The graphic purported to show the changes in alcohol-related fatalities from 2004 to 2008 in three states: New Mexico, which had a adopted a mandatory ignition-interlock requirement in 2005 for alll drunk drivers; Virginia, which had adopted such a requirement only for those drunk drivers found to have blood-alcohol readings of roughly twice the legal level; and Maryland, which had not adopted interlock legislation during that  period. It gave the impression that fatalities in Marylland had fallen at a rate equal to or greater  than the rate in New Mexico -- bolstering Vallario's position that an ignition interlock bill affecting all drunk drivers was not needed in Maryland.

The chart did not use the most relevant and authoritative data: the National Highway Traffic Safety Administration's tabulation of fatalities involving drunk drivers (using the national standard of .08 percent blood alcohol). Instead it pulled the Maryland numbers from a group called AlcoholAlert and focused  on all fatalities  in which the driver  had any alcohol in the bloodstream --  down to the non-intoxicated level of .01.

But the numbers from New Mexico and Virginia, which were charted on the same grid, were taken from some other source. They did not match up with the comparable Maryland numbers from AlcoholAlert. (A staff member for the Judiciary Committee agreed that the  numbers were incorrect and said he would check into the source of the data.)

The effect of using the non-comparable data was to misrepresent the results in New Mexico, which proponents of the interlock bill had been holding up as an example of the positive results of such a law. Had comparable numbers been used, New Mexico's numbers would have shown a decline of 44  percent rather than the 34 percent it did show. (The Virginia  numbers, no matter which data are used, show no positive effect from its law with a .15 threshold -- the number Vallario and the liquor lobby were advocating.)

A table inside Vallario's packet, based on actual NHTSA data, actually makes a strong case for the bill  the chairman killed. It showed a decline of 26.3 percent in the drunk driving fatality rate in Maryland compared with a 39.4 percent decline in New Mexico. (Maryland's decline was exaggerated because of a spike in deaths in 2004. If you take 2005 as the starting point, Maryland's decline was 0.3 percent while New Mexico's was more than 34 percent.)

However you look at these numbers, the data are too sketchy to "prove" anything. At best, they merely give an indication that Virginia's law isn't working and that New Mexico's is. But when they are portrayed accurately, the numbers do buttress the case made by MADD and other advocates of a stiff ignition interlock law.

Vallario provided this chart  to members of his committee late at night when members were exhausted and nobody was in a position to challenge his data. Even if they had, the outcome likely would have been the same. But members of a legislative committee, as well as the public, depend on the information provided by the staff and the chairman of a committee to be  accurate. When it's not, an explanation is in order.

In the House of Delegates, committee chairs answer to  only one person: House Speaker Michael  E. Busch.

Caroline Cash, executive director of MADD Maryland, said the distribution of the inaccurate chart shows that "the public is not getting the facts."

Asked what recourse the group would take, she said: "The first step would be the speaker's office."

It would be unlike the professional staff of the General Assembly to deliberately  take part in any effort to deceive members. My first question would be whether the source of the chart Vallario distributed were a lobbyist. That would explain a lot.