BENV0139 Statistics for Heritage Science
BENV0139 Overview
- Term 1 2026/27
- University College London
- Two-hour closed-book exam
- Postgraduate module
BENV0139 Statistics for Heritage Science teaches statistical thinking through collections, materials, environmental monitoring, surveys and conservation decisions.
- Choose from structure Variable type, sampling and intended action determine the statistical method.
- Interpret uncertainty Effect size and intervals matter more than a binary significance label.
- Diagnose before trusting Residuals, missingness and sensitivity checks qualify every model claim.
- Use heritage detail Materials, environments and stakeholders change what a result means.
What you get in the full BENV0139 Exam Bible
- 8 chapters, 39 pages — every examinable topic
- 28 original practice questions, each with a model answer in the Centrally organised exam format
- 8 worked examples, solved the way BENV0139 marks
- 13 computed figures, redrawn from the course’s own numbers
- Exam-morning plan for the Centrally organised exam (100%) — what to revise last, where marks are lost
Free on this page: assessment table, chapter map, one worked example, key terms, FAQ. The rest is in the full exam bible.
How BENV0139 is assessed
| Component | Weight | Format |
|---|---|---|
| Centrally organised exam | 100% | Term 3, usually May. Two hours, in person and closed book. The exact date is announced during Term 2. |
The Moodle assessment page says the module is evaluated with one centrally organised exam and names no other component, so the table expresses the sole-component weight as 100%. The captured page does not print that percentage as a separate numeral.
Assessment structure
What other students are asking about BENV0139
- Where can I find lecture notes or study guides for BENV0139 at ucl?
- The BENV0139 study guide in AskSia Library covers Statistics for Heritage Science in 8 chapters, with the assessment table, a chapter-by-chapter topic map, key terms and a worked example, free to read on this page.
- Is there a final exam in BENV0139, and how much is it worth?
- Yes. In the offering this AskSia Library guide is built from, Centrally organised exam carries 100% of the BENV0139 mark; the full split is Centrally organised exam 100%.
- What does BENV0139 cover on summaries, samples and visual distributions?
- In AskSia Library's BENV0139 guide, the Summaries, Samples and Visual Distributions chapter explains mean, median, variance, standard deviation, IQR, accuracy, precision, histograms, boxplots, skew and modality.
- What does BENV0139 cover on probability, normal models and the clt?
- In AskSia Library's BENV0139 guide, the Probability, Normal Models and the CLT chapter explains gaussian models, standard scores, one-sided and two-sided areas, binomial counts, sampling distributions and CLT boundaries.
- What does BENV0139 cover on inference and hypothesis tests?
- In AskSia Library's BENV0139 guide, the Inference and Hypothesis Tests chapter explains null hypotheses, p-values, Type I and Type II errors, standard error, t tests, confidence intervals and power.
Current BENV0139 dates
| Date | Item | Control |
|---|---|---|
| Term 3 | Centrally organised exam | Usually in May; the exact date is announced during Term 2. |
Dates are as published in the 2026/27 Moodle assessment page. Confirm exact deadlines and submission settings in the live LMS.
What BENV0139 covers
Eight connected chapters move from heritage data and visual summaries through probability, inference, regression, multivariate methods, study design and evidence-based action.
Heritage Data and Statistical Thinking
Questions, variable types, observational units, uncertainty, sampling bias and the difference between description, association and causation.Heritage Data and Statistical Thinking begins with the decision, observational unit and measurement process.
- Heritage data formats
- Variable and measurement types
- Observational units
- Population and sample
- Sampling bias
- Repeated measurements
Summaries, Samples and Visual Distributions
Mean, median, variance, standard deviation, IQR, accuracy, precision, histograms, boxplots, skew and modality.Summaries, Samples and Visual Distributions connects numerical summaries to the shapes they describe. Mean and standard deviation are sensitive to every observation and work naturally for roughly symmetric distributions.
- Mean and median
- Variance and standard deviation
- Quartiles and IQR
- Histograms and bin width
- Boxplots and outlier flags
- Accuracy and precision
Probability, Normal Models and the CLT
Gaussian models, standard scores, one-sided and two-sided areas, binomial counts, sampling distributions and CLT boundaries.Probability, Normal Models and the CLT turns thresholds into areas under stated distributions. A normal model is defined by its mean and standard deviation, and a z score expresses a value's signed distance from the mean in standard-deviation units.
- Normal distribution
- Standard score
- One-sided tails
- Two-sided areas
- Binomial counts
- Central Limit Theorem
Inference and Hypothesis Tests
Null hypotheses, p-values, Type I and Type II errors, standard error, t tests, confidence intervals and power.Inference and Hypothesis Tests explains how evidence is evaluated against a null model without turning p-values into probabilities that a claim is true.
- Null and alternative hypotheses
- P-value interpretation
- Type I error
- Type II error
- Standard error
- Independent and paired t tests
Regression and Model Diagnostics
Coefficient interpretation, prediction, residual patterns, heteroscedasticity, nonlinearity, R-squared, selection and confounding.Regression and Model Diagnostics interprets fitted relationships through coefficients, predictions and residuals.
