The average salary paid to employees in a job is a representation of the job's internal worth to an organization and external value in the market. Computing an average salary, however, isn't as straightforward as you might think. When you calculate an average salary, calculate the median salary, too. It might be a better statistic to use, especially with smaller sample sizes.
Within an organization, the average salary of a job can be used to gain some understanding of the relative worth of a job compared with others. In salary surveys, the average salary represents an approximate competitive rate of pay, or "going rate," in the market. When a salary survey reports data broken out by company size, industry, state and city, a single job might have several hundred calculations of average salary, one for each breakout group.
Before you calculate the average salary for a job, you must decide how to handle part-time employees. Some organizations exclude part-time employees from the average, because their salaries are determined using a different set of criteria than full-time employees. If you include part-time employees in the average, you must first either convert part-timers to an annual salary or convert full-time employees to an hourly rate. In the latter situation, calculate an average hourly rate and annualize the result.
Differences in Standard Work Weeks
Suppose a company has two controllers in the same job who each earn an annual salary of $80,000. Controller A works in a division with a 40-hour week. Controller B works in a division with a 37.5-hour week. Controller B technically earns more than Controller A because she works fewer hours. However, in reality, the difference in standard work weeks might be irrelevant because they both put in 50 hours per week and are not paid based on the number of hours they work. To reflect the difference in standard work weeks, convert both salaries to an hourly rate; otherwise, average the annual salaries.
The median is usually a better representation than the average for the "going rate" of a job, because it is less sensitive to extreme values than the average. For example, suppose four employees in a job earn $34,000, $35,000, $36,000 and $37,000 per year. A fifth employee earns $54,000 per year because he was grandfathered at that salary. The average salary of $39,200 doesn't accurately represent the going rate for the job because it is skewed by the extremely high salary of one incumbent. The median salary of $36,000, however, is less affected by the outlier.