The phrase, "No significant difference" has been tossed around for years, usually in relation to something education-based: technology in education, technology in higher education, pay-for-merit teacher salaries, distance learning programs, etc.
The kicker is that the phrase itself can be used to prove a particular point or to disprove the same point. For example, a decade ago, studies suggested that students taking classes through distance learning (more specifically, through interactive live video services) produced "no significant difference" in grades, assignment completion, etc compared to students who attended 'traditional' classes (that is, teacher live in the room).
This finding was used to convince schools that putting in distance learning systems would have little to no impact on graduation rates, grades, etc. and that they should jump on the bandwagon.
The same is being said for "target testing" these days, what with NCLB, AYP, API, and a myriad of other acronyms.
Maybe it's just my particular viewpoint, but my response to "no significant difference" surfaces as "then why bother?" Why spend all the money to implement programs that ultimately show no significant difference than the programs already in place? Why add to the already overworked, overstressed workload of teachers, principals and students if there is no significant difference between what they are already doing and instituting some new program that will cost the district (and/or state taxpayers) additional money? What is the point?
As for distance learning, I see the implementation as one of providing opportunity. That is, schools should only use distance learning programs as a way to enhance student learning. Many small schools cannot offer the same variety of classes as many larger ones (different foreign languages, various alternate career courses, etc). I see that as a HUGE benefit. Small schools (also any-sized economically challenged ones, etc) cannot afford to hire a fulltime teacher to reach 5-10 students interested in taking a particular course (or to take college credits while still in high school, etc). In these cases, the cost of a distance learning teacher can be shared among several districts for one teacher to teach multiple sites.
As for testing? I have no idea. I would love some input, insight, direction on that one. If target testing (target assessment or whatever your district may call it) is showing to have no significant difference (NSD), then why bother?
Now, I do understand that the problem with 'no significant difference' centers upon the question(s) being asked and upon the population of respondents/study subjects. That is, NSD would be much more pronounced over a larger population surveyed than a targeted one. I know of many teachers who love the assessments because they use the data to help reach their students where the students lack skill and understanding. Other teachers might read this and have no idea what I'm talking about. Perhaps the issue is not that testing makes NSD, but rather what is *done with* the information the testing provides.
In general, though, if a program (ANY program) is shown to make no significant difference, then I question the costs, the time, the resources involved with implementing and keeping up such a program. Sometimes, NSD is not a good thing, but should instead serve as the point at which administrators, legislators, and taxpayers stop and say, "then why are we doing this?"