25 Comments
In a statement to the GS student body, GSSC President Brody Berg expressed skepticism about the results of the NROTC survey, writing that “I believe that this poll result is meaningless due to the huge number of apparently fraudulent votes.” Although almost 2000 fraudulent votes were removed (including one student who voted 276 times) by matching one ID with one vote, Berg expressed skepticism that this strategy was adequate, further noting that “the votes thrown out were done with no supervision.”
“Even though the side I support ‘won,'” he concluded, “due to the huge number of fraudulent votes, myself and the Executive Board of the General Studies Student Council believe both General Studies and the Columbia community at large would be better served on this controversial issue with a poll whose results we can trust.” Earlier in the day, CCSC president George Krebs had expressed confidence in the result, writing in his CCSC email that “Multiple votes were not counted. If a student attempted to vote multiple times, only the last vote they cast was counted.”
Bwog notes that many of the multiple results probably came from students sharing links with each other, as occurred with the Democrats (who sent a listserve email with one person’s link) and Bwog itself. If the last vote was counted, then the last person to use that link would be the one whose vote counted. In addition, questions about the ability to use others’ links and “hack the survey” remain unanswered.
Berg’s full e-mail after the jump. Bwog thinks this issue may not be over.
Dear Ladies and Gentlemen of General Studies,
The results of the NROTC Survey given to General Studies, Columbia College and the School of Engineering are in:
- 6913 email invitations
- 2971 valid votes (43% of the population)
- 1463 Yes (“Bring NROTC to campus)
- 1502 No (“Don’t bring NROTC to campus)
- 6 abstain
But here is where it gets interesting:
- 4905 votes were recorded
- One person voted 276 times
- And “after eliminating the duplicates and matching the unique IDs with our records, 2971 unduplicated votes are determined as “valid.”
(All this information is from Columbia College SDA Advisor David Cheng)
What this means is that there were nearly 2000 votes thrown out of a race decided by only 39 total votes. More votes were thrown out than the individual size of three out of the four colleges in the survey. Furthermore, the votes thrown out were done with no supervision and with a fraud detection strategy only as good as “eliminating duplicates and matching unique IDs.”
Now, let me be clear, even though I personally voted No, I believe that this poll result is meaningless due to the huge number of apparently fraudulent votes. I say this out of common sense and because, as a software industry veteran of 10 years and senior in the Computer Science department. My professional gut feeling tells me that a significant number of fraudulent votes may not have been removed.
I was quoted in the Columbia Spectator on September 26th saying “My number one thing from the very beginning was to assess student opinion on NROTC in a fair manner.” In the General Studies student body we are proud to have a lot of students on both sides of this issue. Even though the side I support “won”, due to the huge number of fraudulent votes, myself and the Executive Board of the General Studies Student Council believe both General Studies and the Columbia community at large would be better served on this controversial issue with a poll whose results we can trust.
Thanks, and stay tuned,
Brody Berg
Student Body President
General Studies Student Council 2008-09
Columbia University
Loading …
25 Comments
@why does this school talk mad shit about anyone who takes a stand on important issues?
the vote WAS meaningless. someone had to say it. and obviously, the councils will try to cover their asses.
@Breakdown? How did the votes breakdown by college?
Since this started in SEAS and NROTC would benefit them the most, it seems obvious that we should at least know how SEAS students voted.
Also, with only 39 votes separating yes from now in 3 colleges, it’s possible 2 of the 3 colleges actually voted for NROTC. If it’s a 2-2 tie, that changes a lot. Heck, if we know only 1 college – especially if that 1 college is SEAS – voted for NROTC, that still changes things.
@that, or It’s even possible that all three schools voted for it, despite the popular vote losing overall. It’s called Simpson’s paradox.
@it didn't http://commentariat.specblogs.com/index.php/2008/12/02/councils-failed-divide-votes-school/
@brody SUX You’re a douche is harsh but brings up a good point. Brody doesn’t care about GSSC, he only cares about himself. Everybody agress that GSSC is silly with their statements rife with their inferiority complex. I think it’s time for us to have a GS student who understands the undergraduate dynamics to be president. Alfred Davis would do a much better job as President.
