A story bears little relation to reality, and that any statement about the way things are 'in the real world' is a very poor way of figuring out how things ought to be in stories. The simplest argument in support of this belief is the phenomenon of the 'satisfying ending'. If a story were just telling a series of happenings, there would be no reason the last one would be special. Yet, audiences think endings are definitely different from middles or beginnings, they expect endings to be 'satisfying', and random people tend to agree with each other about whether something is an ending and even whether it is a satisfying ending.
My own belief that stories are inherently moral comes from six years of studying what people mean when they say stories are satisfying and important. I studied this by looking at educated people's theories of how stories are structured, uneducated people's choice of what myths and fairytales are worth listening to many times, and psychological studies of language and audience emotions and behavior. My concludion is that conflict in a story comes from an argument over goals and/or methods between characters/faction, and the climax of the story comes when that argument is resolved. So the plot structure of a novel is equivalent to the persuasive argument made in a speech or essay, it just gets a lot longer (but also more subtly persuasive) when you encode the argument into the actions of characters and the behavior of the setting. The conclusion of the argument is the moral of the story. That's what I mean when I say that all stories are inherently moral, because every story is an argument and in deciding who wins the author necessarily promotes the belief that that side is right.
I also believe that gameplay is inherently strategic because it's based on math, and thus any purely strategic decision you want the player to make is thus better handled as gameplay than storytelling; on the other hand it's very hard to convey abstract beliefs through gameplay because it's wordless, so those are better handled with storytelling. Although if one wanted to make a strategic choice have moral consequences you might want to use storytelling to warn the player to expect moral consequences, and of course to present the consequences when they happen.
Taking the Keirsey personality test is an easy, interactive introduction to this personality type system, orienting people to the system by showing them where they stand within it and hopefully demonstrating that it has some accurate things to say about them, and thus will have accurate things to say about characters' personalities. Also, perhaps you have heard the saying that all characters are made out of little pieces of their author? People who have not studied personality understand their own mind best. Also writing is a solitary hobby, most writers are introverts, and introverts usually do not understand other people that well. So, many writers find it easier and more satisfying to write from their own point of view than some hypothetical other person's. A writer can only know for sure how he or she thinks; so a character who thinks just like we do is the only one we can write with surety. Many student authors fall into the trap of assuming that all characters think like they do, which results in all their characters being the same, or assuming that anyone who doesn't think like they do is insane, resulting in inconsistent characters who do things without motivation.
The whole Mary Sue phenomenon is evidence for this. One of the major motivations for writing fiction is that you want to experience safe adventures and escape the unpleasant real world, so you create a character based on yourself, usually with some wish-fulfilling improvements like stunning good looks or magic powers or a loyal telepathic pet, and put this version of yourself into a more interesting world populated by fascinating love interests and opportunities for you to be a hero and be admired.
A personality type system is a, well, systematic way of knowing other people including characters. And that's generally the point of education, to make people look at things in a more effective systematic way rather than the haphazard instinctive way that the average person, and especially artistic types like writers, thinks about personality. With the personality type system it is easy for writers to see how a character might think differently from them in their own logical way. I like the Keirsy system best because it is the most orderly - to create a character you need only make 4 binary choices, and theoretically every character ought to be one of these 16 possible types.
The point of the writing exercise as a whole was supposed to be getting people to study how a character's personality determines their style of interacting with another character. Technogoth thought people needed to work on dialogue, and I thought the most important thing to learn about dialogue is that good dialogue is good primarily because it shows the clearly-defined, consistent, psychology of the characters' thinking and speaking it. Some people naturally develop an understanding of the different ways different people think, and that others have studied character personality as part of learning how to write well. And I have seen some writers succeed very well at writing a character whose personality is completely different from their own, so it's definitely a skill which can be learned.
A sergeant could theoretically be any of the 16 personality types; some would be more likely than others because some are more inclined to like an orderly military environment and some are better at dealing with subordinates, and some are more likely to accept the responsibilities of leadership. But basically, the man's job is irrelevant. The important thing is what role the author wants him to play in the story. Is he supposed to be comically incompetent, charismatic but short-sightedly leading people into trouble, a loner who wants to avoid his leadership responsibilities, or what? If you know what role you want him to play, it should be easy to figure out his personality type, which may give you more insight into how he thinks or, if you already understand that, just provide a rule of thumb to keep his character consistent throughout the story. If you don't know what role you want him to play, or know something but not everything about how you want him to be, browsing through the 16 types you can find one which resonated and use it as a guideline to finish figuring the character out, to help you figure out how the character would think and act in a particular circumstance, and again to help you keep the character consistent throughout the story.
Structural analysis. Basically the idea is to remove all unnecessary details from the story, seeing it as a sequence of symbolic archetypes and transformations. Then you compare several stories to see if the same archetypes or transformations appear in multiple ones.
Like, here's my analysis of beauty and the beast (I left out the curse/rose/time limit part because I don't find that part interesting.):
1. A conventionally beautiful girl whose major trait is kindness
1a. In Disney version the girl does not fit into her society and resists marrying a member of that society.
2. In payment for a crime, the girl is sacrificed by being given to a beast (see also dragons eating virgins, virgins trapping unicorns)
3. The beast's castle represents an alternative society which rewards different traits and abilities.
4. Here Beauty's kindness/openmindedness allow her to prosper where others would fail. (Equates to the testing part of standard hero monomyth.)
5. Having tested the girl, the beast accepts her into his alternative society by falling in love with her.
6. Beauty must confront her old society so she can definitively choose the new one over it.
7. Wedding, barren castle transformed to be domestic and fruitful.
Link to theme-based character design
Link to story analysis and emulation using beauty and the beast as an example
Link to disjunctions and emotional metabolism