I'm going to both answer the questions and comment on the questionnaire. I hope it won't get too confusing. 
Comment: I think you are mixing settings and genres here. You can have a detective game in modern, sci-fi or fantasy setting, for example. Better to split these two.
Comment: That's a good question to ask. People often prefer the same genre, but completely different play styles within it. It might be good to supplement this question with: "give a book or movie example of the style you prefer".
Comment: Once again, a good question to ask. It's good to know if you play with casual gamers (or people that suffer a severe lack of free time) that will bail out if they need to spend some time on the game out of session, or with people who need these hand-outs to really get into the game.
Being able to make important choices, that shape the story and change my character, is crucial for me. Take that away and I will leave the table. Consistent game world and consistent character behavior are important, too, but I may accept a little handwaving. Overcoming tough challenges through luck and ingenuity is fun, but I can have fun in gaming without it.
Comment: Not a good question. Very few people are only one of these. GNS/ Big Model is much more useful in describing a game (as in: what's happening at a table) than in describing people.
Also, if you want to use this kind of classification, it's better to use more detailed and precise terms as defined in Forge essays. And it's even better to ask for some explanation in the answer, like I provided above - some people use the GNS terms in different meanings, so that can easily lead to confusion (eg. one player favors immersion and does not want the system to get in the way, another wants a system that precisely models what happens in the game world - which of them is a simulationist?).
Comment: What did you expect to get from this question?
If you want to challenge my paladin, don't make him fight a dragon. Make him decide if he'll give away an ancient sword that he took an oath to protect, to save children the dragon keeps as slaves.
Comment: A good question, again. I like challenges and would be bored without them, but I played with people who perceive this kind of play as opressive and not fun. It's good to know who you play with.
If each conflict threatens characters' lives, the game ends up strongly biased in characters' favor, because otherwise everybody would die in two combats on average. I strongly prefer having a fair chance to lose in every conflict (on average, losing 30-50% of conflicts every session), but with defeat bringing interesting complications instead of removing my character from game.
Comment: Opinions are strongly polarized on this matter; mine is not an exception. One has to know what the group's preferences are, or there will be painful surprises during play.
The GM may and should help if players overlook or completely misinterpret some facts that were presented in play. It's GM's job to keep everyone on the same page on what is happening and what happened.
But solutions should be suggested only for very inexperienced players or when a player asks for it himself. In other cases it is taking choice away from players.
In general, it's better to avoid situations with "good" and "wrong" solutions whenever possible. Some choices may make things painful for characters - but they were taken for some reason. One shouldn't make them less problematic because of that, but shouldn't treat them as incorrect, either.
Comment: One more good playstyle question.
I don't like innovative games for sake of innovation. It's just that newer games are usually more consistently designed than the older ones.
I want the story to be shaped by players' choices, not pre-scripted. But I also don't want a "sandbox", where freedom is so complete that nothing is really important.
I want the situation to be build around characters, around their goals, their beliefs, their fears. It's GM's job to actively question and challenge, to push characters to see how they react.
Comment: People often focus on the "railroad"-"sandbox" spectrum, forgetting that the game may be player-driven and GM-directed at the same time. If PCs and the game world are not seen as completely separate and immutable, there is no conflict.
Note, however, that:
1. GM is not the only person who decides what is important, and what is not. That's a question for the whole group.
2. The more immersive the game, the more color (descriptions, casual conversations etc.) it needs to build the mood.
3. Combat tactics is unimportant for me. I like fights, because they are dramatic, colorful and clearly show that something important is at stake. But if resolving one takes more than 5-10 minutes, it stops being dramatic and starts being boring.
Considering fictional results is tactical and strategic planning ("how to invade the cloud castle"). If these things fit into a single session, keep them in game. If they are long-term, let players plan their actions OOC, as long as they are not kept secret from the GM. This has the added benefit of giving an advance warning on PC actions and letting the GM prepare for them (as opposed to being unprepared or spending a lot of time guessing and preparing for all possibilities).
But players may also consider metagame results of in-game actions - and this requires discussing them OOC. Sometimes, actions that are natural from character's POV threaten group cohesion, block somebody or take too much spotlight. Mouse Guard RPG has an explicit rule that may be paraphrased as "Your characters may get into arguments or even fights. Do it sometimes - it's fun. But always make sure that other players are OK with it." That is exactly how I do it.
A general comment: A few important questions seem to be missing. There are some things that, in my experience, people feel strongly about and that can cause conflicts during play. They are:
1. Existence of social resolution mechanics and how it interacts with PCs.
2. "Soft system" (game rules are a set of guidelines, a tool for the GM to use; ruling > rule) vs. "Hard system" (playing RAW, rules create the structure of play).
3. Fortune-in-the-Middle vs. Fortune-in-the-End and process simulation vs. metagame resources (see: many discussions on "disassociated mechanics").
Answer: Definitely fantasy. Urban fantasy/ modern occult at second place, sci-fi third.
Comment: I think you are mixing settings and genres here. You can have a detective game in modern, sci-fi or fantasy setting, for example. Better to split these two.
Answer: Story, with a solid dose of character growth and heroism. Action and survival are fine too, sometimes.
Comment: That's a good question to ask. People often prefer the same genre, but completely different play styles within it. It might be good to supplement this question with: "give a book or movie example of the style you prefer".
Answer: Definitely. I don't have too much time, but I'm always willing to learn a new game or a new setting. It really helps me improve my gaming skills, and can be fun in itself. I will not only read GM's emails and hand-outs; I may also write my own if it fits the game (in one campaign I currently play I send short stories about my character between sessions; in other I solve mathematical puzzles that are inspired by in-game events).
