According to the TC bylaws, the "TC chair shall appoint a nomination subcommittee to recommend SPS society awards and SPS paper awards within the technical field of interests of IMDSP TC.  All nominations must be approved by the TC."

The bylaws do not specify mechanisms for generating, culling, or selecting from nominations, nor do they specify the meaning of "approved by the TC."  The Awards Subcommittee is committed to presenting the TC with good slates of candidates generated with as much input as possible from the TC and the rest of the IMDSP community.

The goals of our voting procedures are:

• to choose the best candidate (person or paper); and
• to demonstrate the overwhelming support of the TC (assuming this is the case).

The merit of the first point is self evident.  To understand the second point, note that the TC does not decide, it only endorses.  The strength of any nomination that we forward to the SPS Vice President-Awards and Membership matters.

Adjustments selected by the plurality of voters in July 2006 balloting have led to the following procedure.

Round 1

For each category with more than two candidates, each TC member is asked to vote for at most two candidates.  The top two vote getters advance to Round 2.

Round 2

For each category, each TC member is asked to express preference (if they have one) and approval (or lack thereof) in the same round of voting.

With candidates A and B, the following questions are asked:

• Do you prefer A or B?
• Assuming it is preferred by the plurality of voters, do you approve of endorsing A?
• Assuming it is preferred by the plurality of voters, do you approve of endorsing B?

A candidate will be endorsed by the TC whenever it is both most preferred and approved by a majority of the TC.  Of course, we hope that it will be typical for the most preferred candidate to be approved by an overwhelming majority of the TC.

Comments and responses:

• Why bother to ask so many questions?
  The goal here is to allow voters to express preferences and approval in one step, rather than to first find the "most preferred" candidate and then ask for the TC's approval.  [In the old system, the only way to express a preference for A over B was to vote that B is not worthy of endorsement.]   It might seem complicated, but the complexity has really been shifted from voting twice to the counting.
• Why not ask for ranked lists (including "endorse neither candidate"), as in the circulated proposal?
  What is described above is almost equivalent, and it seems to be easier to ask only binary questions.
• With ranked list voting, couldn't you combine the two rounds?
  Yes, the highly-expressive ranked voting system could be seen to obviate many of the merits of our two-round voting system.  But we recognize that often second round voting will be "more informed" because TC members can take the time to consider two candidate papers more carefully than an initial slate of many papers.
• Some possible votes are undemocractic.  They say, `The paper I like is good enough, but I will subvert the majority if they disagree.'
  It is a striking argument; if you agree, do not vote accordingly.  However, it seems that many of you feel strongly about the "gatekeeping" role of being able to vote that a particular paper is not of award quality.