aiMOOC

ABI - Statistics critically read

QR-Code

ABI - Statistics critically read

Mini-description: Check axes, percentages, baseline, sample and context; then state the limits of the claim. Your goal is not to distrust every statistic, but to read statistical evidence accurately, fairly and in context.

Target group: Englisch als fortgeführte Fremdsprache, Kursstufe Baden-Württemberg, gemeinsamer BF/LF-Kurs. English first; German help comes after context, word-building, visual support, English explanation and dictionary work. Language target: mainly B2, with selected C1-oriented moves where they improve precision.

Media opener: What impression does this graph create before you read any numbers? Now inspect the scale. Write one sentence beginning with: The graph appears to suggest ..., but ...

Compare: The two charts can be based on the same values but create a different visual impression. A non-zero axis is not automatically wrong; the key question is whether the scale is clear, appropriate and likely to distort interpretation.

Einleitung

Statistics can strengthen an argument, but a number never explains itself. A critical reader asks: What is measured? Compared with what? Who is included? What is missing? In an English exam, this skill helps you analyse non-literary material, diagrams and statistics without turning your answer into a mathematics exercise.

For the Baden-Württemberg Abitur 2027, the official Facherlass states that the Leistungsfach writing task may include extended-text material such as images, statistics and diagrams. In non-literary analysis, statistics can also function as a communicative strategy. The Basisfach works with listening, reading, analysis, mediation and different writing tasks during the qualification phase; its Abitur examination is oral. Both BF and LF are based on the applicable Bildungsplan 2016 for this cohort.

Official check, 02 October 2026: Facherlass für die Abiturprüfung 2027, Englisch Bildungsplan Englisch – Leitgedanken Englisch Kursstufe Leistungsfach Englisch Kursstufe Basisfach

I-can goals

I can ...	BF	LF extension
identify what a chart actually shows	use a guided six-lens check	prioritise relevant weaknesses independently
distinguish percentage points from relative percentage change	explain one clear example	integrate the distinction into an argument
judge a sample cautiously	identify size and selection issues	discuss representativeness, bias and uncertainty
analyse visual choices	comment on scale, labels and baseline	link visual design to communicative effect
evaluate a claim	state one justified limit	weigh alternative explanations and evidential strength
communicate findings	structured B2 response	broader, more differentiated B2/C1-oriented response

Prior knowledge quick check

Before you start, make sure you can read a percentage, identify x- and y-axes, distinguish a sample from a population and explain the difference between correlation and causation in simple English.

Think: Which target shows systematic bias? Which shows mainly noise? What would you need to know before generalising from a small set of observations?

The six-lens check

Lens	Ask in English	Typical issue
Axes	What do the axes show? Are labels, units, intervals and scales clear?	truncated scale, uneven intervals, missing labels
Percentages	Percentage of what? What are the raw numbers?	small denominators, percentage points confused with percent change
Baseline	What is the starting or comparison point?	dramatic change created by a narrow baseline
Sample	Who or what was sampled, how many, and how were they selected?	self-selection, small or unrepresentative sample
Context	Who produced the data, when, where and for what purpose?	missing source, outdated period, unclear definitions
Limits	What can the data not prove?	correlation presented as causation, missing variables, overgeneralisation

Core sentence frame: The figure supports the claim that ..., but its evidential value is limited because ...

Percentages: one small distinction, big effect

A rise from 20% to 25% is an increase of 5 percentage points. Relative to the original 20%, it is a 25% increase. Both statements can be mathematically correct, but they answer different questions.

Visual question: Does the design help you compare proportions accurately, or does perspective exaggerate some slices? Explain the effect rather than simply calling the chart “bad”.

Sample: who is actually represented?

A large sample can still be biased. A smaller carefully selected sample can sometimes be more informative than a huge self-selected online poll. Ask who had a chance to be included and who did not.

Listening focus: Note three sampling methods or sources of bias. After viewing, explain one of them without using the video’s wording.

Context and missing cases

This famous visual example reminds you to ask about cases that are not visible in the dataset. The dots show damage on aircraft that returned. A critical interpretation must consider the aircraft that did not return.

Transfer question: Where could a similar error occur in school surveys, product reviews, social media or migration statistics?

Same summary, different pattern

Anscombe’s quartet shows why summary statistics alone may hide important patterns. Similar averages, correlations or regression lines do not guarantee similar data structures.

Viewing task: Write a 30-second explanation of why correlation does not automatically establish causation.

Listening and Viewing

Before viewing: Predict three warning signs of weak statistical reporting.

During viewing: Listen for how study design, headlines and reporting choices can change what readers believe.

After viewing – BF: Give a 60-second summary using claim – evidence – limit.

