Lecture 5:
Stages of behavioural studies research:
Problem
Theory/hypothesis
Research design
Measurement selection/design
Select sample
Collect data/run study
Code and analyse data
Write research report
Measurement:
Can it be quantified
Quality of measures
Changes in variables
Causality:
Correlation is not causation but can show how one thing leads to another
Variables:
Independent variable: the thing that influences the DV.
Dependent variable (outcome variable): a variable that is influenced by
another e.g. weight, drinking.
Validity:
The degree to which a measure accurately measures what it is supposed to
measure (Vogt & Johnson, 2011)
Types of validity:
Face validity: where it looks like it measures what it is supposed to measure
Concurrent validity (Criterion validity): does it correlate strongly with an
already established measure? If you develop a measure for alcohol drinking,
results must match scores already available for alcohol drinking.
Predictive validity (Criterion validity): Does it actually measure the behaviour
it is supposed to represent.
Discriminant Validity: Making sure it doesnt measure something unrelated to
the construct of interest.
Reliability:
Reliability of a measure means that it is consistent and stable.
Types of reliability in questionnaires:
Test-re test: the measure should result in the same scores at two different
time points.
Split-half reliability: if a 10-item scale measuring self esteem was divided in
half (e.g. odd vs even questions) the two sets of 5-items should yield the
same scores.
Scales of measurement:
Stanley Stevens (1946) on the theory of scales of measurement: All
measurement can be done using one of four scales: 1. Nominal (labels,
catergories or catergorical data), 2. Ordinal (ranking, ordering, relative
position), 3. Interval (assesses the distance between things/objects/items
i.e. 25 degrees to 35 degrees is the same as 45 degrees to 55 degress, 4.
Ratio (similar to interval, but has an absolute zero point: e.g length of hair,
unlikely to be used in social sciences) . These are challenged ideas.
Strongly agree to strongly disagree: Likert scale or Interval scale
Ranking food in order of preference: ordinal scale
Where do you live: Richmond, Caulfield, Clayton: nominal scale
Qualitative vs Quantitative:
Qualitative: open-ended format (can provide insights and depth)
Quantitative: can provide degree or level, level of agreement, involve more
people because of its simplicity
Types of scale responses:
Forced choice (dichotomous): yes or no
Forced choice scale (no mid-point): -3 -2 -1 1 2 3
Semantic differential opposite meanings as anchors for ends of scale: Hate
------------------ Love. Issue with this is equivalence between scales. Would
need to make up a quantifiable scale to relate to it.
Types of attitude measures for attitudes and behaviours:
Direct measures: self report, questionnaires, scales, surveys and behavioural
measures
Indirect measures:
o Observational: watch behaviour, reaction, eye movements
o Physiological measure: pupillary dilation, facial muscular activity, brain
activity, skin response
o Implicit measures: word/stem completion (method for assessing
memory and priming tasks), implicit association test (IAT): new
approach, designed to test strength of a persons automatic
association between mental representation of objects.
Explicit (direct): measures may represent conscious attitudes
Implicit (indirect): may represent stored associations built up over time. May
be unconscious.
Socially desirable responding (SDR):
Impression management: we want to appear to be a good person. Try to
please researcher.