Operationism and the hypothetico-deductive model of hypothesis testing has arisen from the history of Viennese Logical Positivism and behaviourism in America.

The golden model of scientific Psychology
Theories are complexes of ideas that give rise to hypotheses which are predictions based upon theories.

Hypotheses are disconfirmed (refuted) by observations (data) but Positivism suggests that verification gives rise to proof. (a logical mistake according to Sir Karl Popper)

Variables & Relationships among variables is another way to talk about hypothetical concepts and actions and the probable relationships between them.

People are seen as "carriers of variables" where we can be measured in many ways but our true nature is hypothesised to exist beyond the theories and hypotheses that we use to think about life and its problems and puzzles.
 

(See Mach-Sensationism)

There are problems surrounding the "external" demands of the theory (assumptions that cannot be denied) which makes it appear like a systemic viewpoint or paradigm that defines reality.

The recognition of the systemic worldview as such is part of the hermeneutical position that recognises that context (of preunderstanding) out of which the theory and data emerge (understanding).

Summary

Dependent - The variable that is measured during the study, it is the responses, behavioural or verbal, . . . data

Intervening - Theoretical constructs representing hypothetical processes) which enable the prediction of dependent variables. For example: B=fx(S, P, H, T, A) stimulus, past, hunger, thirst,...

Hypothetical Constructs - are characteristics or features of theories that are believed to important or influential to phenomena under study. Part of a nomological network (collection of concepts as part of a theory or model)

Operational definitions specific statements of how a phenomenon is going to be measured

Observational vs. Experimental

A distinction has long been made between manipulated and non-manipulated results. By invoking some control over the situation the experimental approach is set to make a stronger statement about causality.

Causality a problem for the ages, is assumed to be found through particular methods, or confirmed through special experience, however, David Hume clearly pointed out the limits of empiricism: causality is but a reflection of habit.

Ultimately it comes down to probabilities and how much confidence and individual observer (scientist) can place in some data or theory. Mach, Bridgman, and Fisher collectively stated: it comes down to personal experience & confidence.

Experiments offer the possibility of eliminating suspect causes or variables. Making use of the falisificationist view, when a hypothesis does not accurately predict the data it can be eliminated or revised, possibly challenging the theory too.

A positivist bias often settles in though, a desire to be certain!

Predictions for the basis of hypothesis testing and causality

Field Studies may contain naturalistic or participant observations. It is possible also to carryout a field experiment or quasi-experiment where natural responses can be observed.

Surveys: through face to face, telephone, internet, and paper-and-pencil questionnaires a great deal can be learned about people by asking them directly (self-report)

Ethics always a concern, as discussed last week need to think about it before hand and take precautions for everyone

Participants may also have a determination of the methods that you select. Depending on their talents and disabilities you may (or not) have access to various kinds of information

Often make use of randomization of participants but sometimes are interested in the social or personal identities as they carry information relevant to the study. Thus group identity is a substitute for the independent variable, however it more easily falls prey to extraneous variables and cohorts.

Case study - Description of Behaviour or personal subjective reports and interpret the meaning of the utterances or actions in terms of a theoretical model in psychology.

Narrative development of a Personal Lived Experience. Often the case study model or autobiographical approach stand as legitimate such as the psychohistory making model

Multiple Methods often provide a superior study, making use of the strengths of various forms in covering for their weaknesses. E.g., triangulation or multi-trait-multi-methods.

Validity of Research - Does the research measure and actually tell you about what it says it does?

Construct validity -does the operational definition and application measure what it attempts to? Hypothetical Constructs are developed in theories and tested in studies.

Face Validity - does the measure appear to measure what it is designed to do. Do it look like it?

Convergent and Discriminant Validity are important when using multi-measures and other variables. Here test the ability to predict the positive and negative correlations or relationships with other variables of constructs.

Criterion Validity - does it act like a predictor variable where it can predict fitness into a category or stream

Internal Validity Cozby suggests that it is about cause and effect - really is about prediction and confidence

External Validity does it predict or meaningful relate to other measures and phenomena? Populations? Settings?

Reliability pertains to the property of tests or measures that indicates how consistently the construct is measured. Reliable measures appear to have low measurement error.

test-retest reliability - pertains to the closeness (correlation) of two scores from one person on two occasions

internal consistency - how consistent the various questions or items on a scale measure the same thing. Cronbach's alpha is a method for testing item-total correlations.

inter-rater reliability - is necessary to determine the consistency of scoring people into categories or classes. When certain criteria are used to classify people one wants to know how much average agreement there is between raters.

Reactivitiy of measures
-does it have an impact on someone's behaviour?
-unobtrusive or non-reactive

Variables and Measurement Scales
nominal - simply naming the data - placing people or actions into categories not giving numerical quantities to data
ordinal - simple rank ordering with no attention to the distance or magnitude of the differences (no some many, ...)
interval - ordered set of categories that have equidistant spaces between them but have no absolute zero or centre
ratio - ordered and interval categories that have absolute '0'

Problems in measurement
 Identification of various types of research methods is often dependent on the type of variables selected.

The variables selected are part of the issue of measurement. Does the type of variable match the phenomenon being measured? Are you mixing variable or scale types?

Must ask "How can we measure the 'variables' or features that we are concerned with once we have chosen a way to measure them what kind of analysis can we do?"
Descriptive / Qualitative

Numerical / Quantitative
There is a rough continuum between quantitative and qualitative methods. At one extreme there exist only numerical analysis and at the other end there are only descriptions and interpretations. In between there are some methods that make use of some (basic) numerical notions while also making use of some (basic) categorical data.