Content
Summary
Histograms and frequency polygons are graphical tools used to visualize the shape, center, and spread of large datasets, typically containing 100 or more observations. While histograms use contiguous bars to show distribution, frequency polygons use line graphs connecting midpoints, making them ideal for comparing multiple distributions. Time series graphs differ by preserving chronological order to reveal trends over time.
Intuition
Think of a histogram as a specialized bar chart where the bars touch each other, signaling that the data is continuous or grouped into intervals. This continuity helps you see the overall 'shape' of the data—whether it is symmetric, skewed, or has multiple peaks. Unlike standard bar charts for categories, histograms require careful boundary setting so that data points fall clearly inside the bars rather than on the edges. Frequency polygons offer a similar view but use a line connecting the center of each bar, which makes it much easier to overlay different datasets on the same graph to compare their distributions directly.
Key equations
Relative frequency is calculated by dividing the frequency (f) of a specific value or range by the total number of observations (n).
Derivation
- Identify the dataset size () and determine the number of bars using the square root rule: round to the nearest integer.
- Calculate boundaries to avoid data points landing on edges: for continuous data, subtract/add 0.05 to the min/max if precision is one decimal; for integers, subtract/add 0.5.
- Compute bin width by dividing the range (ending point - starting point) by the number of bars, rounding up to the next convenient number.
- Plot the histogram by drawing contiguous bars where the height represents frequency or relative frequency.
- For frequency polygons, plot points at the midpoint of each bin's horizontal range and connect them with straight lines, adding zero-frequency points at the ends to touch the axis.
Worked example
Construct a histogram for a dataset of 150 continuous height measurements where the smallest value is 60 and the largest is 74, with precision to one decimal place.
- Determine number of bars: , so use 12 bars.
- Set boundaries: Since precision is one decimal, subtract 0.05 from the minimum (60 - 0.05 = 59.95) and add 0.05 to the maximum (74 + 0.05 = 74.05).
- Calculate range: 74.05 - 59.95 = 14.1.
- Calculate bin width: 14.1 / 12 = 1.175. Round up to a convenient number, e.g., 1.2.
- Plot contiguous bars from 59.95 to 74.05 using the calculated bin width.
Answer The histogram uses 12 bars with a width of 1.2, starting at 59.95 and ending at 74.05.
Common mistakes
- Using separated bars for a histogram; they must be contiguous to indicate continuous data.
- Placing data values exactly on bin boundaries, which causes ambiguity in counting.
- Using pie charts to compare data across different time periods, as fixed areas make relative comparisons misleading.
- Ignoring the need for zero-frequency endpoints when constructing frequency polygons.