The Civil Rights Project at Harvard University published a report on Oct 22, 2002 that better explains the problem of disenfranchised voters.

(Note: It is correct to use either disfranchised and disenfranchised, in case you're curious.)
Democracy Spoiled: National, State, and Local Disparities in Disfranchisement Through Uncounted Ballots

# Spoilage Rates Are Most Prevalent In Counties With High Concentrations Of Minority Voters. Of the 100 counties with the highest spoilage rates, 67 have black populations above 12%. Of the top 100 counties with the lowest spoilage rates, the reverse is true - only 10 had sizeable black populations, while the population of 70 of the counties was over 75% white. There is also a strong correlation between uncounted ballots and black population; specifically, as the black population in a county increases, the uncounted ballot rate correspondingly increases.

# Various Factors Cause the Substantial Disparities in Ballot Spoilage Rates And Mere Technological Improvements Will Not Sufficiently Address These Problems. Evidence from various studies note that while improved voting technology reduces the percentage of discarded ballots across the board, these improvements still do not fully address the disparities between voting precincts, particularly between high-minority and low-minority districts. Indeed, despite popular belief, punch card machines had low ballot spoilage rates in many jurisdictions in 2000, refuting the notion that machine engineering is the critical issue.