The most useful checklists to submit to eBird are complete checklists (every species detected is listed) and every species counted. If at all possible, you should accurately track observer effort--mileage and time. As you use eBird, more and more of your checklists will be as these.
The two main categories of effort-based lists are "Stationary" (you birded from a single location and didn't move more than 100 feet or so) and also a "Traveling" count. Traveling counts can be walked or driven (or both). Traveling counts are recommended to be 5 miles or less. If you change habitats, change to another checklist. On the other hand, if you are traveling through miles of identical habitat then it is not necessary to change checklists after 5 miles (think mono-species grasslands or pelagic trips). No eBird checklist should cross county borders. eBird Reviewers are instructed to weed out long trip lists that span county borders or multiple habitats. Thus, if you enter such lists the data will not be used in the maps and bar charts. But they will appear in your personal lists.
Sometimes, though, you see a noteworthy bird when you aren't birding. Or, you are entering sightings from years past--before you started eBirding--that lack effort information. This is the time to enter an "incidental" (renamed from "casual") list. You may or may not enter a complete list.
I've used incidental checklists when entering old year lists that listed only a new county bird for the year and location. I used an incidental list today. I was driving to a birding location and noted 35 Turkey Vultures on the way. I pulled over and used Bird Log to enter the exact location.
Some birders balk at creating multiple lists for one birding excursion, for instance, while doing a Christmas Bird Count or Migration Count or even a Big Day. However, when you think about it, you are really birding intensely at only a few discrete locations, and then you see a few other noteworthy birds en route that will be entered on individual incidental lists.
An old eBird article that uses the obsolete term "casual" observation is here: