12: Dispersion

12.1 Measuring Variability

Dispersion refers to how data points in a dataset vary in relation to a measure of central tendency, such as the mean, median, or mode. While central tendency summarizes the center of a distribution, measures of dispersion describe the spread or variability of the data. Dispersion is essential for understanding how consistent or scattered data points are around the central value. High dispersion indicates that the data points are widely spread, while low dispersion suggests they are closely clustered.

Understanding dispersion is crucial in quantitative research because it informs how precise estimates are and how much observed values deviate from typical values. It also supports accurate interpretation of patterns and differences within datasets. The most common measures of dispersion include standard deviation, interquartile range (IQR), variance, range, and frequencies (for nominal variables).

12.2 Standard Deviation

The standard deviation is the square root of the variance and is one of the most commonly used measures of dispersion. It provides an intuitive sense of variability because it is expressed in the same units as the original data, making it easier to interpret than variance. The standard deviation indicates how far, on average, each data point lies from the mean of the dataset. It is most appropriate when the data is continuous and approximately normally distributed.

Standard deviation is especially useful when you need a measure of variability that accounts for every value in the dataset. However, it is sensitive to outliers, which can inflate the standard deviation and make the data appear more variable than it actually is for most observations.

12.3 Interquartile Range (IQR)

The interquartile range (IQR) measures the spread of the middle 50% of a dataset. It is calculated by subtracting the first quartile from the third quartile, capturing the range within which the central half of the data lies. The IQR is a robust measure of dispersion because it is less affected by outliers than the full range, making it especially useful for skewed datasets or those with extreme values. It is commonly used when the goal is to focus on typical variability while excluding the influence of outliers.

The IQR is particularly valuable when the standard deviation is misleading due to skewness, as it provides a clearer picture of the central spread without being distorted by extreme values.

12.4 Variance

Variance measures the average squared deviation of each data point from the mean. It reflects how much the data values differ from the mean on average, but because the deviations are squared, the result is expressed in squared units, which makes interpretation less intuitive. For this reason, variance is often not reported as part of basic descriptive statistics, even though it is a foundational concept in statistics.

12.5 The Range

The range is the simplest measure of dispersion, calculated by subtracting the minimum value in a dataset from the maximum value. It provides a quick sense of the overall spread between the lowest and highest values. However, the range is highly sensitive to outliers, meaning that a single extreme value can greatly distort the result.

The range is most appropriate for small or clean datasets where outliers are not present. In larger or more complex datasets, the range may be misleading and is typically supplemented with more robust measures like the interquartile range or standard deviation.

12.6 Nominal Variable Dispersion

For nominal variables, which are categorical and have no inherent order, frequencies serve as a practical way to assess dispersion. Frequencies represent the count of occurrences for each category in the dataset, helping researchers understand how the data is distributed across categories, whether it is concentrated in one or a few categories or spread more evenly.

A high frequency in a single category indicates low dispersion, as most of the data fall into that category. In contrast, a more even spread of frequencies suggests higher dispersion, with the data more equally divided among categories. Frequencies provide insight into the variability of categorical data and help identify how concentrated or diverse the responses are.

12.7 Dispersion in Jamovi

Jamovi can generate values for key measures of dispersion such as standard deviation, interquartile range (IQR), variance, and range, allowing you to assess how much your data varies around the central tendency.

Understanding the Output

The output from the standard deviation, IQR, variance, and range is shown below.

To interpret measures of dispersion, begin by reviewing the sample size (N) and the number of missing values, as these affect how stable and representative the variability estimates may be. Larger sample sizes generally provide more reliable estimates of dispersion.

Next, examine the standard deviation, which indicates how far scores typically deviate from the mean. A larger standard deviation reflects greater spread in the data, while a smaller value suggests that scores cluster more closely around the mean.

The variance represents the average squared deviation from the mean. Although it is less intuitive to interpret directly (because it is expressed in squared units), it provides a mathematical foundation for the standard deviation. Larger variance values indicate greater variability.

The interquartile range (IQR) describes the spread of the middle 50% of scores. Because it is based on quartiles rather than the mean, it is less affected by extreme values and is especially useful when the distribution may be skewed.

The range reflects the difference between the highest and lowest observed values. While it provides a quick snapshot of overall spread, it is highly sensitive to outliers.

When interpreting dispersion statistics together, compare the standard deviation and IQR to determine whether variability is consistent across the distribution or influenced by extreme scores. Examining multiple measures of dispersion provides a more complete understanding of how spread out the data are.

Computing Frequencies

For nominal variables, Jamovi provides a frequency distribution that helps you assess the variability of categorical data.

Understanding the Output

The output from the frequencies is shown below.

To interpret a frequency table for a nominal variable, begin by examining the counts, which indicate how many participants fall into each category. This provides a raw snapshot of how the sample is distributed.

Next, review the percentage of total, which shows the proportion of the sample represented by each category. Percentages are especially helpful for comparing groups because they standardize the counts relative to the total sample size.

Finally, consider the cumulative percentage, which shows the running total as categories are added together in order. For dichotomous variables (those with two categories), the cumulative percentage will reach 100% after the final category. For variables with more categories, cumulative percentages help illustrate how the distribution builds across categories.

Together, counts and percentages allow you to describe the overall composition of the sample and determine whether one category is more prevalent than another.

12.8 Choosing the Right Measure of Dispersion

The choice of dispersion measure depends on the nature of the data and the type of analysis being conducted. For continuous data with symmetric distributions (without outliers), the standard deviation is typically preferred, as it provides a detailed view of how data points vary around the mean. For skewed distributions or datasets with outliers, the interquartile range (IQR) is often more informative because it focuses on the central 50% of the data and is less affected by extreme values. In research, it is often helpful to report both standard deviation and IQR when analyzing continuous variables. This dual approach gives a more complete picture of variability, especially when the data are not normally distributed. For nominal data, frequencies are the appropriate measure, indicating how evenly or unevenly the data is distributed across categories.