Week 2

• Theory (Behavioural Science)


  Set of statements about the mechanisms underlying a particular behaviour, organise and unify different observations of the behaviour and generates predictions about a behaviour,
• Construct


  hypothetical attributes or mechanisms that help explain and predict a behaviour in a theory
• Construct


  External factor: reward

  Construct: motivation

  External behaviour: performance

  
• Operational definition


  Procedure for indirectly measuring and defining a variable that cannot be observed or measured directly

  • specifies a measurement procedure (set of operations)
  • observable behaviour
  • uses the resulting measurements as definition and measurement of hypothetical construct
• Causes of variability


  • Differences in participant characteristics (e.g., age, gender, skill level).
  • Differences in participants' prior knowledge, experience, or motivation can influence outcomes.
  • Environmental influences (e.g., noise, time of day).
  • Situational variables such as interruptions or social context (e.g., group vs. individual settings).
  • Measurement errors due to tools or procedures (faulty equipment, human error)
• Methods to Minimize Variability:

  • Standardizing experimental conditions (controlled variables and settings)
  • Using precise, well-calibrated measurement tools, use digital tools to reduce human error
  • Random sampling to avoid selection biases.
  • Limit participant variability by using inclusion/exclusion criteria (e.g., testing only within a specific age range).
  • Selecting homogenous participant groups when appropriate
  • Use within-subject designs to control for individual differences
  • Repeat and calculate mean
• Types of validity


  • Face Validity: The extent to which a test appears to measure what it claims to measure
  • Content Validity: Covers the entire range of the construct's dimensions (e.g., measuring all facets of "stress").
  • Criterion Validity: Correlation with external criteria, such as established benchmarks (a new test for anxiety is valid if its results are similar to those of an established anxiety test)
  • Construct Validity: Degree to which the measure aligns with theoretical expectations and related constructs.
• Types of reliability

  Test-Retest Reliability:

  • Administering the same test to the same participants at two different times.
  • High correlation between results indicates good reliability.
  • Inter-Rater Reliability:
  • Agreement among multiple observers rating the same phenomenon.
  • Useful for subjective measures like coding behaviors or evaluating performances.
  • Internal Consistency:
  • Ensures all parts of a test measure the same construct.
  • Commonly assessed using Cronbach’s alpha (how closely related a set of items are as a group, which helps assess if they reliably measure the same construct)
• Deviation

  Difference between mean and actual data point

  Take each score and subtract the mean from it

  Can't add deviations, will cancel each other out as some + and some -
• Quantifying error 


  Sum of squared errors (makes negatives into positives, and then add)
• Quantifying error


  Sum of squared errors (makes negatives into positives, and then add)

  
• Problem with SS


  Depends on the number of scores
• Variance (s²)

  Average variability by dividing by the number of scores (n-1)
• Population


  Full collection of units to which we want to generalise a set of findings or a statistical model
• Sample


  A smaller (but still representative) collection of units from a population used to determine truths about the population (n-1)
• Problem with variance


  measured in unit squared ²

  use SD instead

  • large SD: more spread out, further way from mean (less accurate)
  • smaller SD: less spread out, closer to mean (more accurate)

  
• What do SS, variance, and SD represent?

  • fit of the mean to the data
  • variability in the data
  • how well mean represents observed data
  • error
• Test statistics

  
• Relative frequency

  
• Frequencies

  in histogram = when data have more than two possible values (quantitiative data)

  • if you want to see a proportion (how many people sleep 5+ hours) from the dataset = cumulative distribution
• Maths and Logs

  
• Maths symbols


  >> and << = much greater/less than

  ! = factorial (5! = 5x4x3x2x1)

  ∝ = proportional to

  
• Maths on R