@You're a douche I think it’s cute that Brody Berg calls himself “a software industry veteran of 10 years.” If he’s such a hotshot, what’s he doing in a dumping ground of people who couldn’t cut it in the job market?
@... i think it’s cute that you believe that everyone is here because they have to be.
@well brody is the most obnoxious person at columbia
@correction as for my second to last paragraph: I should have said if only 20 more votes were corrected from No to Yes vs the other way around, then the majority would have swung the other way.
@well then to #2: What he means by “decided by 39 votes” is that the difference between Yes and No votes is 39 votes.
Also, to everyone here (and especially #10): while the legitimacy of the GSSC president’s legitimacy and “opinions” may be questionable, his stated facts about the poll are NOT. I feel that the CCSC President’s support of the poll is completely irresponsible. Why?
“If a student attempted to vote multiple times, only the last vote they cast was counted.” …really? The evidence shows that in many cases the SAME LINK was sent to MULTIPLE people. This means that there is a strong potential for many votes NOT TO BE COUNTED. Now, 2000 votes were NOT counted. The vote differential (#NO – #YES) was only 39. Even if a fraction of those 2000 thrown-out votes were real opinions accidentally cast on someone else’s link, there is ample room for the majority to swing the other way if the votes were counted as they should have been. That is, with a secure system that allows one vote per person.
Also, given the opportunity for hacking the system and intentionally voting for someone else, a large number of the “counted” votes may not validly represent the intended vote of the individual. Keep in mind that this is DOUBLY important: correcting each “incorrect” vote takes away one vote from one side and simultaneously adds a vote to the other. This changes the vote DIFFERENTIAL by 2! That means if only 20 (20!) votes were corrected, the majority would swing the other way.
Regardless of whether Berg’s status and participation in this poll’s inception is actually questionable (of which I have no information), all of you should seriously consider the raw statistics and math that he is providing here regardless of his identity.
@huh Columbia Undergrad Student Body Concludes GS Opinion “Meaningless”
@Brody has no right to talk about this issue. He did not get involved in any planning meetings and quite blatantly refused to represent his constituency.
Niko was bad enough but at least he tried. My personal view is that if Brody doesn’t want to represent GS he should just step down or the GSSC needs to start thinking of more serious steps to make sure they aren’t a running joke on campus
@... see: military == annoying.
it follows that: more military == more annoying.
therefore: nrotc == bad.
@irony did this guy forget that GS is meaningless as well?
@daaamn that was a pretty good burn
@how is it possible for people to abstain? did not see that option on my survey
@it's not It’s for people who voted without selecting either one. The correct term is ‘invalid response’.
@Ron Gejman In response to #3: The problem is not deleting the clearly fraudulent or duplicate votes. The issue is that someone apparently used a “scattershot” approach – submitting the form with random identifiers in the hopes of landing upon a valid identifier. That would account for the large number of “unmatched” votes. Unfortunately, it would be impossible to detect how many times he landed on a correct identifier – unless they all came from the same IP address.
@So we don't know How many times this person actually voted? Hell, if he voted that many times, he might have gotten it right 39 times.
@bullshit Guessing identifiers would have been pretty much impossible. Let’s do some simple stats here: note that each identifier is 24 characters long. Each character can be a digit or letter, which gives 36 different possibilities. This allows for 36^24 ~= 10^37 different identifiers. Now, there were ~6000 different identifiers, which means that the chance of landing on a correct one through each guess is about 10^-34. That is not a big number at all.
@Nope Those aren’t the identifiers necessary. See the Bwog comments from a while back. All you needed was the survey ID number, not the identifier at the end. Poor coding on someone’s part.
@276 times? Is that 276 people voted using one link (maybe the Dems or Bwog’s?), or 1 person hitting the buttons 276 times?
@its really not that hard to just plop everything into excel and delete all the rows that have multiple IDs. It’s not like this is rocket science here.
@interesting 1934 votes were thrown out.
and the survey with the most number of duplicate votes received 276 votes.
this leaves 1658 votes. how many individual surveys constituted these 1658 votes? certainly more than 39.
@pollz r ppl 2 “a poll whose results we can trust.”