This is for lookingforoctober [DW].
DISCLAIMER: I'm discussing how it works for me; writing process is so individual that it may (almost certainly) work differently for y'all. If your reaction to my process is OH HELL NO THAT MAKES NO SENSE, you're not wrong! Only you know what process is right for you.
BACKGROUND: My published career started in 1999 with short stories and I continued on my merry way for over 15 years after that. I did actually attempt to write novels before that, of course. I wrote my first novel in middle school. It was completely awful and, so far as I know, no copy of it exists anywhere in the world thanks to technological obsolescence (floppy disks?) and/or flood. I wrote two more novels in high school, of which one was a novel by wordcount only (it was a fix-up consisting of a bunch of connected short stories), and both of which were terrible and also don't exist anymore. And then I spent the next odd decade attempting to write Paper Souls (as its title ended up) until I finalized a draft, and even then I ended up pulling it after a couple months of queries when I realized that it was so fundamentally flawed that it would require a ground-up rewrite. (Remind me to talk more about that sometime.)
I like to think of short stories vs. novels (or more accurately, some short stories vs. some novels, as written by me) in this way:
A short story is a battle. Sometimes not even a battle; a moment in a battle, a short sharp shock. A novel is a war.
Orson Scott Card  has a writing framework which I find helpful because it is reductive. It's called MICE (Milieu, Idea, Character, and Event), and he suggests that most stories focus on one or two of these elements. Milieu is all about the setting/environment, e.g. a travelogue. Idea is what it says on the tin, and a lot of hard sf that focuses on some shiny piece of scitech would fall under this category. Character is also what it says on the tin--category romance is often highly character-focused. And Event is where all the plot-heavy stuff falls, like a lot of epic fantasy.
 I have severe disagreements with what I know of OSC's personal views, although I was not aware of them at the time I read some of his writing how-to books (or, indeed, novels). But, as they say, fas est ab hostes doceri. I consider taking what I learned from him in the domain of craft and swanning off to write queer sf/f to be an entirely satisfactory form of revenge; YMMV.
In any case, the point of MICE is not that (most) stories are a pure expression of any single type, but that you can use this framework as a lens through which to examine the amount of emphasis given to any one or more of them. I'm going to largely disregard Milieu in what follows because it is my impression that the "pure" travelogue is a less common form in sf/f. (I could be wrong? Counterexamples welcome.)
For me, personally, I focused heavily on the Idea element in my short fiction. Part of this was based on genre. Science fiction and to some extent fantasy are both often Idea-based--what-if's and magic systems and the consequences of various kinds of FTL drives (too many examples to list, although I'm thinking of Timothy Zahn's "Cascade Point" right now). This isn't, of course, exclusive to short fiction! Brandon Sanderson's magic systems are a great example of an Idea element in epic fantasy.
But part of this was based on length. You can communicate an Idea and some of its interesting consequences at short story length. Obviously, again, you can also do it at novel length. I point you again at Brandon Sanderson.
On the other hand, at short story length, there are limitations on how much of a Character arc you can delineate. You can certainly get an arc, if you want; there are short stories where the protagonist doesn't change if the focus is elsewhere. In "The Battle of Candle Arc" [Clarkesworld], which is an Event/Idea story, the protagonist doesn't have a character arc at all. He's the same person at the beginning that he is at the end. Or you can make a change for the character(s) a major focus of the story, as in Amal El-Mohtar's "Seasons of Glass and Iron" [Uncanny Magazine]. But it's difficult simply for wordcount reasons to get into really nuanced characterization and character development, and as a gross generalization I would also say that it's difficult to have a character go through more than one or maybe two really large changes. At novel length, you can have a character arc go up and down and up and down and twist into Möbius strips (or whatever nonorientable surface of your choice, I'm going to just leave this metaphor there).
I would say that there are similar constraints on Event or plot, although here the reason is not so much wordcount as the human sense of scale. What's the saying, one is a tragedy but a million is a statistic? You might be able to give a summary of a ten-million-year war in 5,000 words but it's doubtful that, unless you are some kind of genius (and maybe you are, but I'm not), each sentence is going to hit as hard as if you'd picked out some specific event and detailed it. The one genius example I can think of is actually at novel length, which is Olaf Stapledon's Last and First Men; it doesn't have characters in a conventional sense but is rather a future history in which every civilization in a parade of civilizations gets one, maybe two paragraphs. It's brilliantly done; but it's not typical commercial sf.
I started writing during a time when an accepted way to break into writing (or try) was to work your way up through the magazines and then eventually write a novel and try for novel publication. Publishing has changed a lot since then and I actually don't recommend this approach anymore (unless you want to, in which case have at it!). And one of the reasons I don't recommend it, beyond the practical, is that in my case, there were a lot of things about writing a novel that I didn't learn writing short stories.
One of them was characterization. At short story length, characters were temporary resources. I got into their heads as long as I needed to in order to make the plot fire, and then I forgot about them. When I was writing Ninefox Gambit, one thing that I found challenging was that I had to spend so much of the novel inside the heads of Cheris and Jedao  and get to know them in order to write them at all. And writing fanfic hadn't prepared me for this not because there's anything wrong with fanfic, but because for me trying to portray someone else's character, a known quantity, is different from inventing the quantity to be portrayed.
 Cheris is the major POV in that novel, but since Jedao is her antagonist and moreover never shuts up, I had to understand him pretty well--possibly even better, because Cheris is basically honest and Jedao isn't.
The other is structure and pacing. My theory of pacing--of tension and release of tension--mostly comes from musical composition. Which is hilarious because I don't think I've ever composed a single piece longer than three minutes. (Well. I guess the orchestral suite was maybe 10-12 minutes total, I forget.) But knowing how to pace a short story is different from knowing how to manage something as large as a novel. I was helped here by Daedala, who besides being a beta for Ninefox also introduced me to Larry Brooks' Story Engineering. Brooks' book may not work for everyone but it was what I needed. (Obviously, remaining problems with pacing are my responsibility.)
In a short story, you get maybe one arc--rising tension, climax, a dénouement of some sort. Maybe a couple mini-arcs in a longer story. At novel length, this has the possibility to go fractal. Obviously you can structure your novel like a thriller or Ravel's Boléro, one big crescendo of things piledriving forward forward forward as the tension ratchets ever tighter. But that isn't the only way to do it. (I personally find that sort of structure exhausting and a little too predictable, which is perhaps why I don't read many thriller-type books. There's nothing wrong with it, just a matter of de gustibus.) You can have arcs that are made of smaller arcs. And this is something that I'm still struggling with.
Raven Stratagem (forgive me for talking about my own novels but I only know the experience of writing through my own work) was interesting and tricky because I had three POVs with three different storylines and I had to align them perfectly so that the midpoint and climax events fired in tandem--sort of an exercise in synchronization. I had to futz around with adding chapters and rearranging chapters and screwing around with my outline in index-card form while revising in order to try to get it to work. I still don't know if it did, actually, work. Nothing I had done writing short stories had prepared me for this, for reasons of scale. None of my short stories were big enough to have that many moving parts.
Surprisingly, worldbuilding is the thing that I didn't find all that different, possibly because I don't really think I change my approach much between short stories and novels. (I can do exposition; I did a lot of it in Dragon Pearl. I just sometimes choose not to.) But that's another essay, probably.