Dot plots, histograms, box plots Knowledge-Based Study

Study notes

1. Visualizing Numerical Data

   When collecting quantitative data, summarizing the values using visual representations makes it easier to observe patterns, spread, clustering, and center. Three foundational graphical displays used for numerical data are:

   • Dot plots: Best for small, discrete datasets where individual values need to be seen.
   • Histograms: Best for continuous or large datasets grouped into equal intervals.
   • Box plots: Best for summarizing distribution shapes and identifying five-number summaries and outliers.

   Choosing the correct display depends on the size of the dataset, whether the data is discrete or continuous, and which summary features need emphasis.

   Understanding Dot Plots

   A dot plot records data points as dots positioned vertically above a continuous number line. Each dot corresponds to one observation.

   Key Features:

   • Individual observations: Every single data value is visible and accounted for.
   • Frequency visualization: Columns of stacked dots indicate the frequency of each value, making modes immediately visible.
   • Clusters and gaps: Groups of values clustered together or wide empty intervals are readily apparent.

   Example Walkthrough:

   Consider the number of books read by 10 students: 1, 2, 2, 3, 3, 3, 4, 4, 5, 8.

   • The number line extends from at least 1 to 8.
   • Above 3, there is a stack of 3 dots, representing the mode.
   • There is an isolated dot at 8, revealing a gap between 5 and 8, which points to a possible outlier.

   Understanding Histograms

   A histogram groups numerical data into continuous, non-overlapping intervals called bins or classes. The height of each bar represents the frequency (or relative frequency) of data points falling within that specific interval.

   Key Characteristics:

   • Continuous scale: The horizontal axis represents a continuous numerical scale, so adjacent bars touch each other without gaps unless an interval has a frequency of zero.
   • Equal bin widths: Bin intervals must be of equal size (for example: 0 to under 10, 10 to under 20, 20 to under 30).
   • Boundary rules: Values falling precisely on a bin boundary must follow a consistent convention, typically $[a, b)$, where the lower endpoint is included and the upper endpoint is excluded.

   Distinguishing Histograms from Bar Charts:

   While bar charts display categorical data with spaces separating distinct categories, histograms display continuous numerical data with adjacent bars spanning continuous ranges.

   Understanding Box Plots

   A box plot (or box-and-whisker plot) provides a five-point summary of a dataset, emphasizing central tendency and spread without showing individual values directly.

   The Five-Number Summary:

   1. Minimum: The lowest data value that is not an outlier.
   2. First Quartile ($Q_1$): The 25th percentile; 25% of the data lies below this value.
   3. Median ($Q_2$): The 50th percentile, splitting the ordered dataset in half.
   4. Third Quartile ($Q_3$): The 75th percentile; 75% of the data lies below this value.
   5. Maximum: The greatest data value that is not an outlier.

   Interquartile Range and Outliers:

   The width of the central box represents the Interquartile Range ($IQR$):

   $IQR = Q_3 - Q_1$

   A data value $x$ is typically flagged as an outlier if:

   • $x < Q_1 - 1.5 \times IQR$, or
   • $x > Q_3 + 1.5 \times IQR$

   Outliers are plotted as individual detached points, while the whiskers extend to the most extreme data values within the non-outlier limits.

   Analyzing Distribution Shapes

   All three displays enable statisticians to classify distribution shapes:

   • Symmetric (Bell-shaped or Uniform): The mean and median are roughly equal, and the display mirrors itself around the center. In a box plot, the median line sits centrally inside the box, and both whiskers have similar lengths.
   • Skewed Right (Positively Skewed): A tail extends toward larger values. The mean is typically greater than the median. In a box plot, the right whisker and right half of the box are noticeably longer.
   • Skewed Left (Negatively Skewed): A tail extends toward smaller values. The mean is typically less than the median. In a box plot, the left whisker and left half of the box are noticeably longer.

   Choosing the Best Plot:

   • Use a dot plot when the sample size is small ($n < 30$) and exact values matter.
   • Use a histogram when dealing with moderate to large datasets to reveal modality and general shape across intervals.
   • Use a box plot when comparing two or more groups side-by-side or analyzing quartiles and extreme values.