Laboratory experiments

These experiments:

1. Use analytic controls/contrasts to assess alternative explanations

2. Can study variables in isolation

3. A closed system, no other variables present except those intended

4. Use intervention (manipulation) to infer cause and effect

5. Ability to make abstractions about general principles

6. Create situations which may not occur naturally

Scientific method

A systematic, empirical approach to research that involves observation, experimentation, and the formulation and testing of hypotheses to draw conclusions about phenomena.

Relevant facts

Relevant because they inform us about our abstractions.

Irrelevant facts

Irrelevant because:

1. Inadequate protection against interference

2. Inadequate methods

3. Came from outdated or ineffective experiments

4. Discredited in some way

5. No impact on our abstractions

Qualities of experimental results

• Fallible

• Reversible

Also true for all knowledge based upon it.

Abstract concept

Intangible ideas, qualities, or relationships that exist in mental representations. General and may not relate to physical reality.

Concrete concept

Tangible experiences which can be perceived with the senses. Specific and directly relate to physical realities.

Aim of science

To produce and extend reliabke knowledge of events and objects in the world.

• Explain

• Understand

• Predict

• Control

Logic of science

There is a real world → the real world is observable → therefore observational statements are truthful about the real world.

The same is true for other knowledge systems (e.g. common sense, religion, art).

Norms of scientific conduct

• Universalism

• Communalism

• Disinterestedness

• Organised skepticism

Universalism

The acceptance or rejection of claims entering science does not depend on the personal or social attributes of the person stating them.

A norm of scientific conduct.

Communalism

The substantive findings of science are a product of social collaboration and are assigned to the community.

A norm of scientific conduct.

Disinterestedness

Findings are scrutinised, and accepted or rejected without personal or organisational prejudice.

A norm of scientific conduct.

Organised skepticism

Judgement is suspended until the facts are made available.

A norm of scientific conduct.

Assumptions of science

1. There are knowable uniformities in the world.

2. Uniformiies can be explained through causes and effects.

3. Knowledge is obtained via inferences:

   1. From general to specific (deduction).

   2. From specific to general (induction).

4. Experiments are the hallmark of this field.

5. Cumulative progress.

6. Occurs independently from the rest of the world.

7. The best avenue to reliable and objective knowledge.

Caveats about assumptions

• Be aware of what is taken for granted.

• Be aware of what goes unsaid.

• Some assumptions may be difficult/impossible to demonstrate or defend.

Characteristics of scientific knowledge

• Abstract and general - applies across time, place, and people

• Deals with facts but goes beyond them

• Analytic - decomposes things into parts

• Verifiable

• Communicable

• Lawful

Induction

The process of generating laws and principles from a finite set of observables.

Movement from fact to theory.

How to justify an inductive inference

Three conditions must be met:

1. Must be a large number of observations forming the basis of the generalisation.

2. Observations must be repeated under a wide variety of conditions.

3. No accepted observation statement should conflict with the derived law.

Scientific progress

The accumulation of increasingly more inductively-generated laws.

• Made through trial and error, conjectures and refutations, refinement over time

• While a theory can never be proven true, it can be said to be the best currently available over other theories prior

Deduction

Once a law has been generated, it can be assumed that all future cases can be subsumed under them - a conclusion is true if the premises are true.

Movement from theory to fact - therefore important the theory is correct.

Empirical difficulties with induction

• Constant conjunction can be accidental or caused by a third factor

• Causal mechanisms can be inferred, but rarely show their effects

Logical difficulties with induction

• Other hypotheses often account for data equally well

• Constant conjunction necessary but not sufficient to establish causality

• Proving validity and reliability by showing it works is simply proof by induction itself (circular argument)

Difference between facts and theories

Can’t be distinguished!

• What is defined as a fact is itself theory-based

• How we interpret our observations is a function of the concepts available

• Observers cannot record their environment without an initial framework within which they interpret information

Theory

Statements of how the world works. Some are common sense.

Theory-laden observations

The idea that all observations are interpreted through pre-existing theories before they influence new theories, and thus the two cannot be separated.

Working within scientific assumptions is working within a theoretical framework, so it is impossible for our observations to be free of bias even when personal biases are minimised.

The impossibility of discovering the nature of the world without preconceptions determined by experience, context, and/or values.

Theoretical framework

The scaffold of ideas and concepts within which assumptions and inferences can be made about facts and data.

A problem in psychology is that many people fail to use or make explicit these in their work, which inhibits critical evaluation of their work.

• There is an assumption that not using these increases objectivity.

• In reality, it just hinders comparison and robs the work of an easy-to-find foundation of knowledge.

