.5
.
• Problem 1: Given that a coin is a fair coin. What is
the probability that the coin will produce 45 heads
and 55 tails?
• Problem 2: Given that 100 tosses of a coin
produce 45 heads and 55 tails. What is the
probability that the coin is a fair coin?
Why is induction so pervasive
and critical in science?
Science is almost entirely about
unobservables -- about things and times
outside the database of actual observations.
Iron melts at 1,535°C (but everywhere?)
Water boils at 100°C (but everywhere?)
The basis of induction: Aristotle
• Aristotle (384-322 BC) offered 3 methods
of induction
• Unifying concept: in deductive arguments,
which are composed of premises, inductive
arguments are the scaffolds that raise the
status of the deductive argument to a lawlike status
The basis of induction: Aristotle
• Dialectical induction (Topics). Not entirely
relevant to scientific research, but useful:
– mentor to pupil discourse
– “If a skilled pilot is the best pilot and the skilled
charioteer is the best charioteer, then, in
general, the skilled [person] is the best [person]
in any particular sphere” (Perez-Ramos 1988)
The basis of induction: Aristotle
• Enumerative induction (Prior Analytics).
Statements about individual objects provide
the basis or premises for a general
conclusion:
– from observing numerous adult humans, an
inductive argument could conclude that all
humans have 32 teeth
The basis of induction: Aristotle
• Intuitive induction (Posterior Analytics).
Direct intuition of the general principles
exemplified in the data:
– bright side of the mood always faces the sun, so
the moon shines because of reflected sunlight
General principles that pervade
all of the sciences
• Hypothesis generation and testing
• Deductive and inductive logic
• Parsimony
Parsimony
• Shortest path or the less complex
“explanation” to the “true state of nature”
A
B
Parsimony
• Keynes (1962) expressed parsimony as the
law of the limited variety in nature
– Iron melts at 1,535°C
– unlimited nature…unique atoms…unique
properties…no iron, oxygen, no humans (sum
of the parts)
– 100 chemical elements
– related presuppositions of induction
Parsimony
• The principle of parsimony recommends
that from among theories fitting the data
equally well, scientists choose the simplest
theory.
• Thus, the fit of the data is not the only
criterion bearing on the theory choice
Parsimony
•
•
•
•
•
Additional criteria includes:
predictive accuracy
explanatory power
testability
fruitfulness in generating new insights and
knowledge coherent with other scientific
and philosophical beliefs
• repeatability of results
Parsimony
• Q: Why is parsimony an important principle
in science?
.
• A1: The entire scientific enterprise has
never produced, and never will produce, a
single conclusion without invoking
parsimony
• A2: Economy…facilitate insight, improve
accuracy, and increase efficiency
General principles that pervade
all of the sciences
•
•
•
•
Hypothesis generation and testing
Deductive and inductive logic
Parsimony
Science’s presuppositions, domains, and
limits
Science’s presuppositions,
domains, and limits
• Set of beliefs that allow a person to validate
her observations, results, conclusions
(objectivity of science)
– constancy of the universe
– parsimony
• Acceptance and acknowledgement of the
knowable and the unknowable
General principles that pervade
all of the sciences
•
•
•
•
Hypothesis generation and testing
Deductive and inductive logic
Parsimony
Science’s presuppositions, domains, and
limits
General principles that pervade
all of the sciences
•
•
•
•
Hypothesis generation and testing
Deductive and inductive logic
Parsimony
Science’s presuppositions, domains, and
limits…
How do we represent this set of principles
that found in all of the sciences?
General principles that pervade
all of the sciences
•
•
•
•
Hypothesis generation and testing
Deductive and inductive logic
Parsimony
Science’s presuppositions,
domains, and limits
}
SCIENTIFIC
METHOD
General principles that pervade
all of the sciences
There are detractors of the
idea that a scientific method,
upon which we are able to
make claims about the true
state of nature, does not exist
}
SCIENTIFIC
METHOD
General principles that pervade
all of the sciences
There are detractors of the
idea that a scientific method,
upon which we are able to
make claims about the true
state of nature, does not exist
}
Philosophical
&
Scientific
cannot
General principles that pervade
all of the sciences
Paul Feyerabend insisted that
there are no objective
standards of rationality, so
naturally there is no logic or
method to science…“anything
goes” in science…it is no more
productive of truth than
“ancient myth-tellers,
troubadours and court jesters”
Philosophical
cannot
General principles that pervade
all of the sciences
Thomas Kuhn is critical of what he
sees as modernist misrepresentation
of the nature of science:
Modernist definitions of science claim
that science is objective because it is
empirical (based only on the data of
our senses), rational (reasonable, or
logically defensible) and that its
presuppositions are obviously true...
Scientific
cannot
General principles that pervade
all of the sciences
Kuhn claims science is a social
enterprise and as such is also
quite subjective. He argues that,
"every individual choice between
competing theories depends on a
mixture of objective and
subjective factors."