After viewing – LF: Add a judgement about which limitation most strongly affects the inference and justify your choice.

Optional extension: Use the sections on confidence, relative/absolute risk and causation to refine your language of uncertainty. You do not need advanced mathematics to write a precise critical response.

Reading

Fictional practice text – not a real study:

A school newsletter reports: Students in our new study programme are 50% more confident than before. The article compares two voluntary online polls. In the first poll, 4 of 10 respondents selected “confident” or “very confident”; in the second, 6 of 10 did so. The participants were not necessarily the same students. No control group was used.

Poll	Confident responses	Total respondents	Percentage
Before	4	10	40%
After	6	10	60%

Read critically: The relative increase from 40% to 60% is 50%, while the absolute difference is 20 percentage points. The arithmetic does not prove that the programme caused the change. The sample is tiny, voluntary and not matched across time.

BF task: Explain two limits using the six-lens check.

LF task: Evaluate the headline as a piece of communication. Consider mathematical accuracy, sample quality, causal language and omitted context.

Context vocabulary

Rule: Try an English explanation and a monolingual dictionary first. Use the German help only if needed.

Term	English explanation	Word family / useful collocation	Deutsche Hilfe
axis	a reference line used to show values in a graph	x-axis, y-axis, axis label	Achse
scale	the system of intervals used to represent values	scale range, distorted scale	Skala
baseline	the starting or comparison point	baseline value, baseline period	Ausgangs- oder Vergleichswert
percentage point	the arithmetic difference between two percentages	rise by five percentage points	Prozentpunkt
relative change	change measured against the original value	relative increase, relative decrease	relative Veränderung
sample	the cases or people selected from a larger population	sample size, sample selection	Stichprobe
representative	similar enough to the target population for a justified inference	representative sample	repräsentativ
bias	a systematic distortion in data collection or interpretation	sampling bias, selection bias	Verzerrung
correlation	a statistical relationship between variables	positive correlation, correlate with	Korrelation
causation	a relationship in which one factor produces an effect	causal claim, causal link	Kausalität
outlier	a value far from most other values	identify an outlier	Ausreißer
limitation	a factor that restricts what a study or chart can show	methodological limitation	Einschränkung

Word-building for analysis

Base	Word family	Useful sentence
represent	representation – representative – unrepresentative	The sample may not be representative of the wider population.
compare	comparison – comparable – comparatively	The groups are not directly comparable.
vary	variable – variation – variability	The chart does not show how much individual results vary.
infer	inference – inferential	This inference goes beyond what the data can support.
distort	distortion – distorted	The compressed scale may distort the visual impression.

Analysis

Use this short sequence in exam-style analysis:

1. Describe precisely. The chart shows ... over the period ...

2. Quantify selectively. Use only figures that matter to the claim.

3. Analyse the presentation. Explain scale, baseline, labels, proportions or visual emphasis.

4. Check the evidence. Ask about sample, source, definitions and time frame.

5. Connect form and purpose. By foregrounding ..., the graphic encourages readers to ...

6. State limits. However, the data do not establish ...

Avoid: listing every number, calling a graph “manipulative” without evidence, or assuming that a visual flaw automatically makes all underlying data false.

BF and LF differentiation

Shared material	BF route	LF route
one chart plus short source note	guided six-lens grid; select two strong points	independent prioritisation; connect several features
fictional dataset	explain what can and cannot be concluded	assess inference, alternative explanations and uncertainty
speaking	60–90 second structured comment	2–3 minute comparative evaluation
writing	clear paragraph with sentence frames	integrated analysis with stronger material linkage and AFB II/III

BF work is exemplary and manageable. LF work adds breadth, depth, text complexity, independent material connection and more AFB II/III. Difficulty comes from thinking and precision, not from unnecessarily obscure vocabulary.

Speaking

Prompt: A chart can be mathematically correct and still be rhetorically misleading. Discuss.

BF support: use claim – example – limit – conclusion.

LF extension: distinguish between error, selective framing and legitimate design choice; then respond to a counterargument.

Recording: Audio/video recording is optional. Keep recordings private unless everyone involved has explicitly agreed to sharing. No unrequested uploads.

Writing

Practice task: Write an analytical paragraph on the fictional school survey.

Useful opening: At first sight, the figures seem to support the newsletter’s claim. However, a closer look at the percentages and sample weakens this conclusion.

BF scaffold: observation → evidence → limitation → cautious conclusion.

LF extension: integrate visual/statistical analysis with the source’s communicative purpose and evaluate the strength of the inference.

Self-check: Did you explain the effect of a statistical choice, not merely name it?

