Week 15 - Psychology and science Flashcards
What is ‘science’?
- “A systematic enterprise that builds and organises knowledge in the form of testable explanations and predictions about the universe” -Wikipedia
-> need to be able to test our predictions about the universe using knowledge that we have + that knowledge builds upon prior knowledge that has been tested - The word ‘science’ is derived from the Latin word “scientia” which means knowledge based on demonstrable and reproducible data
What is the scientific method?
- First create a hypothesis (hopefully based on existing data/theory)
- Test hypothesis using observation
-> experiments can be used to infer causality - manipulation of A leads to B
-> other observational methods (e.g. correlations) will not allow for causality to be established. But the research might not be mature enough to infer and test for causal relationships - they might simply want to establish whether there is a relationship between A and B
-> Might have an intuition but without testing it we have no solid foundation to believe it - Does the data support the hypothesis or does it not?
- If possible integrate findings into a theory - either showing it can or cannot account for the findings
How can we be certain that this process (the scientific method) works?
- Need to think carefully about study design + eliminating bias
-> control + elimination of extraneous variables and noise influencing results - Need to be clear about what was done and why
-> exactly the methods you use + the controls you put in place and why this was important - Peer review process
-> other experts read and check to ensure it is valid
What are quantitative methodologies?
- Typically concerned with collecting numeric data through experimentation or some other form (e.g. psychometrics)
- Hypotheses are tested by using statistical analysis
- Statistical analysis allows us to establish whether the findings (e.g does changing X lead to Y, or is there a relationship with X and Y) can be generalisable
- How? Because of the underlying assumptions of statistical tests
- But, sometimes data are not in the format that allow statistical analyses
What are qualitative methodologies?
- Related to a range of philosophical approaches
- Conversation analysis - most aligned to positivism (scientific verification) because the data analysis is about the sequential aspects of the speech to understand the meaning of it -BOTTOM UP (the data speak)
- Other approaches are much more the analysts interpretation of the data - much more TOP DOWN
- Some do not believe in reading the literature before carrying out an interview - the theory and model is derived from the data, no pre-existing hypothesis
- Many qualitative approaches do not believe in generalisation - it is about understanding the meaning of the data in relation to a particular cohort or particular phenomenon
- Some people think these methods are important to provide a rich understanding of what theories can be developed and tested (mixed methods approach)
-> do a qualitative approach to begin with, interview ppts + get an idea about subject, then do a quantitative approach which is informed by the info you receive in the initial approach - The approach you use depends on your aim
How does science develop?
Popper - science as reduction
-> he favoured empirical falsification
- You can never prove whether a theory is true, but you can prove that a theory is false - concept of falsification
-> should be scrutinised with experiments, if no evidence to support a theory, that theory can be falsified. But you can never have data to prove the theory is true
What is poppers view on the scientific method?
1) A scientific theory must be falsifiable -> must be possible to conceive of an observation that would contradict it again
2) A scientific theory is not considered to be confirmed by a single positive test, but by its ability to withstand repeated attempts to falsify it
-> best scientific are those that have been subject to the most rigorous tests and have not yet been falsified
Examples of falsification
Theories need to be falsifiable
- Darwin’s theory of evolution
-> any findings of human bones dated at the same time as dinosaurs would falsify this theory
- Have different races developed in isolation?
-> we all have a single genetic marker that shows we are all descended from Mitochondrial Eve, a woman about 200,000 years ago
-> seems to support the “Out of Africa” theory of population spread from one location around 100,000 years ago, all around the world as we exist today
-> Just 1 DNA result (in 8 billion humans) not showing this genetic marker would falsify this theory
Why do we need theories - Lewandowsky & Farrell, 2011: Mars retrograde motion
- You cannot describe or explain this motion of Mars unless you have a theory of the underlying process
-> what needs to happen to Mars to take this pattern in the night sky? - A theory that explains data like this is itself unobservable
-> it cannot be observed (e.g. solar system)
-> the theory is an abstract explanatory tool
-> it exists in the mind of people who use it to explain data
What are alternative models that have been presented to explain similar phenomena?
