Poll Results | Survey Results | About the Voting | About the Results

The second Locus Online Poll and Survey drew 238 responses over a period of a month. This was 16 more than in the 1st poll and survey, but there were apparently fewer distinct voters this time, i.e., more repeat voters. There was nothing like the apparent bloc-voting in the first poll, no pattern of votes all targeting a particular work, but there were more instances of duplicate, or near-duplicate, submissions received, usually within minutes of each other, occasionally separated by days or weeks. The ease with which such votes can be submitted is why I never take ''instant polls'' that you see on various websites (click Yes, No, or View Results) seriously for a moment. Perhaps some of our voters didn't realize that our ballots were not tabulated in such an automated fashion -- they couldn't be, with free-form fields for names and titles that are never formatted the same by one voter to the next. Or there may be innocent explanations; I've been told of network glitches that cause two copies of an email to be sent out. Maybe.

Of the 238 ballots received, 35 were judged to be duplicates, and were discarded from the tabulations, leaving 203. In the 1st poll and survey, there were 222 ballots, of which 13 appeared to be duplicates, leaving 209.

The problem of duplicate online votes probably sounds more serious than it is; with the exception of a couple 1998 categories, the final results would not have been appreciably different even if all the duplicate votes had been counted. On the contrary, I think the results of the two polls are quite reasonable, and I'm particularly pleased with the results of the all-time short fiction poll.

I compiled the votes semi-automatically: the ballots were received by email (in the format voters saw as confirmation after they clicked the 'submit' button) and then exported, from Microsoft Outlook, to an MS Access database file. I wrote a program to parse the body of each email into separate fields, and then, for the fill-in categories, manually filled out a parallel field for each with consistent author/title abbreviations. (The order of ballots received was retained in the database table, so that most repeat votes were obvious.) Tabulations were then done automatically using Access queries and sorts.

Votes in the all-time categories, where five votes were allowed per category, were counted in the same manner as the Locus magazine polls, using the Carr system, devised by the late editor and writer Terry Carr. A 1st place vote gets 8 points, a 2nd place vote 7 points, and so on. This gives equal weight to making the ballot and the placement on it; a 1st place vote is worth twice as much as a 5th place vote, but not five times as much.

Of the 203 tabulated ballots, fewer than 100 voted in any of the 1998 categories; somewhat over 100 voted for 1990s short fiction writer; about 125 voted for all-time short fiction writer; but only about 100 voted in the all-time short fiction categories: novella, novelette, short story. (Many voters checked the 'have you read?' boxes yes or no but didn't vote in any of the fill-in categories.)

Several of the 1998 novella votes were for novels (e.g. Children of God, Komarr) and were discounted from the above figures. (The first ballot received after posting the Nebula Awards winners on May 1st voted for two of the Nebula winners.) Voting in the 1998 categories was scattered, since only one vote per category was allowed, and no lists of suggestions or recommendations were posted. Only two short stories, Sterling's and Carter's, had more than 2 votes; stories by Howard Waldrop, Karen Joy Fowler, Bruce Holland Rogers, and Robert Reed had 2 votes each; everything else had one. A number of the writers nominated for 1990s short fiction writer have not in fact produced much if any short fiction in the 1990s (e.g. Jack Vance) but all votes in that category were tabulated anyway.

The numbers of items voted is higher in the all-time categories than in the 1998 categories, since voters could nominate five items in each all-time category. Votes for all-time stories were tabulated together (regardless of which category a voter had put a particular story into) and then sorted by category for the final rankings..

--Mark R. Kelly

Technical note: these 2nd poll and survey results pages are using a ''style sheet'' for font formatting (it makes life easier especially for material arranged in rows and columns [i.e. using html tables]). If you see the links at the top of the page underlined, it's because you're using a browser that doesn't support style sheets; if you do have a late-model browser, you should be seeing the links without underlining, among other differences. The pages should look OK either way, but if relying on style sheets upsets anyone, let me know.)