16.2 Data Analysis: Frequency Tables, Dot/Line Plots, Histograms & Box Plots

16.2 Data Analysis: Frequency Tables, Dot/Line Plots, Histograms & Box Plots

Statistical literacy in elementary education enables students to formulate investigative questions, collect and organize empirical data, construct appropriate visual representations, and draw reasoned inferences. To teach data interpretation effectively, educators must distinguish between different data types, master the mechanics of elementary graphical displays, compute summary statistics, and critically evaluate visual arguments for distortion or bias.

---

Typology of Data: Categorical vs. Numerical Variables

Before selecting a graphical display or calculating summary statistics, teachers must classify the underlying variable type:

                                  Data Types
                                      │
             ┌────────────────────────┴────────────────────────┐
             ▼                                                 ▼
     Categorical (Qualitative)                         Numerical (Quantitative)
             │                                                 │
     ┌───────┴───────┐                                 ┌───────┴───────┐
     ▼               ▼                                 ▼               ▼
  Nominal         Ordinal                           Discrete        Continuous
(No Order)     (Rank Ordered)                      (Countable)     (Measurable)

1. Categorical (Qualitative) Data

Categorical data describes qualities, characteristics, or attributes that fall into distinct groups:

• Nominal Data: Categories with no inherent mathematical ranking or quantitative order (e.g., student eye color, favorite ice cream flavor, method of transportation to school, state of birth).
• Ordinal Data: Categories possessing a natural sequence or rank order, but where intervals between ranks are non-numeric or unequal (e.g., t-shirt sizes XS/S/M/L/XL, survey satisfaction ratings: strongly disagree / disagree / neutral / agree / strongly agree, finishing positions in a spelling bee: 1st, 2nd, 3rd).

2. Numerical (Quantitative) Data

Numerical data represents measurable or countable quantities where arithmetic operations are mathematically meaningful:

• Discrete Data: Data values that can only take specific, isolated values—typically non-negative integers resulting from counting (e.g., number of siblings, number of library books checked out, goals scored in a soccer match). You cannot have 2.37 siblings.
• Continuous Data: Data values resulting from measurement along a continuous real-number scale, which can take any value within a given interval depending upon tool precision (e.g., student heights, running times in seconds, mass of a classroom pumpkin, daily high temperatures).

---

Organizing Data: Tally Charts and Frequency Tables

Data organization begins with raw observation tracking:

• Tally Charts: Record counts in groups of five (four vertical marks crossed by a fifth diagonal mark) so totals can be counted quickly by fives.
• Frequency Tables: Tabulate distinct categories or numerical bins alongside their raw count (frequency), relative frequency (fraction or percentage of the total sample), and cumulative frequency (running total across successive intervals).

---

Graphical Displays for Categorical and Part-to-Whole Data

1. Pictographs

A pictograph uses visual symbols or icons to represent data frequencies. Every pictograph must feature an explicit key (legend) indicating the quantitative value of each symbol.

• Key Scaling: If one apple icon represents 4 students, a half-apple icon represents 2 students.
• Instructional Consideration: Ensure students multiply the count of icons by the key value rather than simply counting the symbols as individual units.

2. Bar Graphs

A bar graph displays categorical data using rectangular bars whose lengths or heights are proportional to category frequencies.

• Orientation: Can be vertical or horizontal.
• Discrete Separation: In a standard bar graph, visible spaces must exist between adjacent bars to signify that the categories are distinct qualitative entities rather than continuous numerical intervals.
• Double Bar Graphs: Place two contrasting bars side-by-side for each category to compare bivariate categorical data (e.g., comparing boys' versus girls' favorite school subjects).

3. Circle Graphs (Pie Charts)

A circle graph displays how a total quantity is partitioned into proportional component parts. The entire circular disk represents 100% of the dataset or 1.00 total probability.

• Central Angle Calculation: Because a circle contains 360°, each slice's central angle is proportional to its percentage of the total: $\text{Central Angle (Degrees)} = \left(\frac{\text{Category Frequency}}{\text{Total Sample Size } N}\right) \times 360^\circ = \text{Percentage (as decimal)} \times 360^\circ$

• Worked Example: In the transportation table above, 9 out of 25 students ride the bus (36%). Calculate the central angle for the bus sector: $\text{Central Angle} = 0.36 \times 360^\circ = 129.6^\circ$

---

Graphical Displays for Continuous Trends and Numerical Distributions

1. Line Graphs

A line graph plots individual data points on a Cartesian coordinate plane and connects consecutive points with line segments. Line graphs are specifically suited for tracking continuous changes over time (time-series data), such as hourly temperature shifts or plant growth across 8 weeks.

