Histograms, Distribution Shapes, Skewness, and Normal Quantile Plots
Key Concept
- A histogram is often easier to interpret than a frequency table for understanding data distribution.
Histogram basics
- Histogram: bars of equal width drawn adjacent (unless gaps in the data).
- Horizontal axis: classes of quantitative data values.
- Vertical axis: frequencies.
- Heights of bars correspond to frequencies.
Relative Frequency Histogram
- Same shape and horizontal scale as a histogram.
- Vertical axis shows relative frequencies.
- Relative frequency for a class i: rf<em>i=Nf</em>i where fi is the class frequency and N is the total number of items.
Uses of a Histogram
- Visually displays the shape of the distribution.
- Shows the center (location) of the data.
- Shows the spread (variability).
- Identifies outliers.
Distribution Shapes
- Bell-shaped (Normal) distribution.
- Uniform distribution.
- Skewed to the right (positive skew).
- Skewed to the left (negative skew).
Normal Distribution
- Data that form a roughly bell-shaped histogram are said to have a normal distribution.
Skewness (1 of 3)
- Skewness: a distribution that is not symmetric and extends more to one side than the other.
- Right-skewed (positive): longer right tail; left side is shorter.
- Left-skewed (negative): longer left tail; right side is shorter.
- Mnemonic: Skewed left resembles toes on the left foot; skewed right resembles toes on the right foot.
- Right-skew (positive): \text{Median} < \text{Mean}
- Symmetric: Mean=Median
- Left-skew (negative): \text{Mean} < \text{Median}
Assessing Normality with Normal Quantile Plots (1 of 5)
- Criteria for a normal distribution: the pattern of points is reasonably close to a straight line and shows no non-straight systematic pattern.
Assessing Normality with Normal Quantile Plots (2 of 5)
- Not Normal if the points do not lie near a straight-line pattern or show a systematic non-straight pattern.
Assessing Normality with Normal Quantile Plots (3 of 5)
- Normal distribution criterion: points are reasonably close to a straight line and no other systematic pattern.
Assessing Normality with Normal Quantile Plots (4 of 5)
- Not Normal if the points do not lie close to a straight line.
Assessing Normality with Normal Quantile Plots (5 of 5)
- Not Normal if the points show a systematic pattern that is not a straight line.
Quick Reference: Visual and Interpretive Cues
- Use histograms to gauge shape, center, spread, and outliers at a glance.
- Use relative frequency histograms to compare distributions with different totals.
- Check skewness to decide whether mean or median is a better measure of center.
- Use normal quantile plots to assess normality before applying methods that assume normality.