Values

Prescriptive (“should”) statements about behaviour in terms of good or bad, ethical and unethical, moral and immoral.

Enter into science in numerous ways:

• Choice of problem to be investigated

• Planning studies to answer questions

• Collection and analysis of data

• Interpreting findings

• Making decisions about results

Theories

Sets of assumptions about the nature of things to be explained and understood; descriptions of reality through a lens.

Sets of interrelated statements (comprised of definitions and relationships assumed to be true), together with a set of rules for manipulating these statements to arrive at new statements. May also include a set of correspondence rules.

Should give information about how the world behaves, and therefore rule out ways in which is could possibly behave but does not.

Value-laden research

Theory-laden research in which the theories are chosen as a function of an individual researcher’s values, can be to any degree. In this way, accepting the assumptions of science is an expression of a set of values.

Consensual agreement in science

At any time, there is agreement about:

• The nature of the pursuit of knowledge

• This being subject to subject to historical and political fluctuations

• People pursue an understanding of a shared vision of reality

Observational-clarifying role of theories

Role of theories in which theories:

• Allow us to make sense of our sensory observations

• Provide links and connections between disparate observations

• Allow us to interpret and understand results from separate experiments

Analysis and synthesis

Processes by which theories can be built from individual statements into a bigger picture.

• Empirical research is an essential means, but not an end or a main focus of science

• Once at the ‘specific’ level, it is easy to see what measures and statistical analyses are required to test the hypotheses

Aim of science

To strive for abstract, universal, and conditional knowledge claims from which we can deduce statements pertinent to specific circumstances.

Correspondence rules

Rules in a theory that link abstract terms to a set of possible observations.

Concepts

Specialised, well-defined ideas; abstractions from shared experience that promote shared understanding.

Ideas

May refer to things, but are distinct from the things they refer to.

Primitive concepts

The simplest concepts which are assumed to be widely understood. Terms such as these in one theory may need to be defined in the context of another.

Examples:

• Response

• Behaviour

• Organism

• Arousal

• Drive

Defined (derived) concepts

Expressions which use or are descended from primitive concepts.

Examples:

• Response hierarchy

• Dominant/competing response

• Habit

• Excitatory potential

• Learning

Operational definitions

Observable manifestations of the concept.

Assumptions

Relational statements known to be true or accepted as truth.

Scope conditions

Statements specifying the range of applicability of a theory.

Conceptual hypotheses

Defined concepts and primitives which are used to make abstract derived relational statements.

These state what will happen if a given theory applies to the situation.

These remain propositional, and describe relationships between abstract concepts or variables.

Also known as general hypotheses.

Experimental hypotheses

Postulated relationships between empirical variables which have not been extensively tested, and/or have not received extensive empirical support.

These state what observations should be made if the conceptual hypothesis is correct.

Specific observational predictions and statistical hypotheses. (BUT DO NOT equate summaries of statistical relationships among measured variables for theories.)

Theory to hypothesis

Theories do not correspond directly to specific objects and events, and must first be translated, concept by concept and term by term, into statements about observable phenomena.

Initial conditions

The specific circumstances under which observations are made; the conditions of the experimental context.

Small, large

A _____ number of assumptions, definitions, and scope conditions can produce a _____ number of hypotheses.

Linking general to specific

To empirically test hypotheses, we must move from the abstract to the concrete (deduction).

To make predictions, we must show that:

• The scope conditions have been met

• The initial conditions of the experiment represent the involved aspects of the theory

• The theory’s concepts can be represented by the cases used in the experiment

Features of theories

Features:

• Predictive capability

• Links to observables

• Heuristic value

• Parsimony (simplicity)

• Complexity (in reality)

Criteria:

• Internal coherence

• External consistency

Falsification

A quality of a theory in which there are observable circumstances in which it can be possibly proven false, even though no number of observations can prove a theory true (only the best available answer).

A quality of a theory in which there exists a logically possible observation statement or set of statements that are inconsistent with it. If these statements are true, the hypothesis is falsified.

A property of a good scientific theory because it makes definite, refutable claims about the world. Conjectures may be refutable in principle, even if they are not refutable in fact.

Falsificationist

One who believes that observation is guided by and presupposes theory, and that a search for falsification tests a theory more severely.

Bold conjectures

Significant scientific advances are marked by the confirmation of these, or the falsification of cautious conjectures.

Before it can be regarded as an adequate replacement for a falsified theory, a new and ___ theory must:

• Make novel predictions

• Have these predictions confirmed

• Be unlikely in light of the background knowledge of the time - the significance thus depends on the historical context

The ____er, the better - the more it lays itself open to potential falsification, the more it purports to tell us about the world.