Mediation

German source note – fictional: Eine Schülerzeitung meldet, dass „60 Prozent der Befragten“ ein neues Lernangebot unterstützen. Befragt wurden zwölf freiwillige Teilnehmende eines Workshops; sieben stimmten zu. Angaben zur gesamten Schülerschaft, zur Auswahl der Befragten oder zu früheren Vergleichswerten fehlen.

Task: Your English-speaking exchange partner asks whether the result shows that the whole school supports the programme. Mediate only the information needed to answer the question. Do not translate sentence by sentence.

Help: The survey suggests ..., but it cannot be generalised confidently because ...

Help system

Level	Help
1	Highlight axis labels, units, dates and sample size.
2	Use the six headings: axes – percentages – baseline – sample – context – limits.
3	Use sentence frames: The figure suggests ... / This may exaggerate ... / The sample does not necessarily represent ...
4	German rescue help: Frage zuerst „Was wird gezeigt?“, dann „Womit wird verglichen?“, dann „Für wen gilt die Aussage wirklich?“

Feedback and error analysis

After a first draft, mark each sentence with one function: D = description, A = analysis, E = evidence, L = limitation. If most sentences are D, your answer is probably too descriptive.

Common error	Better move
The y-axis is wrong.	Explain exactly how the scale changes the visual impression.
60% is a lot.	State 60% of whom and give the raw denominator if available.
The study proves ...	Use suggests, is associated with or supports unless causal evidence is strong.
The sample is small, so the result is false.	Say the sample limits precision or generalisability; do not invent a stronger conclusion.
AI says the source is reliable.	Verify the source, numbers, date, method and citation yourself.

Responsible AI and source checking

Do the analysis and writing yourself. If you use AI as a support tool, record what you asked it to do, check every statistic and source, and correct unsupported claims. AI output is not a substitute for your own exam response. Do not send private class data to external AI services. Use fictitious or anonymised practice data unless a teacher has approved another source.

Transfer

Statistics may appear in reporting about migration, social change, identity, climate, health, education, economics or technology. The same six-lens check works across topics.

Transfer routine: Before agreeing or disagreeing with a claim, separate three questions: Is the number correct? Is the comparison fair? Does the evidence support the conclusion?

Abi-Check 2027

Leistungsfach: The 2027 written exam includes listening and writing. The writing material may include statistics, diagrams and other discontinuous texts. In analysing non-literary material, statistics can be part of the communicative strategies you discuss.

Basisfach: The Abitur examination is oral, while the qualification phase covers listening, reading, analysis, mediation and different writing tasks. Critical statistical reading therefore supports oral text/media discussion as well as course assessment.

Both levels: The applicable plan for the 2027 cohort expects B2 and, in parts, C1-oriented competence; BF and LF differ in complexity, breadth, depth and differentiation.

Exam habit: Do not spend time calculating everything. Select the figures that matter to the argument and explain their significance.

Follow-up course

Continue with ABI - Graphs and infographics analyse or ABI - Claims, evidence and bias. Reuse the six-lens check with a new topic and less scaffolding.

Interaktive Aufgaben

Quiz: Teste Dein Wissen

Which question best checks the axes of a graph? (Are the labels units intervals and scales clear) (!Who shared the graph on social media) (!How many colours are used) (!Whether the title is short)

A value rises from 20 percent to 25 percent. What is the difference in percentage points? (5 percentage points) (!20 percentage points) (!25 percentage points) (!125 percentage points)

What is the relative increase from 20 percent to 25 percent? (25 percent) (!5 percent) (!20 percent) (!45 percent)

What is the main question when checking a sample? (Who was included and how they were selected) (!Whether the chart uses blue bars) (!Whether the source has a logo) (!Whether the values are whole numbers)

What is a baseline? (A starting or comparison point) (!A list of all respondents) (!A type of pie chart) (!A synonym for correlation)

Which statement is most careful about correlation? (A correlation alone does not prove causation) (!Every correlation proves causation) (!Correlation means the data are false) (!Correlation can only occur in experiments)

Why can a self-selected online poll be misleading? (The participants may differ systematically from the target population) (!Online polls always contain calculation errors) (!Percentages cannot be used online) (!Large samples are never useful)

Which sentence states a limitation appropriately? (The data suggest an association but do not establish a causal effect) (!The graph is obviously lying) (!The result is false because the sample is not perfect) (!The percentage proves the claim for everyone)

What should you do when a chart uses a non-zero y-axis? (Check whether the scale is clear and whether it distorts the comparison) (!Reject the chart immediately) (!Assume the values are invented) (!Ignore the axis completely)

What is the strongest use of a statistic in analysis? (Connect the selected figure to the claim effect and limitation) (!Copy every number from the chart) (!Describe the colours only) (!Replace the analysis with calculations)

Memory

Axis	label unit and scale
Percentage	share out of a whole
Baseline	starting comparison point
Sample	selected part of a population
Context	source time place and definition
Limitation	boundary of a justified conclusion
Correlation	relationship between variables
Causation	one factor produces an effect

Drag and Drop

Ordne die richtigen Begriffe zu.	Thema
Truncated scale	Axes
Relative increase	Percentages
Starting value	Baseline
Self-selection	Sample
Missing source date	Context
No causal proof	Limits

...

