Variable Set #1:
For this set, I looked at the correlations between the death rate due to motor vehicle accidents per 100k people in 1997 (Code #62, CarDeath 97), the percent of the population uninsured from 1999-2000 (Code #83, %No Ins 00), the median age according to the 2000 census (Code #8, Med Age 2000), and the number of males per 100 females in 2000 (Code #10, SexRatio 00). The CarDeath 97 has a moderate, positive correlation of 0.39** with %NoIns 00. This was not surprising to me because if you don’t have insurance, you are less likely to be able to afford care if you are in a car accident. The %NoIns 00 has a moderate, negative correlation of -0.38** with the Med Age 2000 (not surprising because as you get older, you find it more and more important to get health insurance) and a moderate, positive correlation of 0.37** with SexRatio 00 (unsurprising because men are more likely to think they are resilient to injury, and therefore do not need health insurance). Med Age 2000 has a moderate, negative correlation of -0.45** with SexRatio 00. This was not surprising because females live longer than males. There was no correlation between Med Age 2000 and CarDeath 97 (surprising because usually, younger people get into more car accidents), and CarDeath 97 and SexRatio 00 (unsurprising because there is no correlation between sex and the ability to drive well).
I chose to use CarDeath 97 as my dependent variable because it can be influenced by a number of different factors. After running my regression analysis, I found that when I statistically controlled for Med Age 2000 and SexRatio 00, the predicted CarDeath 97 is expected to increase .42 units for every 1 unit increase in %NoIns 00 (Beta = 0.42**). When I statistically controlled for %NoIns 00 and SexRatio 00, the predicted CarDeath 97 is expected to decrease by .08 units for every 1 unit increase in Med Age 2000 (Beta =-.08). When I statistically controlled for Med Age 2000 and %NoIns 00, the CarDeath 97 is expected to decrease by .14 units for every 1 unit increase in SexRatio 00 (Beta = -.14). %NoIns 00, Med Age 2000, and SexRatio 00 account for 17.1% of the variation in CarDeath 97.

Variable Set #2:
For this set, I looked at the correlations between the Marriage rate per 1000 population in 1998 – without Nevada (79.5) (Code #19, Marry Rt 98), the number of teen births per 1000 in 1999 (Code #50, TeenBirth99), the percent of HS grads immediately enrolled in 2-yr or 4-yr colleges in 1998 (Code #87, % College 98), and the percent of the population that is black as of 2000 (Code #24, % Black 00). The Marry Rt 98 has a weak, positive correlation of 0.33* with the TeenBirth 99. This was somewhat surprising to me because I thought there would be more marriage based on the number of teens having babies. Marry Rt 98 has a weak, negative correlation of -0.35* with % College 98. This was interesting to me because as less people got married, more were seeking a college education. The % Black 00 has a strong, positive correlation of 0.54*** with Teen Birth99. I was somewhat sad to hear this because I know that people of color often have less access to resources that white people do, and to me, this looks like minorities are not receiving as much access to sexual education and birth control. There was no correlation between Marry Rt 98 and % Black 00 (not surprising because I have not seen any literature on race being a factor in marriage rate), % Black 00 and % College 98 (this was surprising because of the research done on people from different ethnicities attending college), or Teen Birth99 and % College 98 (surprising because usually, teen births result in deferring higher education).
I decided to make marriage rate my dependent variable because it had both positive and negative correlations. After running the regression analysis, I found that when I statistically controlled for % College 98 and Teen Birth99, Marry Rt 98 is predicted to decrease 0.04 units for every 1 unit increase in % Black 00 (Beta = -0.04). When I statistically controlled for % Black 00 and % College 98, Marry Rt 98 is predicted to increase 0.30 units for every 1 unit increase in Teen Birth99 (Beta = 0.30). When I statistically controlled for % Black 00 and Teen Birth99, Marry Rt 98 is predicted to decrease 0.28 units for every 1 unit increase in % College 98 (Beta = -0.28). % College 98, Teen Birth99, and % Black 00 all account for 19.3% of the variation in Marry Rt 98.