Comment: Once again, a good question to ask. It's good to know if you play with casual gamers (or people that suffer a severe lack of free time) that will bail out if they need to spend some time on the game out of session, or with people who need these hand-outs to really get into the game.
Answer: All, in order of N>S>G.
Being able to make important choices, that shape the story and change my character, is crucial for me. Take that away and I will leave the table. Consistent game world and consistent character behavior are important, too, but I may accept a little handwaving. Overcoming tough challenges through luck and ingenuity is fun, but I can have fun in gaming without it.
Comment: Not a good question. Very few people are only one of these. GNS/ Big Model is much more useful in describing a game (as in: what's happening at a table) than in describing people.
Also, if you want to use this kind of classification, it's better to use more detailed and precise terms as defined in Forge essays. And it's even better to ask for some explanation in the answer, like I provided above - some people use the GNS terms in different meanings, so that can easily lead to confusion (eg. one player favors immersion and does not want the system to get in the way, another wants a system that precisely models what happens in the game world - which of them is a simulationist?).
Answer: Nothing. There's no thing that can give you a guarantee. There are several factors that make a good game more likely, but listing and describing them is not a short answer.
Comment: What did you expect to get from this question?
Answer: Depends on the game. 3 players + GM for intense, immersive play or for rules-heavy, tactical games. 4-5 players + GM for lighter games.
Answer: One campaign, one GM, typically. But I also play a lot of one-shots, each run by a different person.
Answer: "Quite challenging" to "Extremely challenging". And by "challenging" I don't mean using high numbers in enemies' writedowns, but questioning characters' beliefs and putting things they care for at stake.
If you want to challenge my paladin, don't make him fight a dragon. Make him decide if he'll give away an ancient sword that he took an oath to protect, to save children the dragon keeps as slaves.
Comment: A good question, again. I like challenges and would be bored without them, but I played with people who perceive this kind of play as opressive and not fun. It's good to know who you play with.
Answer: Bad rolls alone should never kill characters. Players should be able to put their characters' lives at stake, but they shouldn't be at stake by default.
If each conflict threatens characters' lives, the game ends up strongly biased in characters' favor, because otherwise everybody would die in two combats on average. I strongly prefer having a fair chance to lose in every conflict (on average, losing 30-50% of conflicts every session), but with defeat bringing interesting complications instead of removing my character from game.
Comment: Opinions are strongly polarized on this matter; mine is not an exception. One has to know what the group's preferences are, or there will be painful surprises during play.
Answer: No GMPCs, ever.
The GM may and should help if players overlook or completely misinterpret some facts that were presented in play. It's GM's job to keep everyone on the same page on what is happening and what happened.
But solutions should be suggested only for very inexperienced players or when a player asks for it himself. In other cases it is taking choice away from players.
In general, it's better to avoid situations with "good" and "wrong" solutions whenever possible. Some choices may make things painful for characters - but they were taken for some reason. One shouldn't make them less problematic because of that, but shouldn't treat them as incorrect, either.
Comment: One more good playstyle question.
Answer: If it has sense, the character handles his own business until the player is back. If it has some sense, but not much, handwave a little and explicitly ask for player cooperation ("Mark is not with us, so don't try to contact his character, even though you can"). If the character cannot move away without breaking suspension of disbelief, the GM plays him until he can.
Answer: I like games that support the play style they advertise instead of doing things because it's how they have always been done.
I don't like innovative games for sake of innovation. It's just that newer games are usually more consistently designed than the older ones.
Answer: Both or neither, depending how you look at it.
I want the story to be shaped by players' choices, not pre-scripted. But I also don't want a "sandbox", where freedom is so complete that nothing is really important.
I want the situation to be build around characters, around their goals, their beliefs, their fears. It's GM's job to actively question and challenge, to push characters to see how they react.
Comment: People often focus on the "railroad"-"sandbox" spectrum, forgetting that the game may be player-driven and GM-directed at the same time. If PCs and the game world are not seen as completely separate and immutable, there is no conflict.
Answer: Don't waste my time on unimportant things. I have only so much time for playing, keep it fun.
Note, however, that:
1. GM is not the only person who decides what is important, and what is not. That's a question for the whole group.
2. The more immersive the game, the more color (descriptions, casual conversations etc.) it needs to build the mood.
3. Combat tactics is unimportant for me. I like fights, because they are dramatic, colorful and clearly show that something important is at stake. But if resolving one takes more than 5-10 minutes, it stops being dramatic and starts being boring.
Answer: There are two types of such discussions: considering fictional results and considering metagame results.
Considering fictional results is tactical and strategic planning ("how to invade the cloud castle"). If these things fit into a single session, keep them in game. If they are long-term, let players plan their actions OOC, as long as they are not kept secret from the GM. This has the added benefit of giving an advance warning on PC actions and letting the GM prepare for them (as opposed to being unprepared or spending a lot of time guessing and preparing for all possibilities).
But players may also consider metagame results of in-game actions - and this requires discussing them OOC. Sometimes, actions that are natural from character's POV threaten group cohesion, block somebody or take too much spotlight. Mouse Guard RPG has an explicit rule that may be paraphrased as "Your characters may get into arguments or even fights. Do it sometimes - it's fun. But always make sure that other players are OK with it." That is exactly how I do it.
A general comment: A few important questions seem to be missing. There are some things that, in my experience, people feel strongly about and that can cause conflicts during play. They are:
1. Existence of social resolution mechanics and how it interacts with PCs.
2. "Soft system" (game rules are a set of guidelines, a tool for the GM to use; ruling > rule) vs. "Hard system" (playing RAW, rules create the structure of play).
3. Fortune-in-the-Middle vs. Fortune-in-the-End and process simulation vs. metagame resources (see: many discussions on "disassociated mechanics").
