The Pirates Of Positivism And The Rascals Of

The Pirates of Positivism and the Rascals of Relativism

A guide to the Philosophy of Science for PhD students in Business.

Dr. Steven D’Alessandro Department of Business

Why are you here?• All PhDs in business usually follow a

“scientific approach”.• An understanding of the proof of

knowledge is useful in your argument and writing of your thesis.

• PhD examiners expect to see evidence of scholarship. This includes reason, logic, rationality and a scientific approach.

• Your long term development as a scholar and an active participant in society depends partially on an understanding of the philosophy of science

Why me?• I am an experienced PhD supervisor.

• Three successful PhD completions.• I have published widely as an academic.

• Three books, 50 papers with around 90 citations

• I have an interest in the philosophy of science.

• I believe that a course on the philosophy of science is vital for higher degree students as it teaches critical thinking.

What is this thing called science?• Explanation (laws, theories) based on

testable evidence independent of belief.• Science is not dogma nor authority.• Science is not practice, techniques, data or

empiricism. • Science should be logical, rational and

reasonable.

What is logic?• Scientific arguments to be testable must

be logical.• It is concerned with what follows what. e.g1. Many university lectures are boring2. This is a university lecture.3. This lecture is boring.

• Things can be logical but untrue.1. All cats have five legs2. Mr Darcy is my cat.3. Mr Darcy has five legs.

• Logic is important though as it tests the consistency of arguments.

More observations do not a science make.• Consider this example:• Demand increases when prices fall on

occasion t1

• Demand increases when prices fall on occasion t2

• Demand increases when prices fall on occasion t3

• Conclusion demand increases when prices fall.

• Possible that demand falls when prices fall.

The Inductive Christmas Turkey• The turkey notes

over a long time• “I am always fed at

9.00 a.m”.• The statement is

true till Christmas day.

Charting in finance is an example of this.

We need both logic and evidence to construct a scientific argument

Laws and theories

Predictions and

explanations

Facts acquired through

observation

Indu

cti

on

Deductio

n

Weber’s Law in Psychology and Marketing

∆I/I=K

A price increase of $1 will be more

readily apparent to a 50c item than an

$80 item.

Applies to visual brightness, lifted

weights, tone, smell and taste.

Indu

cti

on

Deductio

n

Problems with the classical scientific model• Induction and Deduction rely still on

interpretation of the facts.• Observation is theory dependent (more on

this later).• Must show how robust theories are. Are

they first falsifiable?• That is, are we sceptical enough to allow

for critical tests of theories that might prove them wrong (extreme examples or evidence).

Karl Popper and Falsification• Some statements are capable of being

tested (falsifiable).1. It never rains on Mondays.2. All substances expand when heated.3. Heavy objects such as your thesis when

released near the surface of the earth fall straight downwards if not impeded.

• Some statements are that are NOT capable of being tested (not-falsifiable).

4. Either it is raining or not raining.5. All points on a Euclidean circle are

equidistant from the centre.6. Luck is possible in financial speculation.

Karl Popper’s model of science

Non falsifiable and therefore psudeo-sciences includes Marxism and Freudism

Implications for your PhD thesis• Nothing is more practical than a good

theory.• You must explain and predict results.• You should allow for the situation that your

theory and data can be wrong.• Showing the robustness of results

(examining situations where the theory is unlikely to hold) strengthens your scientific argument.• Test competing structural equation models

across samples• You do not state beliefs you provide a

possible explanation of the evidence.

Does science progress by falsification?• Science would never had developed

perhaps if theories are rejected early before development.

• Theories can be rejected on minor or trivial grounds

• According to falisificationists all scientific knowledge is unproven till its disproven.

• How then does scientific knowledge develop?

• Perhaps it does by agreement and agrument of scientists.

Thomas Kuhn and the Structure of Scientific Revolutions (1970)

• Kuhn was a physicist at first.• Examined historical evidence of

the progress of science.• Science develops changes as a

result of social process similar to a revolution.

• Key stages are pre-science, normal-science, crisis, revolution, new normal science-new crisis.

