Why? If that's what people want, should WotC be aiming to give that to them. And if Monte is not surprised by the answer, then that means one of two things:
1) He can take confidence in knowing that his design is on the right track because he's correctly judged the mood of the audience
or
2) He should know better than to ask such an obvious question in future. Next time, ask about something people do have strong polarised opinions about, like "dice roll vs point buy".
Problem is, there's always a third choice: "I refuse to answer that question because the options are both bloody stupid."
If they use the information from the poll, this will lead them to designing the game in one of two directions, but whichever way they go the result will be hated. Instead, they will have to go for a middle ground. But since that means they will have to not use the information from the poll, since they can't use the information from the poll, they have actually rendered the poll useless. In which case, why bother?
Perhaps an example might illustrate my point better.
Let's assume we're all playing a little game where I ask a question and you have to pick one of two choices. The question is, which of these two women would you rather date and marry? (You can substitute man for woman if you happen to prefer the other gender.) A woman who is smokin' hot but has a vile personality, or a woman who has a perfect personality but is hideous looking?
Now, clearly these two options aren't very realistic. How often in your life will those be your only choices? Probably never! However, the question does lend insight into whether the person being asked values looks or personality more.
Now assume that I introduce a third option. A woman who has a nice personality and who is pretty. Wouldn't just about everyone pick this third option given the other two? I think so! What does this third option really tell us? I suppose it tells us that the person values both looks and personality, but I expect we already knew that. That statement is probably true of 99.999% of people out there. It doesn't tell us anything about the person that we couldn't already assume was true.
Sure, you could introduce even more nuance. Create more choices. The woman who isn't as good looking as the first but better looking than the third, and whose personality is more pleasant than the first but less than the third. The woman who is better looking than the second but less than the third, and whose personality is less pleasant than the second but more than the third. However, at this point you're arguably creating too many options to be able to gather reliable data.
Sure, the response might be more accurate this way if I only ask one person the question, but what if I plan to ask a thousand people? Too many options can confuse the matter and force people to delay their response long enough that they move beyond their "gut" response. Plus, when I'm analyzing the data later, nuanced responses clutter things up. If I only care about the question, "Do people prefer looks or personality?", then I know how to rate responses one and two. I add one to the corresponding column. How do I rate the other responses though? Should I add 0.5 to both columns for the third response; is it reasonable to add 0.25 to one column and 0.75 to the other for responses four and five? What if you picked option five, but what you intended was 0.333 in one column and 0.666 in the other? What level of simplification am I to allow for?
Perhaps then, it's better to begin with a simple question, with only two responses. That way, peoples' gut reactions can draw them to the choice they like better. After all, the only thing I'm interested in is whether people prefer lone wolf or team oriented characters more. I don't really care that folks prefer both aspects in their D&D characters because I've already assumed that to be the case. I fully intend to design characters that strike some balance between those two poles. You can gather that much just from the article.
If you choose the third choice (not to be counted), then you've simply chosen not to be counted. I seriously doubt that Monte Cook expects L&L to reach every D&D player out there. So you simply become part of the unknown demographic. Again, the existence of such a demographic is reasonably assumed.
In any case, I don't expect this is serious market research. This is a designer trying to get the roughest of measurements on the way the wind is blowing in the D&D community. Listen to the tone of article. It's more like he's spit balling ideas, like he might do at a brainstorming meeting with fellow designers about possible directions D&D might go. The kind of thing they might do before they design a single rule. The only difference is that we happen to be invited to take part in these meetings.
