Table 5. Summary of top 15 variables in large and medium/small areas.
| Variable | Description | Large Metros | Medium/Small Metros | ||
|---|---|---|---|---|---|
| Rank | Correlation | Rank | Correlation | ||
| AA | Percentage of Asian population. | 1 | Negative | 2 | Negative |
| BA | Percent of Black population. | 12 | Mixed | 1 | Negative |
| NH | Percent of non-Hispanic population. | 9 | Positive | 15 | Positive |
| IA | Percent of American Indian, Alaska native population. | 10 | Mixed | 5 | Negative |
| NA | Percent of Native Hawaiian, Pacific Islander population. | 13 | Positive | 6 | Negative |
| Female | Percent of female population. | 7 | Mixed | 4 | Negative |
| Age_1 | Percent of young adults aged below 14 years old. | 6 | Mixed | 14 | Mixed |
| Age_2 | Percent of adolescents aged 15–29 years old. | 4 | Positive | 8 | Negative |
| Age_6 | Percent of elder people aged above 75 years old. | - | - | 7 | Mixed |
| Education_1 | Percent of people with less than a high school degree. | 11 | Mixed | 11 | Positive |
| Education_2 | Percent of people with a high school degree. | 8 | Mixed | 3 | Positive |
| Education_3 | Percent of people with an associate degree. | 14 | Negative | 10 | Negative |
| Unemployment | Percent of unemployed workers in the total labor force. | - | - | 9 | Positive |
| Income | Median household income. | - | - | 13 | Mixed |
| DX90 | Number of days with temperature ≥ 90°F. | 2 | Positive | - | - |
| DX70 | Number of days with temperature ≥ 70°F. | 3 | Positive | - | - |
| HDSD | Seasonal heating degree days. | 5 | Mixed | - | - |
| EMXP | Extreme maximum daily precipitation total within month. | 15 | Mixed | - | - |
| CDSD | Seasonal cooling degree days. | - | - | 12 | Positive |
Note that, positive correlation denotes the relationship between predictor and response variable that changes in the same way (either increasing or decreasing), while negative correlation denotes this relationship changes in the opposite way. Otherwise, a mixed correlation indicates a combination of positive and negative relationship between predictor and response variable.