Symmetry
This video discusses symmetry and skewness as they relate to the shape of a distribution. We use histograms, stem plots, and box plots to talk about these concepts.
A distribution is symmetrical if it can be divided into two halves of equal size and the same shape (mirror images).
In a symmetrical distribution, the plane of symmetry runs through the center, and the mean is the balance point that lies at the center as well.
In the example given, the symmetric case has both the mean and the median equal to the central value:
The mean is described as the balance point of a distribution.
Skewness
Skewness refers to asymmetry in a distribution.
Distributions can be skewed to the left or to the right.
Skewness is interpreted by the direction in which the data cluster or by the tail of the distribution.
A distribution is skewed to the left if it has a long tail that trails toward the left.
A distribution is skewed to the right if it has a long tail that trails toward the right.
The direction of skewness can be read from the plot (histogram, stem plot, box plot) to determine where the data are more spread out.
Stem plots and skewness
The same left/right skew concepts apply to stem plots.
A stem plot can be judged for skewness by flipping it onto its side to resemble a number line, with the leftmost value being the smallest and values increasing to the right.
If the stem plot has a longer tail on the right, it is skewed to the right; if it has a longer tail on the left, it is skewed to the left.
Boxplots and skewness
When determining skewness from a boxplot, outliers can affect interpretation.
Example: a regular boxplot might suggest left skew, but a modified boxplot could reveal right skew for the same data.
A strategy for determining skew in boxplots:
If the two boxes (quartile boxes) are unequal in size, the side of the box that is larger indicates the direction of skew.
If the boxes are equal in size, then look at the whiskers to determine skewness. The longer whisker indicates the skew direction (left or right).
If the boxes are equal in size and the whiskers are also equal in length, the distribution is symmetrical.
Practical example with a histogram
Histograms show frequency by bars.
By counting bars to the right and left of the central value (12 in the example), we can infer skewness.
In the given description, to the right of 12 there are more data values than to the left of 12, implying a shift in the center.
The median in this skewed example is located between 16 and 18, i.e., 16 < \tilde{x} < 18.
The mean, being the balance point, will be influenced by skewness and move toward the tail.
Mean and median in relation to symmetry and skewness
In a symmetrical distribution:
The plane of symmetry is at the median (the middle data point).
The mean is the balance point, so the mean equals the median:
In the symmetric example, both are equal to a central value (e.g., 12):
In a skewed distribution, the mean and median are no longer equal:
Skewed to the left (tail toward the left): the mean is less than the median, i.e., \mu < \tilde{x}.
The mean tends to be pulled toward the left tail; the median sits closer to the right side of the distribution.
Skewed to the right (tail toward the right): the mean is greater than the median, i.e., \mu > \tilde{x}.
The mean tends to be pulled toward the right tail; the median sits closer to the left side of the distribution.
Summary relationships:
Symmetric:
Left-skewed: \mu < \tilde{x}. (mean closer to left tail, median closer to right side)
Right-skewed: \mu > \tilde{x}. ean closer to right tail, median closer to left side)
Key takeaways
Use histograms, stem plots, and box plots to assess symmetry and skewness.
Skewness is about asymmetry and is read from the direction of the tail or the concentration of data.
Boxplots require careful interpretation: unequal boxes point to skewness via the larger side; equal boxes require examining whiskers.
The relationship between mean and median reveals skewness: symmetry implies , left-skew implies \mu < \tilde{x}, and right-skew implies \mu > \tilde{x}.
In an example with central value 12, a symmetric distribution has both mean and median at 12; a skewed distribution shifts these values with the median landing between 16 and 18 in the described histogram.