The McDonald estimates (McD /VEP) are based on the votes cast for president (excluding Overseas votes), divided by the number of voting age citizens minus ineligible felons. 

The CQ /VAP statistic is the votes cast for president divided by the voting age population.

The CPS (Current Population Survey) data based on interviews of a sample of approximately 55,000 households, response rates exceed 90%..

The NES data are based on post election surveys of approximately 1,500 persons, with a 50-65% response rates.

McDonald argues that voter turnout has not declined since 1972 when one looks at the turnout as a percent of the voting eligible population.  In recent years the percentage of the voting age population that is not eligible to vote, due to noncitizenship and felony status has increased sharply.

Sources:

NES /VAP:  American National Election Survey
McD /VEP:  Michael McDonald US Elections Project
CQ /VAP:    Congressional Quarterly Inc. "America Votes"; see US Statistical Abstract 2002, table 395
CPS /VAP /VAC:  Current Population Survey

previous weeks:  Iraq Causalities  Budget deficits.