Copernicus
-> retrograde motion is a result of planets moving at different speeds along their orbits. It occurs when the Earth overtakes Mars
Ptolemy’ Geometric model (predominant model for 1300 years)
-> planets go round the Earth - Earth is centre of universe
-> while orbiting around the Earth, each planet also circles around a point on the orbit
-> if you slightly offset the Earth from the centre, this provides a very good account for the data
Why did the Copernicus model replace the Geometric model?
- The fit of the data was only slightly better
- Mathematically, the predictions of the heliocentric model (Copernicus model) provided a better account of the observed data than the geometric model
e.g. it could explain why the planets appear to move backward at certain times of the year + why the apparent sizes of the planets change as they orbit the sun
BUT - It was a simpler model. It was more simple and elegant
- For the geometric model explanations are needed about why the Earth is slightly off centre of the orbit and about why planets circle round the point on the orbit
Why do we need theories?
- Theories allow us to make predictions or account for data
- We have a preference for simple theories
Occams Razor
- Plurality is not to be posited without necessity
-> when evaluating competing hypotheses, the one with the fewest assumptions should be selected
- Related to falsification. If our theory does not capture the data there are an infinite number of “add-ons” we could make, one which would make the theory work
-> this makes falsification impossible
-> concept is called parsimony
What is the function of theorising in psychology?
Can we explain everything?
- No, we’ll always come up against limits
- These limits might appear relatively quickly (psychology is a young science)
- But theories are helpful in building up our knowledge
Estes (1975)
-> its important to work out what can, and what cannot be explained
-> classification of what a theory can and cannot explain is an important part of developing a theory
How robust is the scientific method?
- Every study is wrong to some extent -> all have weaknesses which limits the ability
-> specific sample - much psychology is based on WEIRD samples (Western. Educated, industrialised, Rich, Democracies)
-> tests a theory in a very specific context
-> validity
-> reliability - the role of chance
Scientific publication is a human endeavour + with anything human there are going to be errors
-> checks and balances to avoid accepting dodgy research may not always work
-> the publication process itself may distort what is found
-> can be biased by dominant views
What is the structure of scientific revolutions? - Thomas Kuhn
- In a field of enquiry there may be a dominant paradigm
Paradigm definition ->a typical example or pattern of something; a pattern or model - In a field of enquiry - a dominant way of doing and thinking about what is being explored
- These come with their own set of tools and ways of measuring things
All fields have a dominant paradigm and this might shift overtime
- Within a paradigm, over time we repeatedly try and refine our tools and measures to provide explanatory power
- Ontology, methodology, and epistemology combine to form a research paradigm
- However, bit by bit we might begin to build a picture that tells us our framework for understanding and approaching a phenomenon is wrong
- A new idea is put forward and begins to gain traction
- At this point - a scientific revolution and paradigm shift
- The new paradigm might have a whole new set of tools and ideas and ways of measuring things with it that make it incomparable to the previous paradigm
- We think of science being measured against a common set of standards to decide what idea/explanation is better
- This is not quite true - scientists are a community with ideas and some are hard to shake off
What is an example of a revolution in psychology?
Cognitive revolution
Prior to the cognitive revolution in the 1960s, the dominant paradigm was behaviourism
Behaviourism:
->psychology can only be a science if it is based on observable behaviour (influenced by philosophical preoccupations about what is a science)
-> as such should not investigate unobservable mental processes
Cognitive psychology - you CAN hypothesise the existence of inner mental states if you can think of ways to test them
- Cognitive psychology and behaviourism are not as diametrically opposed as they seem and are linked in various ways. Behaviourists didn’t believe there was not an inner mental state, they just didn’t want to study it
What is a smaller example of a revolution in psychology?
Autism
-> seen as a general developmental disorder. Known issues around verbal and non-verbal communication
Baron-Cohen (1985) - Does thee autistic child have a theory of mind?
-> Reconceptualised autism as a cognitive issue about conceptual perspective taking
-> this means totally new ways of thinking about how to explore and test it. So new methods + new measures
