Question

I need to test a hypothesis regarding drunk driving: i.e., that a higher blood alcohol content of 0.08+ causes a higher rate of drunk driving fatalities, than a lower blood alcohol content of 0.01-0.07. This is to prove that the American BAC level should be lowered from 0.08 to a range of 0.01-0.07 to reduce drunk driving fatalities. I chose a random sample of 10 U.S. states: 1) Total Fatalities for 2016 and 2015, 2) Total fatalities with a driver’s BAC in the range of 0.01-0.07 for both years, and 3) Total fatalities with a driver’s BAC >/= 0.08. Also, if someone wouldn’t mind helping me test this hypothesis, step-by-step, that would be a huge help.

Can I use this as valid null and alternative hypotheses?

Ho : 0.08+ Blood Alcohol Content fatalities = 0.01-0.07 Blood Alcohol Content fatalities

Ha : 0.08 Blood Alcohol Content fatalities > 0.01-0.07 Blood Alcohol Content fatalities

Here is the data:

State	Total Driver Fatalities 2016	Total Driver Fatalities 2015	BAC .01-0.07 2016			BAC .01-0.07 2015				BAC .08+ 2016	BAC .08+ 2015
Alabama	1,352	1,180	406			271				351	232
California	5,020	4,382	1,306			1,005				1,053	853
Florida	4,580	4,137	1,032			877				850	736
Illinois	1,566	1,357	491			350				411	291
Kansas	556	465	128			92				103	79
Louisiana	1,067	998	415			276				341	234
Maine	206	190	68			66				60	52
Massachusetts	501	409	188			118				149	95
New Jersey	826	756	247			135				190	107
Texas	5,257	4,836	1,469			1,525				1,247	1,282
Mean	2,093	1,871	575			471.5				475.5	396.1

Answer 1

Lets perform the test in excel for the year 2015

We enter the data in excel and then goto data > data analysis tab and select t test

Please note that this is is independent samples t test . The hypothesis can be as follows

H0 : the American BAC level is not lowered from 0.08 to a range of 0.01-0.07 to reduce drunk driving fatalities

H1 : the American BAC level is lowered from 0.08 to a range of 0.01-0.07 to reduce drunk driving fatalities

BAC.01-0.07 2016 406 1,306 1,032 491 128 415 BAC.08+ 2016 351 1,053 850 411 103 341 Analysis Tools OK Histogram Moving Average Random Number Generation Rank and Percentile Regression Sampling t-Test: Paired Two Sample for Means t-Test: Two-Sample Assuming Equal Variances t-Test: Two-Sample Assuming Unequal Variances z-Test: Two Sample for Means Cancel Help 247 1,469 575 149 190 1,247 475.5

t-Test: Two-Sample Assuming Equal Variances ?23 Input Variable 1 Range: Variable 2 Range: Hypothesized Mean Difference: OK SAS1:SA$12 Cancel SBS1:SBS12 Help Labels Alpha: 10.05 Output options Output Range: O New Worksheet Ply: New Workbook

t-Test: Two-Sample Assuming Equal Variances BAC.01-0.072 BAC.08+ 2016 475.5 160622.85 575 Mean Variance Observations Pooled Variance Hypothesized Mean Difference df t Stat P(T<-t) one-tail t Critical one-tail P(T<=t) two-tail t Critical two-tail 231787.4 196205.1 0 20 0.526804 0.302062 1.724718 0.604124 2.085963

as the p value is 0.302 , which is not less than alpha of 0.05 , hence we fail to reject the null hypothesis and conclude that the American BAC level is not lowered from 0.08 to a range of 0.01-0.07 to reduce drunk driving fatalities

Remember that this is a directional test as we are interested in the "lowered" part of the question