WASHINGTON - One out of every four military personnel and other Americans living abroad may have been thwarted in their efforts to vote in the 2008 election because of communications and bureaucratic problems, according to a congressional report released yesterday.

"Registration deadlines, notary requirements, lack of communication, mail delays, poor address information, and state laws that put in place untenable mailing dates are all severe problems," Sen. Charles E. Schumer, chairman of the Senate Rules and Administration Committee, said at a hearing.

Schumer (D., N.Y.) said the study prepared by the committee and the Congressional Research Service, while providing only a snapshot of voting patterns, "is enough to show that the balloting process for service members is clearly in need of an overhaul."

He plans to work with Sen. Saxby Chambliss (R., Ga.) in crafting legislation dealing with the issue.

The study surveyed election offices in seven states with high numbers of military personnel: California, Florida, North Carolina, Pennsylvania, Texas, Washington, and West Virginia.

It said that of 441,000 absentee ballots requested by eligible voters living abroad - mainly active-duty and reserve troops - more than 98,000 were "lost" ballots that were mailed out but never received by elections officials. Taking into account 13,500 ballots that were rejected for such reasons as a missing signature or failure to notarize, a quarter of those requesting a ballot were disenfranchised.

The study found that an additional 11,000 ballots were returned as undeliverable.

Gail McGinn, the Defense Department's acting undersecretary for personnel and readiness, told the committee that the Pentagon "has taken extraordinary steps to ensure that members of the uniformed services, their family members, and overseas citizens have an opportunity to vote."

She said the department would issue a report to Congress in December on the results of the 2008 election and said other surveys, while providing useful information, should be viewed with caution if their results were based on nonrandom populations.