Statistical Associates Publishing is a website that offers guides focused primarily on correctly executing various statistical techniques. The guides are short and provide screenshots and step-by-step instructions for for some of the major commercial statistical packages. One can buy the books - each at very low prices $5USD or less per book from Amazon - or request a free copy from the publisher with some reasonable restrictions on use.
I should note that since I don't spend much time doing statistics myself, I feel a bit unqualified to attest to the quality of the materials. However, I'm a pretty good judge of user guides in general (having created a lot of really bad ones over the years) and I was able to glean some information from Amazon reviews and from researching the publisher himself. Therefore, I feel pretty confident in what I say below, but do take it with a grain of salt.
The publisher is G. David Garson. In addition to what he shares at the Statistical Associates website, he continues to publish pretty regularly and co-authored at least one entry in the Springer International Encyclopedia of Statistical Science (Toma, Roxana, and G. David Garson. "Research Designs." International Encyclopedia of Statistical Science. Springer Berlin Heidelberg, 2011. 1224-1227.)
The guides themselves are straightforward, understandable to the novice and most importantly, given that they include software-specific instructions, updated often. The use of screenshots to supplement written instructions is always a good idea as it helps the user orient herself and breaks up blocks of instructions. Most Amazon reviews were positive and the overriding theme was that the works are concise and good for users at different levels. One complaint from some reviewers however, was while the books help users figure out specific techniques or steps in the major commercial tools, they didn't necessarily tell you why you should use that technique. So, good for novices, apparently yes, but users with no exposure to statistics at all? Probably not.
An open question is whether these books would do more for a user than simply finding a similar video on YouTube. Again, I don't feel qualified to answer this from a content perspective, but I do think it's nice to have alternatives to video presentations for those who learn best from written instructions.
Libraries may purchase the collection as a package - the cost is extremely low - but I'm not sure how urgent the need would be. Much of the site is pitched to instructors and it seems to make more sense to view these books as supplementary course readings when the course involves working with particular statistical techniques. So as a possible source of low-cost course materials, Statistical Associates Publishing is probably a good source.
Update from the publisher:
I posed a few questions in my original post and the publisher responded as summarized next.
Statistical Associates does aim for introductory graduate students and, relative to YouTube, they vouch for their accuracy. Dr. Garson also made an interesting observation that "...most researchers use a single package and therefore never have occasion to cross-verify results with different packages. It is very common for researchers using the default settings on different packages to come to different results. We not only provide worked SPSS, SAS, and Stata solutions, but we reconcile them. It is very hard to find that information elsewhere on the net."
Certainly, in other contexts I've seen users go with defaults across programs and end up confused. This particular characteristic seems like an excellent reason for keeping this reource in mind when working with your students/users.