- Linear and multiple regression
- Coefficient units
- Prediction and extrapolation
- Residual diagnostics
- Heteroscedasticity
- R-squared and model fit
PCA, Factor Analysis and Clustering
Loadings, scores, explained variance, scaling, latent factors, clustering choices and multivariate interpretation.PCA, Factor Analysis and Clustering separates three multivariate tasks that are often blurred. PCA constructs orthogonal directions capturing decreasing amounts of total variation.
- Principal component directions
- Loading plots
- Score plots
- Explained variance
- Scaling and standardisation
- Latent factor interpretation
Sampling, Study Design and Robustness
Outcome definition, longitudinal structure, missingness, confounders, grouped data, diagnostics, sensitivity and generalisability.Sampling, Study Design and Robustness builds an analysis plan from a measurable outcome to a decision. Longitudinal environmental records do not create longitudinal evidence about deterioration when condition is rated only once.
- Outcome definition
- Cross-sectional and longitudinal designs
- Missing-data mechanisms
- Confounding and interactions
- Grouped observations
- Sensitivity analysis
Statistical Evidence for Heritage Decisions
Effect size, interval overlap, study comparability, practical importance, risk, stakeholder constraints and conditional recommendations.Statistical Evidence for Heritage Decisions integrates effect estimates, intervals, study quality and practical stakes.
- Effect size
- Confidence interval overlap
- Study comparability
- Statistical power
- Practical significance
- Decision costs and reversibility
The module welcomes varied backgrounds and begins with data typologies before moving through distributions, probability, inference, regression and multivariate analysis.
Its central habit is to choose a method from the research question and data structure, not from a memorised list.
The current assessment page states that the module is evaluated with a centrally organised exam in Term 3. The exam is two hours, in person and closed book. Its format is expected to resemble the recent examples, with an emphasis on narrative and critical discussion of statistical ideas in a heritage context.
Sample papers differ in their number of questions, so revision should focus on the reasoning that survives layout changes: define the statistical target, show the calculation or output interpretation, test assumptions, and connect uncertainty to a heritage action.
The module's marking criteria join two essential areas.
Statistical methods must be explained precisely, selected appropriately and interpreted with their assumptions and limitations. Heritage examples must be realistic and detailed enough to show how materials, sampling constraints, stakeholder needs and conservation consequences change the analysis.
Clear scientific writing and structure matter because the marker must be able to follow the chain from data to decision.
This guide therefore mixes numerical drills with narrative planning.
It covers robust summaries for skewed or uncertain data, normal probabilities and z scores, null hypotheses and t tests, residual diagnostics, regression coefficients, PCA loading plots, missing-data mechanisms, confounding and evidence synthesis. Every worked answer ends by stating what the number can support and what would require stronger design or additional evidence.
Worked example · free
Standardise a temperature threshold and select the tail
- 1Translate exceeding into an upper-tail probability and sketch the region to the right of 22.0°C. This seven-mark allocation is original practice weighting, not a published UCL marking scheme.
- 2Standardise: z = (22.0 - 19.0) / 2.0 = 1.50.
- 2Use the standard normal upper tail: P(Z > 1.50) is approximately 0.0668.
- 2Interpret the result as about 6.7% of comparable days under the stated normal model, and note that seasonality or control failure could invalidate that model.
Key terms
- Observational unit
- The object, site, person or time period represented by one independent unit of analysis.
- Sampling bias
- Systematic distortion caused when the observed sample does not represent the target population.
- Standard deviation
- A measure of the typical distance of observations from their mean, expressed in the data's units.
- Interquartile range
- The width of the middle half of ordered observations, calculated as the upper quartile minus the lower quartile.
- Standard score
- The signed distance between a value and a mean measured in standard-deviation units.
- Sampling distribution
- The distribution a statistic would have across repeated samples generated by the same process.
- Null hypothesis
- The parameter claim used to calculate how surprising the observed test statistic would be.
- Type I error
- Rejecting a null hypothesis when that null claim is true.
- Type II error
- Failing to reject a null hypothesis when a specified alternative is true.
- Confidence interval
- A model-based range produced by a procedure designed to cover the target parameter at a stated long-run rate.
- Heteroscedasticity
- A residual pattern in which variability changes across fitted values or predictor levels.
- Confounder
- A variable related to both an exposure and an outcome that can distort their observed association.
- Principal component
- A weighted direction through variables that captures as much remaining variation as possible subject to orthogonality.
- Robustness check
- A planned alternative analysis used to test whether a conclusion depends on a defensible modelling choice.
BENV0139 FAQ
How difficult is this module?
The published material does not provide a difficulty rating. It does show a broad progression from descriptive statistics and probability through inference, regression and multivariate methods. The closed-book exam also rewards narrative reasoning in a heritage context, so success depends on explaining assumptions and consequences as well as carrying out calculations.
What is the assessment structure?