Kreuzworträtsel

Baseline	What English term names the starting or comparison point?
Sample	What English term names the selected group studied?
Context	What term covers source time place and definitions?
Outlier	What term names a value far from most other values?
Correlation	What term describes a statistical relationship between variables?
Percentage	What term describes a share expressed per hundred?

LearningApps

Lückentext

Offene Aufgaben

Leicht

Axis detective: Find a freely licensed chart and mark axis labels, units, scale and baseline. Add two English sentences on what the design makes easy or difficult to see.
Percentage switch: Create three fictional examples that distinguish percentage points from relative percentage change. Check every calculation.
Headline repair: Rewrite an exaggerated fictional statistical headline into a cautious, evidence-based English headline.
Sample card: Design a one-page visual card that explains sample size, selection and representativeness using invented data.

Standard

Chart commentary: Record an optional private 90-second spoken analysis of a chart using claim – evidence – limit. Do not upload it unless you explicitly choose to.
Survey redesign: Improve a fictional self-selected school poll so that its sampling procedure becomes more defensible. Explain the changes in English.
Media comparison: Compare two freely usable visualisations of similar data and analyse how scale, proportion and labelling affect interpretation.
Mediation briefing: Turn a short German statistical note into a concise English briefing for a specified audience without translating line by line.

Schwer

Evidence audit: Analyse a current public statistic from a reliable source. Verify source, date, sample, definitions and limits before writing your conclusion.
Counterclaim task: Write two plausible interpretations of the same fictional dataset, then explain which parts are supported and which go beyond the evidence.
BF LF redesign: Build a BF version and an LF version of the same statistics task. Increase independence, material linkage and AFB II/III for LF without adding artificial language difficulty.
Responsible AI audit: Ask an approved AI tool to explain a fictional chart only if your school rules allow it. Compare its output with the source data, document errors and write your own corrected analysis. Never submit personal class data or use AI as a replacement essay.

Lernkontrolle

Transferanalyse: A newspaper chart starts its y-axis close to the lowest value. Explain when this choice may be defensible and when it may mislead. Refer to purpose, scale and reader impression.
Inference check: A voluntary poll of 2,000 app users finds 78 percent approval. Explain why the large number of responses alone does not establish representativeness.
Percentage reasoning: A rate rises from 8 percent to 12 percent. Explain the change in percentage points and as relative change, then discuss which wording would be clearer in a headline.
Causation challenge: Two variables rise at the same time. Develop at least two alternative explanations that show why correlation alone is insufficient for a causal claim.
Material link: Combine a short non-literary paragraph and a fictional chart. Explain how the chart strengthens, qualifies or conflicts with the written argument.
Communication effect: Evaluate how a source can remain numerically accurate while still creating a one-sided impression through selection, baseline or missing context.

Lernnachweis

For a strong learning record, you should be able to identify the six lenses without prompts, distinguish percentage points from relative percentage change, analyse axis and baseline choices, discuss sample quality without overclaiming, connect statistics to communicative purpose, separate correlation from causation, state limits in precise English and verify sources independently.

A BF learning record should show secure, structured application to manageable material. An LF learning record should additionally show broader material connection, independent prioritisation, differentiated evaluation and stronger AFB II/III performance.

Your work must be your own. If AI support was permitted and used, document it and verify its claims and sources.

OERs zum Thema

Commons media used in this aiMOOC: Truncated Bar Graph, Bar graph, Misleading Pie Chart, Survey-survey, Survivorship-bias, Anscombe's quartet and the statistical bias/noise illustration. Check each file description page for author and licence details before reuse outside the wiki.

Video sources: TED-Ed and CrashCourse via YouTube embeds. No full transcripts, compulsory literary texts or unauthorised film sequences are reproduced here.

Verknüpfte Lernbereiche

ABI - Statistics critically read

aiMOOC-Projekte

Inhalte werden geladen ...

Mediathek wird aus dem Wiki geladen ...

Frag mich zum aktuellen Lerninhalt oder wähle eine kurze Hilfe.

Textmodus bereit

Sprache