Editor's note: The Free Press is releasing these articles by Ron Baiman that generally support the analysis by Robert F. Kennedy Jr. in the current issue of Rolling Stone. Also see Richard Hayes Phillips' Through a Glass Darkely.

Connally Spreadsheet (Excel Spreadsheet)

Was the 2004 Presidential Election Stolen?: The History, The Crime, The Cover-Up, and Conclusions (Adobe PDF)

Analysis of Connally Spreadsheet

1) Of only 14 out of 88 counties where Bush did abnormally well relative to Moyer (better than a 1.43 Bush/Moyer ratio - Column E) 9 of them (M) just happen to be the same counties where Kerry did implausibly poorly relative to Connally (R).

In the other 3 K/C<1 counties (Highland, Miami, Van Wert) Bush didn't do too shabily either, his lowest ratio is in Miami which at 1.36 is still well above his overall state average (relative to Moyer) of 1.21 (Cell D92). I set the cutoff at 1.43 to select out most of the counties where Bush did well relative to Moyer but that were not K
2) Now, some might claim that these abnormally high C/K ratio and B/M ratios are a result of voters switching from Moyer to Connally (reducing the denominator of B/M and increasing the numerator of C/K). To check this I looked at the counties where M/(M+C) was below the state average Moyer vote share of 53.2% (indicating a less than average vote for Moyer relative to Connally). Of the 12 K/C<1 counties this only occurs in one (Butler) (Column K).

3) The flip side of this is that this is also the only county where C/(M+C) was above the state average for Connally of 46.8% and K/C<1 (U).

4) 11 of the 12 K/C<1 counties are among the only 20 (out of 88) counties with K < 34.1% (W94 and X94).

5) These same 11 K/C<1 counties are among only 19 counties with B> 65.6% (AB94 and AC94).

6) The "above average" Bush vote in B/M>1.43 and K/C<1 counties is 68,928 (F) which is very close (considering that these are all estimates from state averages) to the "Lost" Kerry vote (H) of 81,599.

7) If "above average" Bush vote from the other 3 K/C<1 counties is added to this it comes to 75,766, even closer to the 81,599 "lost" Kerry vote (G and F96).

8) A county by county comparison of the "lost" Kerry vote (H) to the "extraordinarily above average" Bush vote in these same counties (F) and the other 3 K/C<1 Counties (Column G) shows a very close county by county order of magnitude correlation.

Conclusion: about 75,000 to 82,000 votes were shifted from Kerry to Bush. I don't know how you could possibly explain this remarkable series of "coincidences" any other way!