kurtosis
C2Technical / Academic
Definition
Meaning
A statistical measure that describes the shape of a probability distribution, specifically the "tailedness" (the presence and extremity of outliers) and the peak's sharpness compared to a normal distribution.
In broader contexts, it refers to the degree of concentration of data points in the tails versus the center of a distribution. High kurtosis indicates heavy tails and a sharp peak (more outliers), while low kurtosis indicates light tails and a flatter peak.
Linguistics
Semantic Notes
Primarily used in statistics, finance, risk management, and data science. The concept is often contrasted with 'skewness' (which measures asymmetry). Not to be confused with general 'peakedness'; modern interpretation focuses more on tail weight.
Dialectal Variation
British vs American Usage
Differences
No significant lexical differences. Conceptual focus in finance may vary slightly: US texts more frequently link it to 'tail risk' in financial modelling, while UK texts might emphasise its role in quality control and engineering statistics.
Connotations
Neutral technical term. In finance, carries a connotation of risk assessment (high kurtosis = higher risk of extreme events).
Frequency
Extremely rare in general language. Exclusively found in advanced technical literature in both varieties.
Vocabulary
Collocations
Grammar
Valency Patterns
The distribution shows/ exhibits/has a high kurtosis.We calculated/measured the kurtosis of the dataset.Kurtosis is defined as...compared to a normal distribution with zero excess kurtosis.Vocabulary
Synonyms
Strong
Neutral
Weak
Vocabulary
Antonyms
Usage
Context Usage
Business
Used in quantitative finance and risk management to assess the probability of extreme market movements ("black swan" events).
Academic
Core concept in statistics, econometrics, and data science for describing distribution properties. Common in research papers analysing non-normal data.
Everyday
Virtually never used.
Technical
Essential in statistical software output, Six Sigma quality control, signal processing, and any field modelling data with outliers.
Examples
By Part of Speech
noun
British English
- The kurtosis of the returns was markedly positive, suggesting a greater risk of extreme losses.
- Pearson's measure of kurtosis is still commonly cited in older textbooks.
American English
- The software automatically calculates the kurtosis and skewness for any imported dataset.
- A high kurtosis value warned the quants of potential tail events in the model.
adjective
British English
- The kurtosis coefficient was reported in the appendix.
- (As part of compound adjective) A heavy-tailed, high-kurtosis distribution.
American English
- The kurtosis statistic is sensitive to outliers.
- They analysed the kurtosis properties of the waveform.
Examples
By CEFR Level
- In basic statistics, we learn about the mean and variance, but advanced topics include kurtosis.
- A normal distribution has a specific kurtosis value, which we use as a benchmark.
- The researcher noted the leptokurtic nature of the data, indicated by an excess kurtosis greater than zero.
- Financial asset returns often exhibit high kurtosis, violating the assumption of normality in many models.
Learning
Memory Aids
Mnemonic
Imagine a KURT (a name) who is an extreme rock climber. He loves the tails of distributions (the extreme ends) and the peak of the mountain. KURT-osis tells you about Kurt's favourite parts: the tails and the peak.
Conceptual Metaphor
THE SHAPE OF A DISTRIBUTION IS A LANDSCAPE. High kurtosis is a tall, sharp mountain with long, sloping foothills (tails). Low kurtosis is a wide, flat hill with short, abrupt edges.
Watch out
Common Pitfalls
Translation Traps (for Russian speakers)
- Do not confuse with 'куртоз' (non-existent). The standard Russian translation is "эксцесс" (excess kurtosis) or "куртозис". The concept of 'эксцесс' specifically refers to 'excess kurtosis' (kurtosis minus 3).
Common Mistakes
- Pronouncing it as /ˈkɜːrtəsɪs/ (wrong stress).
- Using 'kurtosis' to mean 'skewness'.
- Thinking high kurtosis means 'more peaked' without considering tail weight.
- Using 'kurtosis' in everyday language.
Practice
Quiz
What does a high kurtosis value most directly indicate about a probability distribution?
FAQ
Frequently Asked Questions
It is neutral statistically but has implications. In finance, high kurtosis ("fat tails") means a higher risk of extreme events, which is typically undesirable for risk-averse investors.
Skewness measures the lack of symmetry (lopsidedness) of a distribution. Kurtosis measures the combined weight of the tails relative to the rest of the distribution, focusing on outlier propensity.
It is kurtosis minus 3 (the kurtosis of a normal distribution). So, excess kurtosis = 0 for a normal distribution, >0 for heavy tails, and <0 for light tails.
Yes. For a standard kurtosis formula, the minimum value is 1. For *excess* kurtosis, values can be negative (e.g., -1, -2), indicating a distribution with lighter tails and a flatter peak than a normal distribution (platykurtic).