Cautious conjectures

The confirmation of these is uninformative, indicating only that some well-established theory has been, once again, successfully applied.

Three stages of theory-testing

1. A new idea (bold conjecture) is put forward, formulated as a tentative hypothesis or theory

2. Certain consequences are deduced

3. The deduced conclusions are examined to see if they are compatible with what is observed

Theories that have been falsified must be rejected - measurement must therefore be very accurate.

After falsification

Alternatives to discarding a theory or hypothesis:

• Modification is allowed if it produces a change in the original theory which increases its empirical content

• Falsifying observations can lead to changes which are developments of the theory

Methodological rules forbit the adoption of auxiliary hypotheses just to save the theory.

Null hypothesis

Almost always wrong. Not a bold conjecture.

Verification

Happens to theories if they are shown to be true. However, no scientific theory or hypothesis can be proven true - we can never reach truth, but can aspire to greater knowledge.

Naive falsification

A theory is falsified by a observation that the scientist decides to interpret as conflicting with it.

Tests are two-cornered fights between the theory and the experiment in which these two face only each other, and the only interesting outcome is conclusive falsification.

Sophisticated falsificationism

Tests are three-cornered fights between a theory, rival theories, and the experiment. Some of the most interesting confrontations result in confirmation rather than falsification.

A theory is acceptable or scientific only if it:

• Has corroborated excess empirical content over is predecessor/rival

• Explains the previous successes of its predecessor/rival

• Leads to the discovery of new facts

Challenge: Resolving inconsistencies between theories, not replacing a refuted theory.

• The counter-evidence of one theory may corroborate another theory

A theory cannot be regarded as falsified until a better theory is found.

Hard core

The defining characteristics of a research program, which are comprised of fundamental principles (primitive concepts).

Rendered unfalsifiable by the methodological decisions of its protagonists (scientists).

Cannot be modified - if it is, a new program will be created and the person altering it has opted out of the original program.

Protective belt

The supplementary assumptions and hypotheses which augment a research program’s hard core and provide a means by which definite predictions can be made, thereby protecting it from falsification.

Blame for failure is placed on these less-fundamental components. Thus, science can be seen as the programmatic development of the hard core’s implications.

Assumptions here are modified in attempts to improve the match between a research program’s predictions and the results of obesrvation and experiment.

• Successful efforts to solve problems by modifying peripheral assumptions will contribute to the development of the research program

• These modifications must be independently testable and, in themselves, be open to new tests and new discoveries

• More tests means more discoveries and greater chance of success

• Ad hoc modifications are ruled out

Positive heuristic

What scientists of a given research program ‘should’ do. Consists of a partially articulated set of suggestions on how to change and modify the refutabe protective belt.

• More difficult to characterise specifically

• Gives guidance

Negative heuristic

What scientists of a given research program ‘should not’ do. Commonly involves not tampering with the hard core.

Research program

A collection of theories and assumption a scientist works within. Consists of an untouchable hard core of foundational principles, and a protective belt of supplementary laws and principles which can be altered and falsified.

• Early work takes place regardless of falsifications of observation

• Protective belt constructed to be suitable, sophisticated, and adequate - failed predictions indicate more work needs to be done

• When appropriate to subject it to experimental tests, confirmations (not falsifications) are of paramount significance

Its worth is indicated by the extent to which it leads to confirmed and novel predictions.

Progressive research program

A research program that retains its coherence and, at least intermittently, leads to novel and natural predictions that are confirmed (natural as opposed to contrived or ad hoc).

Degenerating research program

A research program that loses its coherence and/or fails to lead to confirmed novel predictions.

Progress involves replacing this with a progressive program which has been shown to be a more efficient predictor of novel phenomena.

Science as problem-solving

Aim of science: Secure theories with high problem-solving effectiveness. Science progresses when successive theories solve more problems than their predecessors.

Types of problems:

• Conceptual

• Empirical

The worth of a theory depends partially on how many problems it solves.

Conceptual problems

When a theory:

• Is internally inconsistent, or the theoretical mechanisms it postulates are ambiguous

• Makes assumptions about the world that run counter to other theories

• Fails to utilise concepts from other, more general theories to which it should be logically subordinate

• Violates principles of research traditions of which it is a part

A change from an empirically well-supporte theory to a less well-supported theory could be progressive, provided that the latter resolved significant conceptual difficulties confronting the former.

Empirical problems

Three types:

• Potential problems: What we take to be the case about the world, but for which there is not yet an explanation

• Solvable/actual problems: Accepted claims about the world which have been solved by some viable theory

• Anomalous problems: Rival theories have solved them but are not yet solved by the theory in focus

A theory should solve a maximal number of these problems while generating a minimal number of anomalies.

