Drinking, Driving, and Seatbelts

This dataset contains several variables related to the automobile fatalities, drinking behavior, and seatbelt use. There is no "correct" analysis for these data, but there may be some interesting lessons to be learned. Perhaps the most important lesson is not to jump to conclusions. Some of these relationships may suggest conclusions that are misleading.

The data below contain information on the rate of traffic fatalities per 100 million vehicle miles traveled, the percentage of fatalities that involved blood alcohol levels above .10 g/dl, the percentage who were killed will using (and not using) restraint devices such as seatbelts and airbags, the state's estimated seat belt use rate, and the state's yearly per capita consumption of alcohol.

Most of these data came from the National Highway Traffic Safety Administration. They were found at http://www.nhtsa.dot.gov/people/ncsa/ in a file on State Traffic Data. That location also has a number of other interesting and useful datasets. The Data on Alcohol consumption were for 1989, and were taken from a document entitled "Quick Facts," published by the National Institute of Alcohol Abouse and Alcoholism, 1989.

STATE	RATE	ALC_10	KILLREST	KILLUNRE	BELTUSE	ALCCONS	PERMILES
AL	38.46	34.3	29.8	64.9	52	1.91	2.2
AK	19.76	42.4	42.4	48.5	69	3.52	2.1
AZ	35.77	33.7	36.3	52.5	90	3.15	2.6
AR	35.49	23.4	26.6	62.3	41	1.64	2.4
CA	20.56	31.2	47.9	37	85	3.12	1.5
CO	23.34	35.1	37.3	60.5	56	2.88	1.8
CT	13.52	40.9	30.7	59.2	72	2.8	1.1
DE	23.27	31	38.5	61.5	60	3.13	1.6
DC	15.85	43.8	28.1	56.3	63	4.77	1.7
FL	25.2	31.1	38.1	59.3	59	2.97	2.2
GA	30.54	26.9	27.8	59.2	53	2.44	1.7
HI	17.24	31.5	43.1	35.3	80	2.89	1.6
ID	32.35	26.5	35.6	61	59	2.33	2.1
IL	20.9	34.7	31.9	52.8	69	2.68	1.7
IN	24.58	27.4	32.3	58.2	64	2.15	1.5
IA	27.46	30.1	39.9	49.5	76	2.05	2
KS	24.67	34.4	28.6	58.3	54	1.89	1.8
KY	33.35	26.7	30.1	66.7	52	1.85	2.1
LA	33.65	39.9	27.3	55.7	59	2.43	2.3
ME	20.24	23.6	24.3	68.5	50	2.56	1.5
MD	20.05	26.3	50.6	40.1	70	2.76	1.5
MA	9.85	33.4	22.3	54	53	2.97	0.9
MI	22.91	31.6	41.6	48.5	67	2.57	1.8
MN	21.81	36	38.2	45.2	56	2.56	1.4
MS	53.33	35.2	23	75.6	46	2.05	2.9
MO	32.42	40.6	26.8	63.3	71	2.37	1.9
MT	40.41	36.8	36.8	60.9	70	2.74	2.3
NE	21.9	25.2	30.9	54	64	2.28	1.6
NV	30.72	40.7	41.7	51.2	71	5.07	2.2
NH	13.3	25.4	25	65.3	57	4.52	1.1
NJ	14.06	31.5	39.1	49.9	61	2.78	1.3
NM	41.07	41.7	39.8	57.5	86	2.7	2.3
NY	15.95	24.2	41.7	44.6	72	2.55	1.5
NC	29.58	27.6	44.4	40.8	81	2.16	1.9
ND	16.67	43.9	22.5	60	42	2.4	1.1
OH	18.91	25.2	38.1	60.4	63	2.18	1.4
OK	28.22	30.7	26.2	72.6	46	1.81	1.7
OR	22.24	30.7	55.7	36	80	2.54	1.9
PA	18.03	32.7	27.5	55.5	71	2.23	1.6
RI	9.91	31.5	27.8	69.4	58	2.87	1
SC	34.95	26	38.1	61.3	64	2.5	2.3
SD	20.74	39.7	16.7	66.7	40	2.24	2.1
TN	32.53	33.3	27.6	68.1	64	1.96	2.2
TX	25.97	44.2	44.8	53.6	72	2.63	1.8
UT	25.89	21.2	32	59.2	56	1.58	1.7
VT	23.5	31.2	38.5	56.9	67	3.18	1.7
VA	19.34	30.3	33.7	57.6	70	2.53	1.3
WA	17.1	38	36.8	59.7	83	2.66	1.3
WV	28.64	35.2	29.4	65.2	58	1.84	2.2
WI	20.72	35.3	37.8	54.6	64	3.16	1.4
WY	49.13	37.2	25.6	71.8		2.64	2.4

These data can be used to look at a number of relationships. We can ask about fatalities as a function of alcohol consumption, seatbelt use, and other variables. Contrary to what we might expect, these correlations are not all that high. Can you think of some reasons why this would be the case? Why should we not be surprised that the more compliant a state's population is with seatbelt use, the more fatalities will involve seatbelts? How can we best look at the relationship between fatalities and seatbelt use? (Interestingly, that answer may not involve a correlation coefficient.)

For more data on these relationships, you can go to the Internet address given above and click on any of there reports. You will find some interesting results, and some useful data.