In the data provided by U.S Census bureau, American community survey, 2011; the difference between mean and median income for varied races is listed. From the table, it is clear that mean income is higher than median income for every race mentioned and also for category with ‘All workers’. This data concludes that there are few extremely high values in each individual category meaning some individuals in each group are getting paid way too higher salary than others. Consequently, this is a positively skewed distribution.

Each group has a higher mean income than median income however, white group has highest mean income than any other group. There is a difference of $20651 between mean income and median income for the same group which is far more than black with $15026 and Latino with $13728 difference in mean and median income. From this analysis we can confirm that most of the higher paid job positions belong to people in white race group.

This table represents interval-ratio variable which is positively skewed with some extreme high values. Using the mean as central tendency would provide misleading information about the ‘average’ income in given group of workers. For example, the central tendency for white race group would show value of $76063 which also considers the extreme high values while if median is used, the income is considered for the (N+1)/2th case which is more appropriate while providing central tendency for such skewed distribution. Considering mode as the measurement for central tendency is not a good option as it will only consider the highest income which will again use the extreme high values of the income and may provide false information. So, I would use median as a measure of central tendency as it is not influenced by the extreme positive values of the income.

You did a really good job breaking everything down by category. I thought median was the only way to go with this data as well.

Thank you Heidi. Yes! In my opinion median is an appropriate measure of central tendency for this data.

Very good analysis of this chart. I think that most people thought along the same lines, noting the amount of difference between the mean and median for the races listed. As the first reply to your blog stated, median was also the best way to display this data. The way the data is displayed on this particular graph was somewhat disparaging to the minorities listed.

Thank you..