Tom Reeves and Ron Oliver have presented points that make me think that there may be hope for the field after all. Even more gratifying has been the quality of the responses.
I agree with Tom that most of the IT literature is of little lasting value. But I also agree with Ron that the main reason for the low quality is that this literature is mostly motivated by the survival needs of the academics who generate it. You only have to read the insightful reactions of people like Bev Garcia to know that there is no lack of understanding of ways to address the needs of the larger society.
Gatekeepers have certainly influenced the literature to some extent. If you tried to publish in the early 60s, you will recall rejections from the theoretical journals that claimed that your research was "too applied" and from the application journals that claimed that your research used "resources that are too expensive for anyone to use."
But, as a former gatekeeper, I can testify that few deserving studies are being rejected these days because they fail to meet overly narrow views of what constitutes "useful research." During my last decade of serving as editor of the Journal of Computer-Based Instruction, I was also reviewer for several other journals and peer reviewer for most of the funding agencies that support IT research. I typically read about 250 manuscripts and 50 research proposals each year. Most were neither published nor funded. Exemplary work was rare, whatever its approach. It was simply not the case that good research was being turned away from journals and funding because it used techniques unfamiliar to the gatekeepers. In general, the bulk of research using innovative research approaches was as poor as that following textbook quantitative methods. The reason is simple. It takes time and effort to "do it right." Most parties do not have the resources to support that time or effort.
The academic in need of documentation of her/his research prowess can crank out a half-dozen toy studies in the time that it would take to do one proper study that would have real impact on the field. Under present constraints, most academics would perish as researchers if they attempted to do only socially responsible research. The quick generation of publications that demonstrate only mechanical competence in finding references and analyzing data is rewarded not only by academic departments (including many who claim to be looking for quality rather than quantity in publications), but also by journals (which are now in such overabundance that most would quickly go out of business without a ready supply of reasonable-appearing articles).
Funding agencies are little better. Even many of the private agencies are under pressure to fund a wide variety of projects, even if the quality is marginal. An agency that fails to use all of its appropriated funds risks having its appropriation cut (along with the positions of the staff that decides what to fund). Funding agencies nominally seek to improve some aspect of society, but that goal is usually measured in the short term and the survival of the agency depends on its having a reasonable ratio of successes. That translates into funding of projects and investigators that are geographically diverse, and that look like they have a reasonable track record. How do you tell that an investigator has a reasonable track record? The "easy" way is to count publications.
In truth, it is not quite as grim as the above might suggest. Editors prod authors to do a bit more than the minimum in demonstrating competency and relevance. Sometimes a string of seemingly random studies coalesces into a nicely programmatic sequence that leads to a useful contribution. Funding agencies do risk support of investigators who are willing to expend effort over a long period of time in search of information that has a potential for a major payoff.
But the trend of rewards is not toward support of socially (or even technically) useful work. Salomon's complimentarity of analytic and systemic approaches works, but at a greater cost than many can afford to expend. A moratorium on analytic studies is more likely to lead to discovery of ways to do cheap, inappropriate systemic research than to production of truly useful research. In fact, the proliferation of "new" theory and research techniques is one symptom of the problem, rather than a source of hope. Even when the groundwork is valid, these approaches rapidly generate a bandwagon effect. They are quickly re-interpreted in a superficial fashion that allows their use as a basis for quick and simplistic publications by those who do not have the time to come up to speed in a more established quasi-scholarly niche.
What is the solution for those who really care and want to leave the world a better place? If you are an academic in an institution that has a strong research orientation, there are no simple solutions. Your short-term survival may depend on your playing the game enough to bring in grants and contracts that help support the institutional overhead. It is not enough to be a good teacher or to be widely published in good journals.
One solution is to engage in guerrilla research. Find ways to do studies that do not rely on institutional or government funding. That means that you will not normally be able to work with the latest hardware, but it also means that you will not have to spend time as the sacrificial test site for half-baked hardware designs. It means that you will have more difficulty in finding cooperative institutions and colleagues, but it also means that the people you work with will be genuinely interested in improving learning, and not simply looking for a way to get new equipment or released time. It means that you will rarely be able to crank out six papers a year, but it also means that the papers you do write will have impact on something beyond the length of your vita.