• Critical Distinction: A line graph should never be used for unrelated categorical items (e.g., connecting bars of favorite pets with a line), because intermediate points between distinct categories have no physical meaning.

2. Dot Plots (Line Plots)

A dot plot (historically called a line plot in elementary standards) displays individual numerical data values as dots, Xs, or circles stacked vertically above a continuous number line.

• Visual Diagnostic: Dot plots immediately reveal distribution shape, clusters (groupings of data), peaks (modes), gaps (intervals with zero data), and isolated outliers.

Number of Books Read by Students During Reading Month:

         * 
         *        * 
         *   *    *   * 
     *   *   *    *   *        *
   ──┼───┼───┼────┼───┼───┼────┼───
     1   2   3    4   5   6    7

• Looking at this dot plot: The mode is 2 books (peak of 4 dots). There is a gap at 6 books. A cluster appears between 1 and 5 books.

3. Stem-and-Leaf Plots

A stem-and-leaf plot organizes numerical data into leading place-value digits (stems) and trailing place-value digits (leaves).

• Preservation of Raw Data: Unlike histograms, stem-and-leaf displays retain every original data value while presenting a distribution shape equivalent to a horizontal histogram.
• Leaf Ordering: Leaves must always be sorted in ascending order from the stem outward, with equal spacing.
• Mandatory Key: Every stem-and-leaf plot must include an explicit decoding key to communicate place value (e.g., 4 | 3 = 43 points, or 4 | 3 = 4.3 cm).

Worked Example: Test scores from an elementary math assessment: Data: 58, 62, 65, 71, 74, 74, 78, 83, 85, 87, 88, 92, 96, 99.

  Stem │ Leaves
  ─────┼────────────────
     5 │ 8
     6 │ 2  5
     7 │ 1  4  4  8
     8 │ 3  5  7  8
     9 │ 2  6  9

  Key: 7 │ 1 = 71 points

4. Histograms

A histogram displays continuous numerical data that has been binned into contiguous, uniform numerical intervals (class intervals).

• Histograms vs. Bar Graphs (Critical Conceptual Differences):
  1. Data Type: Bar graphs represent discrete categorical groups; histograms represent continuous numerical ranges.
  2. Bar Spacing: In a histogram, bars must touch without gaps between adjacent intervals, visually communicating the continuity of the underlying scale (a gap appears only when an interval frequency is zero).
  3. X-Axis Labeling: Bar graphs label distinct category names; histograms label continuous interval boundaries (e.g., 60–69, 70–79, 80–89).

---

Box-and-Whisker Plots: Five-Number Summary and Outlier Analysis

A box-and-whisker plot (box plot) visually summarizes a numerical dataset by partitioning ordered observations into four equal sections (quartiles), each containing exactly 25% of the data.

            ┌───────┬───────────────┐
      ├─────┤       │               ├───────┤        *
   ───┼─────┼───────┼───────────────┼───────┼────────┼───
     Min   Q1    Median (Q2)       Q3      Max    Outlier

The Five-Number Summary

1. Minimum (Min): The lowest observed value that is not an outlier.
2. First Quartile (Q₁): The 25th percentile; the median of the lower half of data.
3. Second Quartile / Median (Q₂): The 50th percentile; the physical middle score of the complete ordered dataset.
4. Third Quartile (Q₃): The 75th percentile; the median of the upper half of data.
5. Maximum (Max): The highest observed value that is not an outlier.

Computing Quartiles and the Interquartile Range (IQR)

• Interquartile Range: The spread of the middle 50% of the data: $\text{IQR} = Q_3 - Q_1$

The 1.5 × IQR Outlier Identification Rule

An outlier is an extreme observation that deviates significantly from the overall pattern of the distribution. In formal statistics and elementary educator assessments, outliers are identified using the standard mathematical fences:

• Lower Outlier Fence: Any data value strictly less than Q₁ - (1.5 × IQR)
• Upper Outlier Fence: Any data value strictly greater than Q₃ + (1.5 × IQR)

Complete Worked Example of a Five-Number Summary and Outlier Check

Dataset (N = 11 ordered scores): 52, 68, 72, 74, 76, 80, 84, 86, 88, 90, 98.

