Statistical Analysis, Biostatistics, and the Role of Forensic Scientists

Roles of the Forensic Scientist

  • Investigative Role (Investigator):
    • Focuses on the crime at the beginning of the legal process.
    • Observes evidence, generates hypotheses/leads, and suggests explanations to guide police inquiries.
    • Acts to answer questions such as "What happened?" and "What is this?"
  • Evaluative Role (Evaluator):
    • Focuses on the suspect and acts at the end of the legal process.
    • Remains impartial, working for both prosecution and defence to evaluate findings given opposing propositions.
    • Quantifies uncertainty and provides objective assessments to assist the court in decision-making.

Differences between investigative and evaluative roles

Deductive vs. Inductive Inference

  • Deductive Reasoning:
    • Moves from general laws or premises to specific outcomes or predictions (Causes→Effects\text{Causes} \rightarrow \text{Effects}).
    • Certainty: If the premises are true, the conclusion must necessarily be true.
    • Limitation: Does not add new knowledge beyond what is contained in the premises.
  • Inductive Reasoning:
    • Moves from specific observations or data to general laws or underlying causes (Effects→Causes\text{Effects} \rightarrow \text{Causes}).
    • Uncertainty: Adds new knowledge beyond the premises, but conclusions remain probabilistic rather than certain.
    • Forms the primary logical framework for scientific evaluation by estimating how strongly observed evidence supports a hypothesis.

Measures of Center

  • Mean (Arithmetic Mean):
    • Calculated by summing all data values and dividing by the total count.
    • Sample Mean Formula:xˉ=∑xn\bar{x} = \frac{\sum x}{n}
    • Population Mean Formula:μ=∑xN\mu = \frac{\sum x}{N}
    • Properties: Uses every data value; non-resistant (highly sensitive to extreme outliers).
    • Weighted Mean Formula:xˉ=∑(w⋅x)∑w\bar{x} = \frac{\sum (w \cdot x)}{\sum w}
    • Frequency Distribution Mean Formula:xˉ=∑(f⋅x)∑f\bar{x} = \frac{\sum (f \cdot x)}{\sum f}
  • Median:
    • The middle value in a dataset sorted in ascending or descending order (denoted x~\tilde{x}, MM, or Med\text{Med}).
    • Odd Sample Size (nn): Exact middle value.
    • Even Sample Size (nn): Mean of the two middle values.
    • Properties: Resistant to outliers; does not change when extreme values are altered.
  • Mode:
    • The value(s) occurring with the greatest frequency.
    • Data sets may have no mode, one mode (unimodal), two modes (bimodal), or multiple modes (multimodal). Applicable to qualitative data.
  • Midrange:
    • The value halfway between the maximum and minimum values.
    • Formula:Midrange=maximum+minimum2\text{Midrange} = \frac{\text{maximum} + \text{minimum}}{2}
    • Properties: Non-resistant; highly sensitive to extreme values.

Measures of Variation

  • Range:
    • Difference between the maximum and minimum values (Range=maximum−minimum\text{Range} = \text{maximum} - \text{minimum}). Highly sensitive to extreme values.
  • Standard Deviation:
    • Measures how much data values deviate from the mean. Always non-negative (s≥0s \ge 0).
    • Sample Standard Deviation (ss):s=∑(x−xˉ)2n−1s = \sqrt{\frac{\sum (x - \bar{x})^2}{n - 1}}
    • Shortcut Formula:s=n(∑x2)−(∑x)2n(n−1)s = \sqrt{\frac{n(\sum x^2) - (\sum x)^2}{n(n - 1)}}
    • Population Standard Deviation (σ\sigma):σ=∑(x−μ)2N\sigma = \sqrt{\frac{\sum (x - \mu)^2}{N}}
  • Variance:
    • Equal to the square of the standard deviation (s2s^2 for sample variance, σ2\sigma^2 for population variance).
    • s2s^2 is an unbiased estimator of σ2\sigma^2, whereas ss is a biased estimator of σ\sigma.
  • Range Rule of Thumb:
    • Determines significantly low or high values:         Significantly Low Values≤μ−2σ\text{Significantly Low Values} \le \mu - 2\sigmaSignificantly High Values≥μ+2σ\text{Significantly High Values} \ge \mu + 2\sigma
    • Roughly estimates standard deviation from range:         s≈Range4s \approx \frac{\text{Range}}{4}
  • Empirical Rule (Bell-Shaped Distributions):
    • ∼68%\sim 68\% of data fall within xˉ±1s\bar{x} \pm 1s
    • ∼95%\sim 95\% of data fall within xˉ±2s\bar{x} \pm 2s
    • ∼99.7%\sim 99.7\% of data fall within xˉ±3s\bar{x} \pm 3s

Empirical rule for bell-shaped distribution

  • Coefficient of Variation (CV):
    • Expresses standard deviation as a percentage relative to the mean.
    • Sample CV:CV=sxˉ×100\text{CV} = \frac{s}{\bar{x}} \times 100
    • Population CV:CV=σμ×100\text{CV} = \frac{\sigma}{\mu} \times 100

Measures of Relative Standing and Boxplots

  • zz Score:
    • Standardized value indicating the number of standard deviations a data point xx lies above or below the mean.
    • Sample Formula:z=x−xˉsz = \frac{x - \bar{x}}{s}
    • Population Formula:z=x−μσz = \frac{x - \mu}{\sigma}
    • Significance Criteria: A value is significantly low if z≤−2.00z \le -2.00 and significantly high if z \ge +2.00$.\n\n![z score range for significant values](https://assets.knowt.com/pdf-flow-prod/8be34712-cce6-4030-b3c1-f9f966bc1c8f-figures/41.png)\n\n* **Percentiles (P_k):**\n * Measures of location dividing sorted data into 100 groups with roughly 1\% of data per group.\n * **Percentile Calculation for Value x:**\n        \text{Percentile of } x = \frac{\text{number of values less than } x}{\text{total number of values}} \times 100\n * **Locator Formula:**\n        L = \left(\frac{k}{100}\right) n\n * If Lisawholenumber,is a whole number,P_kistheaverageoftheis the average of theL\text{th}andand(L + 1)\text{th} sorted values.\n * If Lisnotawholenumber,roundis not a whole number, roundLuptothenextwholenumber;up to the next whole number;P_kistheis theL\text{th} sorted value.\n* **Quartiles & 5-Number Summary:**\n * **Quartiles:** Q_1 = P_{25},,Q_2 = P_{50}(Median),(Median),Q_3 = P_{75}.\n * **Interquartile Range (IQR):**\n        \text{IQR} = Q_3 - Q_1\n * **5-Number Summary:** Consists of Minimum, Q_1,,Q_2,,Q_3, and Maximum.\n* **Boxplots & Outliers:**\n * A graphical summary constructed using the 5-number summary.\n\n![Boxplot representation of 5-number summary](https://assets.knowt.com/pdf-flow-prod/8be34712-cce6-4030-b3c1-f9f966bc1c8f-figures/53.png)\n\n * **Outlier Criterion (Modified Boxplots):** A value is identified as an outlier if it is greater than Q_3 + 1.5 \times \text{IQR}orlessthanor less thanQ_1 - 1.5 \times \text{IQR}$$.