Research traditions

These:

• Indicate what assumptions can be regarded as uncontroversial

• Help identify portions of a theory that are in difficulty and should be modified/amended

• Establish rules for the collection of data and testing of theories

• Pose conceptual problems for any theory

• Not directly testable

• Tend to be enduring entities

Worth of a theory

Partially determined by how many problems the theory solves. Mainly determined by comparing the theory’s effectiveness and progress to that of its rivals.

For every theory, we should:

• Assess the number and weight of the empirical problems it is known to solve

• Assess the number and weight of its empirical abnormalities

• Assess the number and centrality of its conceptual difficulties or problems

• Prefer theories that come closest to solving the largest number of empirical problems while generating the smallest number of significant anomalies and conceptual problems

• Evaluate using a cost-benefit analysis and compare to a wide context

However, the idea of accepting or rejecting a theory exists on a continuum - theories not yet worthy of acceptance may still warrant pursuit.

Methods

These:

• Express means-ends relationships

• Should be characterised in terms of their usefulness

• Free to invent new or discard old ones depending on their usefulness

• Differ depending on the research program

Self-determination theory

Theory stating that the most favourable outcomes of individuals’ endeavours emerge from actions that are most self-determined. These are understood to be derived from intrinsic motivations which come from a core, individualistic ‘true self’.

Three fundamentally-assumed needs determine one’s motivational orientation:

• Competence

• Autonomy

• Relatedness

Self-categorisation theory

A theory stating that people can and will act in accordance with the norms and values of their salient self-category/group. The self is dynamic and can be both individually and collectively represented.

Three suggested needs (not well-supported):

• Self-esteem

• Certainty

• Meaning-making

Needs can be socially constructed.

Surrogates for theories (Gigerenzer)

Vague, imprecise, often unfalsifiable attempts to replace theory.

Four types:

• One-word explanations

• Redescription (near-tautologies)

• Vague/muddy dichotomies

• Null hypothesis testing

Supported by:

• Insitutionalisation of null hypothesis testing (rather than testing against alternatives)

• Disciplinary isolation (preventing cross-discipline sharing of knowledge and creating intellectual ‘inbreeding’)

One-word explanations

A broadly-defined noun chosen to relate to the phenomenon it describes, which does not specify any underlying mechanism or theoretical structure. The temptation in using them is that they can be used to fit any observation that falls vaguely into the same domain of meaning.

A surrogate for theory.

Tautology (redescription)

A circular argument in which the term itself is a description of the behaviour/phenomenon it defines.

A surrogate for theory.

Near-tautology (Wallach)

A position derived from a circular argument, in which the logic may be sound but the foundational principles are circular in nature, thus making studies less robust.

Muddy dichotomies

The idea that strict dichotomies do not naturally exist, and so their constructed existence ‘muddies’ analysis to the point where things are impossible to interpret clearly as time goes on and knowledge is constructed around these ideas.

A surrogate for theory.

Data fitting

The danger of using mathematical tools to make models fit evidence post hoc, rather than testing if the data fits the theory or for modelling new theories.

A surrogate for theory.

Direct replication (Crandall and Sherman)

A type of study replication which aims to fully replicate the methods of a previous study. These aim to enhance confidence in the operationalisation and the precision of estimation. Type 1 and 2 error refer to false positives and negatives about the operations themselves. These are necessary and ethically mandated to ensure quality and monitor potential participant harm.

Schwartz & Strack: It is equally important to replicate the psychological circumstances of the original study, not just to ‘go through the motions’ in exactly replicating the methods - circumstances effect outcomes, and psychological equivalence is just as important as technical equivalence.

Conceptual replication (Crandall and Sherman)

A type of study replication which aims to replicate the concepts captured in a previous study, often with different methods. These aim to enhance confidence in theoretical hypotheses, and in ideas across a variety of circumstances. Type 1 and 2 error refer to false positives and negatives about theoretical hypotheses.

Klein: Failure to replicate is a symptom of under-developed theory in psychology as a field - theories lack specificity, and adherence to materialist tenets may have a limiting effect on the appreciation of the studied phenomena.

Gergen

Argued that all of psychology is and is determined by history and context - what determines the norms of human behaviour and its research is a product of time and place. It is impossible to escape bias, or the unpredictable human reactions to the implicit and explicit expectations of experimentation they are part of.

This author also asserts that the pursuit of knowledge and what is deemed important is shaped by the ideals and values of a place and time - knowledge may be used for propaganda purposes.