This exploration looks at wage income in Minnesota as reported in the 2019 American Community Survey (ACS) administered through the US Census Bureau. Since wage income was the focus, only people 20 years old or more were included. The data was obtained from IPUMS USA, an online source of US Census microdata that also provides documentation and harmonization of variables across time periods. IPUMS USA has a web tool for the selection of data samples along with variables of interest. In this case, 41 variables were selected, although not all were used in the analysis. The 2019 ACS sample obtained (filtered to include only people 20 years old or more) consisted of 43,143 cases (i.e., rows). The data from IPUMS USA is already fairly clean -- for example, there were no null values in the variables obtained.
The 2019 ACS is a 1% sample of the US population obtained using techniques of clustering and stratification. The result is a sample with weights in order for the sample to be representative. For analyses concerning wage incomes, only employed wage earners were included in the population of interest, reducing the effective sample size (and operative data frame) to 25,597 cases. Through the use of weights, these cases were taken to be representative of 2,746,113 people in Minnesota in 2019.
Some visualization tools include built-in parameters for including weights in order to make the data graphic representative of the population and some do not. For the latter case, an expanded data frame (having 2,746,113 rows) was created such that there were as many copies of a row from the original data frame as corresponded with the weighting of that row in the original data frame. This expanded data frame could then be fed into the visualization tool to produce a representative data graphic.
Wage Income & Educational Attainment. One of the clearest observations from this investigation was the positive correlation between wage income and educational attainment. Increasing degrees of educational attainment correlated with higher wages on average. Given our background knowledge that many higher paying jobs often require greater education, this is not surprising yet the visual display of this correlation is still striking. Also, given our background knowledge that educational achievements can produce direct affects on salary, it is reasonable to interpret these graphics as showing the causal impact of educational attainment on salary.
Wage Income & Age. Average wages increase from the 20s to the 30s and into the 40s where they plateau and stay until the 60s and then begin to decline in the 60s and thereafter. It was investigated whether the increase in average wages in the sub-50s period can be explained solely through increases in educational attainment over that period. The result was that educational attainment cannot be the sole driving factor. Also, the decline in average wages starting in the 60s has an easy explanation given the increase in retirement starting in that period if people with higher wages leave the work force disproportionately earlier.
Wage Income & Race/Ethnicity. It was observed that Whites and Asians had higher average wages than the other major racial/ethnic categories (Blacks, Hispanics, and American Indians). It was investigated whether differences in educational attainment could explain the difference in average wages and there was evidence that proportionate differences in educational attainment may be a factor.
Wage Income & Marital Status. It was observed that higher average wages were associated with being married with the spouse present. At the other end, with lower average wages, were those who were widowed or never married/single. The high-low difference in median wage incomes between the marital status categories was found to be $$$24,000, a meaningful difference. What accounts for this difference would be an interesting topic to explore.
Wage Income & Sex. Differences in average wage income based on sex (males earning higher wages on average than females) were observed to exist across the various categories of the factors of age, educational attainment, race/ethnicity, and marital status, highlighting the overall fact of difference as fairly robust. Why that difference exists would be an interesting topic to explore.
The presentation focuses on the interaction of race/ethnicity and educational attainment on 2019 wage incomes in Minnesota. The graphics from the exploration were already largely polished, so only a little polishing was necessary.
The 2019 racial/ethnic distribution in Minnesota (of people 20 years old or more) is displayed with a horizontal bar chart.
The distribution of wage incomes by race/ethnicity is presented using box plots and the observation made that Whites and Asians had higher average wages than the other major racial/ethnic categories (Blacks, Hispanics, and American Indians).
Box plots across the increasing levels of educational attainment are used to show the positive relationship between educational attainment and wage income -- namely, increasing degress of educational attainment correspond to higher wages on average.
Box plots clustered by race/ethnicity across the levels of educational attainment are used to show that the aforementioned positive relationship between educational attainment and wage income holds across the major racial/ethnic categories, although there are some local subversions of the overall trend.
The previous observation is also made with a heatmap table of median wage incomes.
A horizontal stacked bar chart is used to compare the proportionate distribution of educational attainment obtained by the major racial/ethnic categories in Minnesota in 2019. Differences in average wage income between Whites and Asians, on one hand, and the other major racial/ethnic categories (Blacks, Hispanics, and American Indians), on the other hand, are observed to coalign with the observable difference in the proportion of educational attainment as it concerns a 4 year college degree and beyond. Roughly 43% of Whites and Asians obtain a 4 year college degree or higher whereas the percentage is in the low 20s for the other three major racial/ethnic groups.