MOD 3.3 | VID 7 | Comprehensive Study Guide on Outlier Detection and the Interquartile Range Method

Outlier Definitions and Sources

• Definition of an Outlier:
  • An outlier is a data value that is considerably larger or smaller than most of the other values in a data set.
• Causes and Sources of Outliers:
  • Data Entry or Measurement Errors: Outliers frequently occur due to operational errors, such as a misplaced decimal point, which causes a recorded number to become vastly larger or smaller than the surrounding data.
  • Natural Population Extremes: Outliers can also be correct, valid measurements that reflect the true presence of extreme values within a population distribution.
• Corrective Action:
  • If an outlier is confirmed to be an error, it should be corrected or removed from the data set whenever possible.

Temperature Data Case Study

• Data Context:
  • Summer temperature measurements were recorded in Fahrenheit ($^\circ\text{F}$) at a downtown location across eight consecutive days.
• Dataset Observations:
  • Measured values: $81.2$, $85.6$, $89.3$, $91.0$, $83.2$, $8.45$, $79.5$, and $87.8$.
• Outlier Identification:
  • Identified Outlier: $8.45$
• Diagnostic Analysis:
  • The value $8.45$ is unquestionably an error resulting from a misplaced decimal point (where the intended value was likely $84.5$).
  • The value must be corrected prior to performing further statistical analysis.

The Interquartile Range (IQR) Method for Outlier Detection

• Interquartile Range (IQR) Definition:
  • The interquartile range measures the spread of the middle $50\%$ of a dataset and is calculated as the mathematical difference between the third quartile ($Q_3$) and the first quartile ($Q_1$).
  • Mathematical Formula: $IQR = Q_3 - Q_1$
• Systematic 4-Step IQR Procedure for Detecting Outliers:
  • Step 1: Calculate the first quartile ($Q_1$) and third quartile ($Q_3$) of the dataset. Technology such as the TI-84 Plus calculator can be utilized to obtain these values.
  • Step 2: Compute the interquartile range using $IQR = Q_3 - Q_1$.
  • Step 3: Establish the outlier boundaries, which serve as explicit cutoff thresholds:
  • Lower Outlier Boundary Formula: $\text{Lower Boundary} = Q_1 - 1.5 \times IQR$
  • Upper Outlier Boundary Formula: $\text{Upper Boundary} = Q_3 + 1.5 \times IQR$
  • Step 4: Screen all dataset values against the boundaries. Any observation strictly less than the lower outlier boundary or strictly greater than the upper outlier boundary is classified as an outlier.

Middle School Absence Case Study

• Data Context:
  • Attendance monitoring tracking the number of daily student absences at a middle school located in Northwestern Montana for every school day during the month of January.
• Summary Statistics:
  • First Quartile ($Q_1$): $45$
  • Third Quartile ($Q_3$): $59$
• Interquartile Range Computation:
  • $IQR = Q_3 - Q_1 = 59 - 45 = 14$
• Boundary Computations:
  • Lower Outlier Boundary: $Q_1 - 1.5 \times IQR = 45 - 1.5 \times 14 = 45 - 21 = 24$
  • Upper Outlier Boundary: $Q_3 + 1.5 \times IQR = 59 + 1.5 \times 14 = 59 + 21 = 80$
• Outlier Evaluation:
  • Minimum Threshold Test: No absence count in the dataset is less than the lower boundary of $24$.
  • Maximum Threshold Test: The observed daily absence count of $100$ exceeds the upper boundary of $80$
  • Conclusion: The value $100$ is definitively identified as an outlier.