Trying to compute the effect of stimulus is guesswork. The approach Holtz-Eakin used was simplistic and a complete shot in the dark. He compared the actual GDP with an estimate that assumed the GDP linearly continued its decline (see the charts here -
http://www.americanprogress.org/issues/2011/08/holtz_eakin.html). There is no reason to assume the GDP change would continue linearly.
This type of "analysis" is simplistic and has no basis in reality. The economy is dynamic. And to really analyze the impact the cost of the spending has to be taken into account. If the spending is from taxes then the lost opportunity cost and the cost of removing the taxes from the economy has to accounted for. If its borrowed money (or printed) then inflation, interest, and debt payments have to be considered.
Applying this approach to the revised data (as in the reference
http://www.americanprogress.org/issues/2011/08/holtz_eakin.html) shows a multiplier of 4. That is absurd, a total multiplier of 4 is extremely high. Even the CBO reports of the stimulus do not show multipliers that high. The largest multiplier is due to money spent on education because a little education pays a lifetime dividend. The spending has to be to actually educate people (as in student loans/grants), not on supporting the education system/bureaucracy. Little of the stimulus went to educating people.
Research in peer reviewed journals before this subject became so politicized shows the multiplier ranges from 0.1 to 2, and several studies indicate that the total multiplier is above 1 only during wartime, which is believable because in wartime a large part of the extra spending truly goes to infrastructure - roads, airports, ports, trains, planes, ships to transport the war material; factories to build the vehicles, guns, aircraft, etc to fight the war. Much of this infrastructure has dual military/civilian uses and serves civilian purposes when the war ends. Employment increases during war as well.
And if people are going to follow Keynes, then follow it all the way. The Keynes approach is to deficit spend (that's deficit spend, not increase taxes) during bad times in order to boost the economy, and pay off the accumulated debt during good times. It is not to deficit spend like a fool all the time.