This is an old topic in model railroading, so I will be brief, but I think there are important points which deserve repeating.
Most of us do build layouts primarily for ourselves, and that’s often stated if any aspect of a layout gets criticized. Of course we have modeling friends and perhaps an operating group, and knowledge of those folks’ perspective will affect what we do and how we do it. And it is natural to regard these layout viewers as the “expert” audience. But I think in a number of ways, the real test comes when the layout is viewed by what one of my friends calls “civilians,” people with no particular knowledge of the hobby.
I know from having shown my layout to neighbors, friends and relatives over the years, that such visitors indeed do not know very much about any prototype railroad, and so could accept just about any railroad equipment they might see. But they have a surprisingly good sense of era. They do grasp if your era is the 1930s or the 1950s or the 1970s, partly by motor vehicles, partly by signage, partly by the clothing of figures. And they do notice anomalies within those eras, such as a girl in a miniskirt in a 1940s layout scene, or a 1980s minivan in a street scene from the 1950s. I remember visiting a layout set in “the present,” and I heard another visitor, not a modeler, say, “How come there are no graffiti?” And they readily notice out-of-era logos for familiar consumer products, for example on billboards, that are too modern or too old-fashioned.
So would I advocate that all you need to do is be “scenery consistent” in developing a layout? Not entirely, no. Those same “civilian” visitors will notice an unweathered freight car in a train of weathered ones. They will sure notice an unpainted brass engine (though they may find it more impressive looking than your painted and weathered ones). On a pre-1955 layout, which means nearly all freight cars will be boxcar red, or black, or yellow/orange reefers, they will sure notice any brightly colored cars (a few, of course, are prototypical). And they will probably notice steam locomotives mixed with modern wide-nose diesels
This leaves to one side the more flamboyant examples of freelancing. I vividly recall an NMRA convention at which one model entry was a trio of Model Die Casting ore cars, somewhat upgraded in details, but painted lime green, lettered for the Jimmy and Suzy Railroad (the entrant’s children, as he proudly explained), and of course entirely unweathered. I believe your average visitor will recognize that models of that type do not represent part of the real railroad world.
But even with nominally realistic modeling, issues can arise. In thinking about my hypothetical visitor to your layout, whether a modeler or not, you may well think, “Hey, I can explain.” Sometimes that is natural. Yes, the MKT really did have yellow box cars in the 1940s and 1950s, and so forth. But I maintain that every time you have to explain something, you are losing some of the plausibility of your layout that you have (presumably) worked pretty hard to achieve.
That is the core point I’m trying to make in this post. Be as consistent and “mainstream” and “average” in your modeling as you can, consistent of course with your particular prototype era and location. That’s the basic reason I try to avoid modeling very many rarities or oddballs, whether rolling stock, or automobiles, or billboards, or structures, or anything.
But as I mentioned above, many modelers think of their primary audience as the fellow modeler who knows something about the railroad being modeled, especially for the era which has been chosen for the layout. Now the things that might call for “explanation” are more specialized.
A number of modelers, including my friend C.J. Riley, have placed photos, drawings, or other materials on the fascia about the layout, to silently document what is seen to be modeled nearby. Then the modeler can say to himself or herself, “Oh, this model packing house really looks like its prototype,” or whatever the case may be.
These same considerations are part of what made me change my former layout, conceived as an imaginary short line, the Lompoc & Cuyama, which interchanged with the Southern Pacific, into an SP branch line. I discussed this in a post about my choice of layout locale (you can view it at this link: http://modelingthesp.blogspot.com/2011/01/layout-design-locale.html ). The branch line itself is still imaginary, but the Southern Pacific, and the area of California modeled, are real.
Best of all, it seems to me, is if the knowledgeable visitor, likely a fellow modeler, walks into your basement or layout room and says, “Ah, Hinton!” or “Oh, it’s the Oakland Mole,” or even “Hey! Sherman Hill!” Ideally, they would then go on to observe further, something like “I see it’s 1940” or whatever corresponds. You have not had to “explain” anything, because your modeling speaks for itself.
Chuck Hitchcock in Kansas City used to have a fine layout (now replaced) which modeled the transition-era Santa Fe in Kansas City and environs. [See Kalmbach’s Great Model Railroads, 1991.] I have heard Chuck relate how, when Richard Hendrickson visited for the first time, he walked into the layout room, where about the first thing you saw was the Kansas City depot. Noticing the train standing there, Richard promptly said, “Fast Mail, 1951,” and as Chuck said, sounding a little surprised, “He got it right, too.” I think many of us wish for both those things: a layout which is that accomplished in place and era, and a visitor who can appreciate what we have done.