• Step 1: Find the Median (Q₂) With N = 11 (an odd count), the median is the 6th score: $\text{Median } (Q_2) = \mathbf{80}$

• Step 2: Find the First Quartile (Q₁) The lower half consists of the 5 scores strictly below the median: 52, 68, 72, 74, 76. The median of this lower half is the middle score: $Q_1 = \mathbf{72}$

• Step 3: Find the Third Quartile (Q₃) The upper half consists of the 5 scores strictly above the median: 84, 86, 88, 90, 98. The median of this upper half is the middle score: $Q_3 = \mathbf{88}$

• Step 4: Compute the Interquartile Range (IQR) $\text{IQR} = Q_3 - Q_1 = 88 - 72 = \mathbf{16}$

• Step 5: Calculate Outlier Fences

  • 1.5 × IQR = 1.5 × 16 = 24
  • Lower Fence = Q₁ - 24 = 72 - 24 = 48
  • Upper Fence = Q₃ + 24 = 88 + 24 = 112

• Step 6: Evaluate Extreme Values

  • The lowest score is 52. Since 52 ≥ 48 (it is above the lower fence), 52 is NOT an outlier.
  • The highest score is 98. Since 98 ≤ 112 (it is below the upper fence), 98 is NOT an outlier.
  • Five-Number Summary: Min = 52, Q₁ = 72, Median = 80, Q₃ = 88, Max = 98.

---

Comparison of Elementary Data Displays

---

Identifying Misleading Graphs

Media reports, marketing claims, and standardized assessments frequently feature graphs engineered to mislead viewer interpretation. Elementary educators must teach students to detect four common visual distortions:

1. Truncated Vertical Axis (Broken Baseline):

   • The Distortion: The vertical frequency axis begins at a non-zero number (e.g., starting at 80 instead of 0) or includes an axis break (zig-zag symbol) that is easily overlooked.
   • Visual Effect: Exaggerates minor differences between categories. A bar representing 82 and a bar representing 88 will appear four times taller than each other, creating the false impression of a dramatic change when the real increase is only about 7%.

2. Non-Uniform Scale Intervals:

   • The Distortion: The axis increments are inconsistent (e.g., jumping from 0 to 10, then 20, then suddenly 50, 100, 500 without changing visual spacing).
   • Visual Effect: Distorts the slope of line graphs or flattens rapid exponential increases into seemingly linear trends.

3. Violations of the Principle of Proportional Ink (Pictograph Area/Volume Distortion):

   • The Distortion: When data doubles, the illustrator doubles both the height and the width of a 2D icon, or doubles the height, width, and depth of a 3D icon.
   • Visual Effect: Doubling both dimensions of a 2D symbol quadruples the visual surface area (2² = 4). Doubling all three dimensions of a 3D symbol octuples the volume (2³ = 8). The viewer perceives a four- to eight-fold increase rather than a two-fold doubling.

4. Three-Dimensional Perspective Distortion:

   • The Distortion: Circle graphs or bar charts rendered with dramatic 3D angles or tilted perspective.
   • Visual Effect: Slices positioned at the "front" of a tilted 3D pie chart appear significantly larger to the human eye than equal or larger slices placed at the "back" due to perspective foreshortening.

---

Classroom Error Analysis & Diagnostic Scenarios

Scenario 1: The Histogram Gap Error

• Student Work: A fifth-grader constructing a histogram of student reading times (0–14 min, 15–29 min, 30–44 min) draws bars with wide spaces between them, identical to a categorical bar graph.
• Diagnostic Assessment: The student does not understand that histograms represent contiguous numerical intervals along a continuous scale. The spaces imply discrete qualitative categories rather than continuous measurement.
• Instructional Remedy: Guide the student to examine the interval endpoints. Since the scale measures continuous time, the bars must touch to show that as soon as one interval ends, the next begins immediately. Emphasize: Bar graphs have spaces because apples and oranges don't touch; histograms have touching bars because continuous numbers flow without interruption.

Scenario 2: The Whisker Length Misconception in Box Plots

• Student Work: Looking at a box-and-whisker plot where the right whisker (from Q₃ to the Max) is twice as long as the left whisker (from the Min to Q₁), a student states: "There are twice as many people in the right whisker as in the left whisker."
• Diagnostic Assessment: The student confuses whisker length (variability/spread) with data frequency (sample size).
• Instructional Remedy: Reinforce the definition of quartiles. Every segment of a box-and-whisker plot—the left whisker, the left box half, the right box half, and the right whisker—contains exactly 25% of the data points. A longer whisker does not mean more people; it means those 25% of people have scores that are spread out over a wider numerical range (lower data density).