Understanding and Comparing Distributions

Displays for Comparing Groups

Statistical displays such as histograms and boxplots are essential tools for visualizing and comparing quantitative data across groups. They help reveal the shape, center, and spread of distributions, allowing for meaningful comparisons.

Describing Distributions: Shape, Center, Spread

• Shape: Refers to the overall appearance of the distribution (e.g., symmetric, skewed, unimodal, bimodal).

• Center: Indicates where most values cluster (commonly measured by the mean or median).

• Spread: Describes the variability of the data (measured by range, interquartile range (IQR), or standard deviation).

• Example: The histogram below shows the average wind speed in Hopkins Memorial Forest in 2011. The distribution is right-skewed, with most days having low wind speeds.

Comparing Groups: Seasonal Wind Speeds

To compare wind speeds between Spring/Summer and Fall/Winter, histograms for each season are examined. Key features to describe include modality, symmetry, and unusual features.

• Modality: Number of peaks in the distribution.

• Symmetry: Whether the distribution is balanced or skewed.

• Unusual Features: Outliers or gaps in the data.

• Example: The histograms below show that winter is substantially windier than summer, and winter wind speeds are more variable.

Summary Statistics for Seasonal Comparison

Summary statistics provide numerical measures for comparing groups. The table below summarizes mean, standard deviation, median, and IQR for summer and winter wind speeds.

Using Histograms to Make Comparisons

When comparing groups with histograms, it is crucial to use the same horizontal scale for each group. This ensures that differences are visually meaningful and not due to scale discrepancies.

• Example: The histograms below compare the number of cigarettes produced by two machines over 30 days. The clarity of differences depends on consistent scales.

The Five-Number Summary

The five-number summary is a concise description of a quantitative variable, consisting of:

• Minimum value

• First quartile (Q1): 25th percentile

• Median: 50th percentile

• Third quartile (Q3): 75th percentile

• Maximum value

Boxplots: Construction and Interpretation

Boxplots are graphical displays based on the five-number summary. They provide a balance of information and simplicity, making them ideal for comparing groups.

Steps to Construct a Boxplot

1. Draw a vertical axis spanning the data range.

2. Mark horizontal lines at Q1, median, and Q3.

3. Connect these lines to form a box.

4. Erect "fences" at Q1 - 1.5(IQR) and Q3 + 1.5(IQR) to identify potential outliers.

5. Draw "whiskers" from the box to the data values within the fences.

6. Mark outliers beyond the fences with special symbols.

Boxplot Example: Average Wind Speed

• Five-number summary: Max = 6.73, Q3 = 2.28, Median = 1.12, Q1 = 0.46, Min = 0.00

• Interpretation: The boxplot shows the spread and center of daily wind speeds, with whiskers and outliers clearly marked.

Comparing Histograms and Boxplots

Histograms provide detailed information about the shape and distribution, while boxplots summarize the data and are ideal for comparing multiple groups side by side.

Boxplots for Comparing Groups

Boxplots are especially useful for comparing several groups, such as monthly wind speeds. They display the median, quartiles, and outliers for each group on a common scale.

Histograms vs. Boxplots

• Histograms: Best for comparing two groups and visualizing detailed distribution shape.

• Boxplots: Ideal for comparing multiple groups, showing summary statistics and outliers.

Outliers

How to Handle Outliers

Outliers are data points that differ significantly from other observations. They require careful attention, as they may indicate errors or important phenomena.

• Investigate: Research the cause of the outlier.

• Data errors: Outliers may result from data entry mistakes, misunderstanding survey questions, confusion about units, or dishonesty.

• Action: Fix errors if possible, justify removal if necessary, or analyze with and without outliers.

Outlier Example: Kentucky Derby Winning Times

Winning times in the Kentucky Derby show a wide range, with some values standing out as outliers. The histogram and boxplot below illustrate this.

Summary Table: Handling Outliers

Additional info: Outliers may be the most informative values in a dataset, revealing unusual events or errors. Proper handling and transparent reporting are essential for credible statistical analysis.