Below is the editorial from the June 16, 1997 edition of InfoWorld. Because what's in InfoWorld Electric isn't always the same as what's in the print edition, I've retyped it here. It's, of course, prone to my typing mistakes.
Readers determined winners of this year's Readers' Choice awards

When we invalidated the results of our March Web-based Readers' Choice poll, some readers were outraged (See "Help us reclaim Readers' Choice Product of the Year awards from OS/2 zealots," March 31, page 55, and "New poll will determine winners of Readers' Choice Product of the Year awards", April 21, page 67.) Many of those same people will have the same reaction to the results of the voting that we are reporting this week. (See article on page 92.)

In both cases, the flash point is OS/2. In the March poll, which was conducted only on InfoWorld Electric, OS/2 "won" by an overwhelming majority in the categories of Client, Server, and Overall Product of the Year, receiving as many as six times the number of votes of the second place finishers.

That unusual voting patter -- plus the fact that we failed to ensure that ballots were cast only by InfoWorld readers -- led us to nullify the results and commission an outside research company to conduct a statistically valid survey among our readers. The research company used a ballot identical to the one that had been posted on the World Wide Web as the basis for a telephone survey.

In the telephone survey, OS/2 received only 2 percent of the votes in the Client category, 5.7 percent in the server, and combined 2.3 percent in the Overall Product of the Year category. I'm sure the OS/2 community will see a conspiracy in those results, particularly because two Microsoft products are among the winners. We are confident that the telephone survey reflects the opinions of InfoWorld readers, both print and Electric. So this year's Readers' Choice awards will be awarded to Microsoft Windows NT Workstation 4.0 (Client), Microsoft Windows NT Server (Server), Norton Administrator Suite Premiere Edition 3.0 (Networking), and Netscape Navigator 3.0 (Internet and Overall Product of the Year).

If you'd like to express your opinion on our Readers' Choice Awards, join our forum on InfoWorld Electric.

Sandy Reed is the editor in chief of InfoWorld. Contact her at sandy_reed@infoworld.com.



Personal Comments:

I would like to take this opportunity to point out a few things in Ms. Reed's editorial. Before I start, I want it known that I do not consider the first poll to necessarily be valid. I believe there is good reason to suspect the first poll.

She makes no mention that the telephone poll excluded anyone that worked for a company that produced hardware, worked for a company that produced software, or considered themselves an IT professional.

One of Ms. Reed's reasons for invalidating the previous poll (she stated this in the earlier forum, of which I don't have an archive so you'll have to trust my memory) was that OS/2 shouldn't win because it doesn't have the market share. She fails to mention here that Windows NT has a smaller market share than OS/2.

Ms. Reed spends far too much time in this editorial telling us how poorly OS/2 did in the repoll.

Ms. Reed changed her story concerning OS/2's votes in the previous poll. In her March 31 , 1997 editorial she states that "In all three categories, OS/2 received at least six times the number of votes of its nearest competitors". Now she tells us that "receiving as many as six times the number of votes of the second place finishers" which tells us that it only had to happen in one of the categories. Ms. Reed has yet to release the actual numbers from the previous poll.

Ms. Reed tells us that "We are confident that the telephone survey reflects the opinions of InfoWorld readers, both print and Electric. " Yet she limited the new poll to only print subscribers and even then she limited the audience as I described above.

In her March 31, 1997 editorial, Ms. Read specifically asked the print edition readers to not let OS/2 win the second poll. If that doesn't indicate a bias, I don't know what does.