Understanding Measures of Dispersion and Data Transformation

Measures of dispersion tell us how spread out the data points are in a dataset. When we perform a transformation on the data, such as adding a constant to each observation, it's important to understand how these measures are affected.

Effect of Adding a Constant on Data

Let's consider a dataset with observations \(x_1, x_2, \ldots, x_n\). If we add a constant \(c\) to each observation, the new dataset becomes \(y_1, y_2, \ldots, y_n\), where \(y_i = x_i + c\).

This transformation shifts the entire dataset along the number line. How does this shift impact different measures of dispersion?

Analyzing Specific Measures of Dispersion

Range

The Range is the difference between the maximum and minimum values in the dataset.

Original Range = \(x_{max} - x_{min}\)

New Range = \(y_{max} - y_{min} = (x_{max} + c) - (x_{min} + c) = x_{max} + c - x_{min} - c = x_{max} - x_{min}\)

Adding a constant to each observation does not change the Range.

Standard Deviation

The Standard Deviation measures the typical distance of data points from the mean. It is calculated based on the deviations from the mean.

Let the original mean be \(\bar{x}\). The new mean is \(\bar{y} = \frac{\sum (x_i + c)}{n} = \frac{\sum x_i}{n} + \frac{\sum c}{n} = \bar{x} + c\).

Now, let's look at the deviations from the new mean for the transformed data:

\(y_i - \bar{y} = (x_i + c) - (\bar{x} + c) = x_i + c - \bar{x} - c = x_i - \bar{x}\)

The deviation of each transformed point from the new mean is exactly the same as the deviation of the original point from the original mean. Since standard deviation is calculated using these deviations (specifically, their squares), the standard deviation remains unchanged.

Adding a constant to each observation does not change the Standard Deviation.

Mean Deviation About Mean

The Mean Deviation About Mean is the average of the absolute deviations from the mean.

Original Mean Deviation = \(\frac{\sum |x_i - \bar{x}|}{n}\)

New Mean Deviation = \(\frac{\sum |y_i - \bar{y}|}{n}\)

We found that \(|y_i - \bar{y}| = |x_i - \bar{x}|\). Therefore, the sum of absolute deviations and the mean deviation about the mean remain unchanged.

Adding a constant to each observation does not change the Mean Deviation About Mean.

Coefficient of Variation

The Coefficient of Variation (CV) is a relative measure of dispersion. It is the ratio of the standard deviation to the mean, usually expressed as a percentage:

\(CV = \frac{Standard\ Deviation}{Mean} \times 100\%\)

We know that adding a constant \(c\) to each observation:

• Does not change the Standard Deviation (\(SD_{new} = SD_{original}\)).
• Changes the Mean (\(\bar{y}_{new} = \bar{x}_{original} + c\)).

So, the New Coefficient of Variation is:

\(CV_{new} = \frac{SD_{new}}{\bar{y}_{new}} \times 100\% = \frac{SD_{original}}{\bar{x}_{original} + c} \times 100\%\)

Unless the standard deviation is zero (all data points are the same) or the original mean is zero and the constant is also zero, adding a non-zero constant \(c\) to observations with a non-zero mean \(\bar{x}_{original}\) will change the mean, and thus the Coefficient of Variation will change.

Adding a constant to each observation will change the Coefficient of Variation (assuming the original mean is not zero and the standard deviation is not zero).

Summary Table

Based on this analysis, the measure of dispersion that will change when a constant is added to each observation of a data set is the Coefficient of Variation.

Revision Table: Measures of Dispersion Properties

Additional Information: Location vs. Scale Parameters

Adding a constant to data is a change in the location of the data. Measures of dispersion like Range, Standard Deviation, and Mean Deviation are scale parameters; they are invariant to changes in location. They measure the spread, which doesn't change if you just shift the whole distribution.

Multiplying data by a constant, on the other hand, is a change in the scale. This would affect measures like Range, Standard Deviation, and Mean Deviation, typically multiplying them by the absolute value of the constant. The Coefficient of Variation is the ratio of a scale parameter (SD) to a location parameter (Mean), making it sensitive to changes in both scale and location (via the change in mean).