Statistics: A matter of interpretation?

Deaths among bicyclists younger than 16 have declined 84 percent since 1975, while deaths among bicyclists 16 and older increased 80 percent.

Percent of bicyclist deaths by age, 1975-2007
Year	<16		≥ 16		Total*
	Num	%	Num	%	Num
1975	675	67	323	32	1,003
1976	605	66	306	33	914
1977	613	66	306	33	922
1978	570	64	320	36	892
1979	552	59	373	40	932
1980	529	55	431	45	965
1981	479	51	438	47	936
1982	413	48	446	52	864
1983	434	52	392	47	830
1984	410	49	415	50	838
1985	424	49	435	50	869
1986	435	47	485	52	929
1987	443	47	494	53	940
1988	393	44	502	56	901
1989	371	45	449	55	822
1990	292	34	551	65	853
1991	300	36	530	63	836
1992	297	41	418	58	717
1993	303	38	500	62	806
1994	294	37	495	62	796
1995	278	34	542	65	828
1996	248	33	507	67	761
1997	248	31	560	69	811
1998	229	30	526	69	757
1999	211	28	537	72	750
2000	191	28	496	72	689
2001	155	21	570	78	729
2002	153	23	503	76	663
2003	143	23	481	77	626
2004	149	21	569	79	722
2005	142	18	636	81	784
2006	111	14	653	85	769
2007	106	15	583	84	696
*Total includes other and/or unknowns

There's lots of fascinating info here. A bicycle segregation advocate recently visited, and pointed out how bicyclists' deaths have decreased since the advent of more and more segregated facilities. I pointed out my belief that the decline in fatalities was primarily the result of the abandonment of bicycling by children. This graph from the Insurance Institute of Highway Safety shows how the ratio between adult and child fatalities has flipped in the last three decades, and how adult fatalities have continued an upward climb (a roughly 100% increase to 624 per year for a three year average), while child fatalities have declined dramatically (averaged roughly to 80%, or 120 per year for the last three years).

Now, what we have also seen over this period is the dramatic decrease in children cycling (schools now consider bike racks to be an option, not a necessity), with an increase in adult cycling. Without reliable counts for bicycle usage, much of this is pure speculation. But what is clear is the decline in cyclist fatality rates has nothing to do with segregated designs, and every thing to do with the cessation of children riding their bicycles.

The segregationist I mentioned posited this question: "If bike lanes and side paths are so dangerous, why has the fatality rate declined?" I would suggest that a 100% increase in adult fatalities is no decline at all... and adult cyclists are the very ones utilizing the segregated designs that I (and others) believe offer no safety advantage at best, lure cyclists into a false sense of security, and are clearly more dangerous than employing vehicular cycling techniques.