The Moodle page states that the module is evaluated with a centrally organised exam and lists no other component. It takes place in Term 3, usually in May, lasts two hours, and is in person and closed book. The exact date is announced during Term 2.
What is the current exam format?
The current page says the style will be closer to the 2023-24 and 2024-25 examples, with narrative and critical discussion of statistical concepts in a heritage context. Those samples have different question counts, so the current section count should be taken from the live paper rather than assumed from one past year.
Are past papers and solutions available?
Yes. The Moodle assessment section lists solved examples from several years and a separate practice archive. This guide uses their question types and marking cues but writes original questions. Older papers remain useful for content practice, while their conditions do not override the current two-hour closed-book format.
How much mathematics should I revise?
Revise the meaning and use of each relation rather than symbols alone. You should be able to calculate and interpret summaries, z scores, standard errors, t statistics, coefficient intervals and simple regression predictions. A correct number still needs units, assumptions and a heritage interpretation.
What makes a strong heritage example?
A strong example names the object or building, the measured variables, the sampling constraint, the decision-maker and the conservation consequence. It explains why heterogeneity, rare samples, uncertain dates or material sensitivity changes the method or conclusion. Replacing a generic noun with museum does not create heritage depth.
How should I revise in the final week?
Alternate one calculation set with one narrative plan each day. Classify errors by tail, denominator, units, assumptions, confounding or missing heritage context. Rework the same reasoning with fresh numbers or a new material setting, then practise a timed closed-book answer that ends with a conditional recommendation.
Questions students ask in each BENV0139 chapter
- Why begin with the heritage question?
- The question defines the outcome, population and decision. Those choices determine which variables matter, whether observations are independent, and what kind of uncertainty could change the action. Starting with a test name can produce a technically correct calculation that answers the wrong conservation question. Chapter 01 · Heritage Data and Statistical Thinking
- When should I use median and IQR?
- Use them when a distribution is skewed, contains uncertain extremes or has outliers that would pull the mean and standard deviation away from the typical observation. Explain the shape and measurement context rather than invoking robustness as an automatic rule. Chapter 02 · Summaries, Samples and Visual Distributions
- What does a z score tell me?
- It tells you how many standard deviations a value lies above or below a mean under the stated scale. It is not itself a probability. Convert the threshold first, then use the question wording to select the correct tail or interval area. Chapter 03 · Probability, Normal Models and the CLT
- What does a p-value mean?
- It is the probability, under the null model, of obtaining the observed result or one at least as extreme according to the test statistic. It is not the probability that the null is true and does not measure the practical importance of the effect. Chapter 04 · Inference and Hypothesis Tests
- What makes a residual plot acceptable?
- Look for a roughly random band around zero with no systematic curve, trend or changing spread. Also inspect influential observations and grouping. A plot need not look perfectly random, but visible structure should be explained or addressed before ordinary inference is trusted. Chapter 05 · Regression and Model Diagnostics
- How do I interpret a loading plot?
- Start with vector direction and length. Similar directions indicate positive association in the displayed plane, opposite directions indicate negative association, and near-right angles suggest weak relation there. Then use heritage knowledge to propose a cautious axis label. Chapter 06 · PCA, Factor Analysis and Clustering
- Why is one condition survey not longitudinal?
- Repeated environmental predictors do not make the outcome repeated. One final condition rating supports mainly between-museum comparison and remains vulnerable to baseline differences. Repeated standardised condition assessments are needed to estimate within-museum change directly. Chapter 07 · Sampling, Study Design and Robustness
- Can significant and non-significant studies agree?
- Yes. Their estimates and confidence intervals may overlap substantially even when p-values fall on different sides of 0.05. Compare magnitudes, uncertainty, populations and methods before describing conflict. A binary label discards much of the evidence. Chapter 08 · Statistical Evidence for Heritage Decisions
How to study for the exam
Study the module as a chain from question to decision. Begin with variable types, observational units and distribution shape because these choices determine every later calculation. Build a compact relation sheet from memory, but attach each formula to its target: individual observation, sample mean, difference in means, regression coefficient or principal component.
For probability, draw and shade the requested area before using a table. For inference, state the null, tail, threshold, effect estimate and limitation. For regression and PCA, interpret units and geometry before naming a story, then inspect diagnostics and preprocessing. Keep a two-column error log: the statistical error on the left and the heritage consequence on the right.
A wrong tail changes an exceedance decision; ignored clustering makes uncertainty too small; a generic model claim can misdirect treatment or monitoring. Use solved papers to learn question families and marking cues, never to memorise proprietary wording. In the final week, practise under closed-book timing. Draft five-part plans that define, apply, challenge, interpret and decide.
Check each example names a material, collection, building or conservation action. Finish by explaining what result would change the recommendation.
Your AI Statistics tutor for BENV0139
Stuck on a hard BENV0139 question? Sia is AskSia’s AI Statistics tutor — ask any BENV0139 Statistics for Heritage Science question and get a clear, step-by-step explanation grounded in how the course is actually taught and assessed. Read this whole study guide free, then take your hardest questions to Sia.