Statistical correlation of paired data

From Single Variables to Paired Data

In single-variable statistics, we summarize one measurement at a time using histograms or box plots. However, real-world data often comes in pairs to explore relationships between two quantitative characteristics observed on the same subjects.

• Paired Data: Observations recorded as ordered pairs $(x_i, y_i)$, where $x$ is the explanatory variable (independent) and $y$ is the response variable (dependent).
• Scatter Plot: A two-dimensional graph where each observational unit is plotted as a single point at coordinates $(x_i, y_i)$.

---

Direction and Form of Association

When examining a scatter plot, we first describe its form and direction:

• Form: The overall geometric pattern formed by the points.
  • Linear: Points cluster around a straight trend line.
  • Non-linear: Points follow a curve (e.g., quadratic, exponential).
• Direction: The orientation of the relationship.
  • Positive: As $x$ increases, $y$ tends to increase.
  • Negative: As $x$ increases, $y$ tends to decrease.
  • No Direction: Changes in $x$ show no systematic change in $y$.

---

Strength of Association and Outliers

Beyond direction and form, visual assessment requires checking strength and unusual observations:

• Strength: Measures how closely the points adhere to the underlying form or trend.
  • Strong: Points lie very close to a single imaginary line or curve.
  • Weak: Points display high scatter/variability around the general trend.
• Outliers: Individual points that deviate sharply from the main pattern of the rest of the data.

---

Worked Example: Analyzing Study Hours vs. Exam Scores

Problem: A teacher records weekly study hours ($x$) and exam percentage scores ($y$) for 12 students. Characterize the association between $x$ and $y$ using the scatter plot below. Identify any unusual points.

---

Worked Example: Step-by-Step Solution

Solution Steps:

1. Form: The overall trend follows a straight path. It is Linear.
2. Direction: As study hours increase, exam scores increase. The slope of the trend line is positive, indicating a Positive Association.
3. Strength: Ignoring the single outlier, points cluster very closely around the dashed line, showing a Strong relationship.
4. Outlier: Point A at $(1.5, 92)$ is an outlier — the student achieved a high score despite studying very little.

Common Pitfall Note: Correlation describes visual and statistical association, not causation. A strong positive correlation does not guarantee that increasing study hours alone directly causes high scores for every individual.