Statistics in the Courtroom

Important Statistical Concepts

Causality vs Correlation:

  • For Causality, one action must cause the other – Smoking causes lung cancer.
  • For Correlation, two actions occur at the same time (a measure of the relationship between two things) – Smoking is correlated with alcoholism.
  • The complication arises because when there is causality there is almost always correlation, but the opposite is not true!
  • e.g. Eating disorders and Soap Operas
  • Most studies, especially those in the social sciences are very good at establishing correlation, however, it is extremely difficult to prove causality.

Sample Size / Selection Biases

  • Sample size is the number of individuals that you are testing
  • The sample size is an important feature of any empirical study in which the goal is to make inferences about a population. Larger generally better – what is acceptable tho?
  • Overgeneralizing based on Small Samples
  • Sample of convenience
  • Attrition

Effect Size

  • The effect size is a measure of the strength of a phenomenon….i.e. The strength of a correlation. It emphasizes the size of the association between two groups.
  • Effect sizes are important when the metrics under analysis do not have intrinsic meaning, like a score on a personality test.
  • When they are reported - r .10 = small effect, .30 = medium, .50 = large.
  • Weight loss advertisements.

Confirmation Bias

  • Confirmation bias is a tendency for people to favor and seek out information that confirms their beliefs or hypotheses.
  • People do this consciously and unconsciously
  • The way we ask the question also influences the way we answer the questions

P Values

  • The chance the results are incorrectly deemed significant when they are not
  • Or the probability the result you obtained was due to chance and not actually real
  • False positive rates…Speed on Green Tickets

Understanding the Baseline

  • Need a foundation for measurement…a reference point
  • For comparisons to be valid, it is important to make some sort of adjustment for the starting point. It is an initial collection of data which serves as a basis for comparison with the subsequently acquired data.
  • Aboriginal and Low SES populations

Different Types of Research & Data

Qualitative Data

  • Data that approximates or characterizes; describes information.
  • Data that can be observed but not measured
  • Self-report measures
  • Observational analysis

Quantitative Data

  • Systematic empirical evidence that can be analyzed with statistical, mathematical, or computational techniques.
  • Measurement processes are critical to quantitative data
  • Quantitative data defines…whereas qualitative data describes
  • Peer Reviewed Journals
  • RCT
  • Meta-Analysis

Opinion

  • These are based upon personal opinion and belief.
  • May contain facts, but not necessarily
  • Generally will also include the opposing viewpoints
  • Caution – Books fall into this category
  • Look for references

Figures in Research Papers

-Scale

-Error Bars

How to Challenge Information?

1.)Empirical Studies

  1. Sample Size
  2. Sample Demographics
  3. Empirical Evidence vs Self Report
  4. Does the study say what the expert says it says
  5. Competing Interests

2.)Psychometric Testing

  1. Screening vs. Diagnosis
  2. Assessment
  3. Baseline Measures
  4. Validation Population
  5. Correlational Value
  6. Respondent Burden
  7. Defensive Responding/Lie Scales

1