Notes on Mean vs Median (Transcript)

Mean vs Median in Data

• Definition of the mean (average): the sum of observations divided by the number of observations.
  • Formula: $\text{mean} = \frac{\sum<em>{i=1}^n x</em>i}{n}$
• Everyday use vs statistical term: what people call the "average" is the mean in statistics.
• Definition of the median: the middle value when data are ordered from smallest to largest.
  • If the data are sorted as $x<em>{(1)} \le x</em>{(2)} \le \dots \le x_{(n)}$, then:
  • For odd $n$, the median is $x_{(\frac{n+1}{2})}$
  • For even $n$, the median is the average of the two middle values: $\text{median} = \frac{x<em>{(n/2)} + x</em>{(n/2+1)}}{2}$
• Key difference: mean can be pulled by outliers or skew; median reflects the middle of the distribution and is less affected by extreme values.
• Conceptual takeaway: mean and median can tell different stories about the data depending on distribution shape.

The Feet Data Example

• Data context: 9 people observed; total number of feet counted = 17.
• Compute the mean:
  • $\text{mean} = \frac{17}{9} \approx 1.888… \approx 1.89$
  • The transcript states the mean as 1.89 feet per person.
• Narrative devices: mentions of two-left-feet jokes and a high-tech prosthetic to illustrate data collection and variation.
• Median concept in this dataset:
  • If you arrange the feet counts from smallest to largest for the 9 people, the median is the middle value (the 5th value since $n=9$), i.e., the central observation.
  • The transcript notes the idea of the "middle two feet" as a way to illustrate the middle of the distribution; formally, for odd $n$ you take the single middle value, for even $n$ you average the two middle values.
• Important implication:
  • It is possible for the mean to be a non-integer that no one actually has (e.g., 1.89 feet), while the median is an actual observed value.
  • This underscores that the mean and the median can diverge in real data.
• Takeaway about the two measures:
  • Mean can imply an overall average that doesn’t reflect typical individuals if the data are skewed.
  • Median represents a typical observation in the ordered list.

Mean vs Median in Medicine and Risk

• In medical research, the mean risk is often higher than the median risk due to a small number of individuals with very high risk (outliers).
  • Example concept: heart disease risk distribution where some individuals have high cholesterol, diabetes, smoking history, or family history that elevates risk.
• Consequence of mean being pulled up by high-risk outliers:
  • More than half of the population may have a risk below the mean, even though the mean is elevated by a few high-risk cases.
• Practical and ethical implications:
  • Relying on the mean risk can overstate the typical risk and lead to unnecessary pills or tests for many individuals.
  • In decision-making, the median risk may provide a better sense of what a typical person faces when the distribution is skewed.
  • Highlights the importance of choosing the appropriate measure of central tendency for medical guidelines, patient communication, and policy.

Two Key Points About Scientific Statements (from the transcript)

• Statements must be precise: they should be stated clearly with the exact quantities involved (e.g., the mean risk, the median, etc.).
• They require careful thinking about truth and representativeness:
  • The concept of sampling is introduced as a foundational principle for making inferences from data.
  • The transcript playfully mentions sampling Doritos as a joke, indicating that sampling is a topic to be addressed later, but the underlying idea is to consider how samples reflect a population.

Formulas and Definitions (Summary)

• Mean:
  $\text{mean} = \frac{\sum<em>{i=1}^n x</em>i}{n}$
• Median (sorted data $x<em>{(1)} \le \dots \le x</em>{(n)}$):
  $\text{median} = \begin{cases} x<em>{(\frac{n+1}{2})}, & n \text{ odd} \ \frac{x</em>{(n/2)} + x_{(n/2+1)}}{2}, &amp; n \text{ even} \end{cases}$
• Key interpretation:
  • Mean is sensitive to outliers and skewness.
  • Median is robust to outliers and provides a central tendency that better reflects the typical observation when data are skewed.

Real-World Implications and Takeaways

• In skewed data (like many medical risk distributions), prefer the median to describe typical values.
• Use the mean when the distribution is roughly symmetric and not heavily influenced by outliers.
• Always consider the distribution shape before choosing which center measure to report for informing decisions and policy.
• Be cautious of relying solely on the mean in contexts with potential outliers or highly skewed data, especially in healthcare and risk assessment.