Scientific
cannot
General principles that pervade
all of the sciences
Instead, science occurs in
revolutions where old ideas are
thrown out and new ones
accepted. Science is therefore
capricious, and each discipline of
science cannot share a set of
pervading principles
Scientific
cannot
General principles that pervade
all of the sciences
These revolutions are called
PARADIGM SHIFTS
Scientific
astronomy
geology
chemistry
physics
biology
astronomy
geology
chemistry
physics
biology
General principles and technologies are distinct to each
scientific discipline
Thought experiment
• You have been awarded a $500,000 grant
and can spend it on any type of equipment
that is relevant to your research.
• Make a list of what you will buy and justify
it (don’t worry about EXACT price values
as you essentially can afford almost
anything!)
• (don’t forget about Gregorio’s research!)
Thought experiment
• Can you safely say that you will not rely on or
utilize any of the following principles by using
your new equipment?:
hypothesis generation and testing
Deductive and inductive logic
Parsimony
Science’s presuppositions, domains, and limits
“you”
“them”
Pervasive in
all sciences
Based on Greek
philosophers &
many others
Non-negotiable
presuppositions
of perception:
“you see what I
see…you feel
as I feel”
}
Unique to our fields but after
the same thing...
}
Common
to all
of us
The Scientific Method
(“mission statement”)
• The scientific method is the process by
which scientists, collectively and over time,
endeavor to construct an accurate (that is,
reliable, consistent and non-arbitrary)
representation of the world
Bayesian and Frequentist
Approach to Scientific Research
• Bayesian Statistics have been developed for
a variety of purposes, such as designing
experiments, estimating the values of
quantities of interest, and testing hypothesis
• Useful because this family of statistics takes
into account prior results as opposed to
assigning independence to each result,
thereby introducing efficiency
Bayesian and Frequentist
Approach to Scientific Research
• For a loaded dice (biased for “6”):
• The frequentist views dice throws as
independent events, each number or face
having an equal probability: each value has a
1/6 probability of appearing.
• The Bayesian, the probability of getting a “6”
will be more than just 1/6 (as will the
probability of being thrown out on your ear!)
Bayesian Approach to Scientific
Research
• The search for patterns in data will be “more
realistic” as you do not discard “prior”
knowledge -- it helps you get to the “answer”
much faster
• Calculations are not very difficult for small
sample sizes, but can get complicated for
large ones…let’s see an example:
Bayesian Example
• Coin toss determines the configuration of the
marbles that go into an opaque urn:
• heads: place 1 white + 3 blue marbles
(WBBB)
• tails: place 3 white + 1 marbles blue
(WWWB)
• Only “coin-tosser” knows
Bayesian Example
Ratio of the likelihood of “heads”
to the likelihood of “tails”
Number of
draws
Posterior
probability
Bayesian Approach to Scientific
Research
• Your confidence in the results (and hence
your hypothesis) increases tremendously with
each draw of a marble
• If trials are expensive then using likelihood
values are important
• Can be computationally complex (trade off)
The Humanistic side of Science
• Your perceptions of the humanistic side of
science:
• It can lie between one’s research and one’s
beliefs
• It may not be realized at the outset
• It may change during your career
• You may not want them to intermingle
Science as a Liberal Art
• The search for and the advancement of
knowledge and truth is a common goal
among scientists
• The “truth” will (hopefully) be used to
improve the world in which we live in
• The “truth” will be used for just and moral
purposes
Science as a Liberal Art
• As scientist, we may be in dilemmas that
will challenge out personal beliefs
• A strong conviction in what one believes
should reflect the kind of work one
undertakes
• May or may not reflect current social
climate
Science as a Liberal Art
Examples of controversial research:
• stem cell research
• genetic engineering / GM food
• nuclear sciences
• control systems (used by the defense)
• biological control
• alternative fuel research
Science as a Liberal Art
Do you have ethical “boundaries” that you
have considered in your work?
Ethics Case Studies
•
•
•
•
•
•
Isa and Senait will lead discussion
Introduce the paper
Break into groups
Read and discuss paper
Develop topics for big discussion
Introduce second ethics issue (no break-out
groups)
Ethics in Research
• Reporting of data accurately is seen not
only as a high professional quality, but also
a moral one.
• Why?
Ethics in Research
• Ethical researchers do not plagiarize or claim credit for the
results of others;
• They do not misrepresent sources or invent results;
• They do not submit data whose accuracy they have reason
to question, unless they raise the question;
• They do not conceal objections that they cannot rebut;
• They do not caricature or distort opposing views;
• They do not destroy or conceal sources and data important
for those to follow
Research Ethics and Science
Writing: Example
Research Ethics and Science
Writing: Example
Research Ethics and Science
Writing: Example
Research Ethics and Science
Writing: Example
Research Ethics and Science
Writing: Example
Research Ethics and Science Writing: Example
Research Ethics and Science Writing: Example
Concept Map:
Water Example