• Normal scientists work within a paradigm (world-view).

Khun’s cogs of science

The critical stage is the choice of one paradigm over another• This about argument and convincing

people rather than evidence.• Each paradigm has different standards and

meta-physical principles.• Observation is paradigm /theory

dependent.• A scientist though may hold beliefs about

different paradigms (hedging bets).• There seems to be no place for objective

or better knowledge in his reasoning.

Implications for your thesis and PhD• You, your supervisor, department, parts of

the discipline all share paradigms of the world that may differ from examiners, journal reviewers, editors and ARC assessors.

• Therefore how well you convince, write and state your argument can be as important as the facts.

• Paradigms and world views change.• Economics. Keynes, Monetarism• Marketing. SEM, BWS, PLS. Qual/Quant.

• Even good evidence can be rejected by those from other paradigms.

Theories as research structures II: Research Programs.

• Imre Lakatos believes that science progresses by how well competing research programs confirm what was a novel prediction or explain new facts.• Newton’s laws and Halley’s comet.

• Science is progressive.• More difficult to do in the social sciences

where people react to theories ( finance, accounting, economics, psychology and marketing).

• A link to a recorded lecture on the topic is available at http://www.lse.ac.uk/collections/lakatos//

Implications for your thesis• You must show the superiority of your own

research program by comparing it with others.• Examine other schools of thought.

• The superiority is how well it predicts novel and new facts over competing programs.

• Examine competing theories/ programs where possible.

• Your research program represents scientific progress at a point in time. • Stating how you see your research program

advancing or not is important in your conclusion.

Feyrabend’s anarchistic theory of science.• Science has been given the status of

religion (dogma) when it comes to knowledge.

• Scientists will use a number of arguments to get their theories accepted.• Including propaganda and trickery.• Legitimate and illegitimate means.

• As long as knowledge works, anything (method) goes.• Voodoo and black magic are just as

good as science.• This is an issue of freedom!

Implications for your thesis• Non-scientific knowledge can be considered

as good as scientific knowledge. As long as it works!• Show competing non-scientific explanations.

• We should be open to new approaches, methods of learning.• Question the method you are using.

• Science and theories can become dogmas, thus reducing freedom (Marxism and Economics).• Keep an open mind to change.

• People can be made to fit the model.• Social implications of scientific knowledge

are important

The Bayesian Approach.• A theory has a probability of being correct

from 0 to 1.• Baye’s theorem

• P(h/e)=P(h).(Pe/h)/P(e).• Where• P(h/e) = Probability of the hypothesis (h) in

the light of the evidence (e).• P(h) = prior probability ascribed to the

hypothesis prior to the consideration of the evidence.

• P(e/h) = posterior probability after the evidence, e is taken into account.

Problems with Baye’s theorem.• A theorem not an explanation for the

development of scientific knowledge.• Still relies on assumptions and theory

dependent measures.• There are subjective and objectivist

Bayesians.• However Quantum Mechanics would seem

to follow a Bayesian approach. As does much statistical argument in business.

Implications for your thesis and PhD• Statistical evidence is only ever true at

one point of time, given the evidence and hypotheses considered.

• One never really proves the theory 100%.• Instead there is a possibility of calculating

the superiority of one theory over another.• No such thing as a bad theory, some are

better (have higher probabilities than others).

• Explanation (deduction and induction) are still crucial.

To sum up this is a debate about realism and ant-realism (relativism)• We are trapped in a language• Observation is theory dependent.• How does one know the truth?• Are there multiple truths?• Need to consider the existence of an

objective reality and prediction. • This more difficult to do in the social

sciences (business) than in the physical sciences.

• Consider the importance of falsification and competing research programs.

Implications for your PhD thesis• If you are interviewing respondents there

are multiple realities.• They are not passive data points.

• Be aware that you can never confirm or deny a theory just present evidence for and against it from your perspective.

• Have people outside your discipline read your thesis.

• Some theories (explanations) are better than others. You task is to find evidence for or against a set propositions in the most objective fashion.

• Even qualitative researchers should follow the scientific method.

Questions?

That concludes my talk.