This article from teachonline.ca draws from a report by JISC (the UK academic network organization) to provide 5 'principles' for assessment. I put the scare quotes around 'principles' because they are mostly descriptive labels for trends and they are woefully non-inclusive. There is also a subtext here - that I do understand is incredibly hard to avoid because I failed to fully do so myself in my own post last week - that assessment is primarily concerned with proving competence for the sake of credentials (it isn't). Given these caveats, most of what is written here, however, makes some sense.

Principle 1: authentic assessment. I completely agree that assessment should at least partly be of authentic activities. It is obvious how that plays out in applied disciplines with a clear workplace context. If you are learning how to program, for instance, then of course you should write programs that have some value in a realistic context and it goes without saying that you should assess the same. This includes aspects of the task that we might not traditionally assess in a typical programming course such as analysis, user experience testing, working with others, interacting with StackOverflow, sharing via GitHub, copying code from others, etc. It is less obvious in the case of something like, say, philosophy, or history, or latin, though, or, indeed, in any subject that is primarily found in academia. Authentic assessment for such things would probably be an essay or conference presentation, or perhaps some kind of argument, most of the time, because that's what real life is like for most people in such fields (whether that should be the case remains an open issue). We should be wary, though, of making this the be-all and end-all, because there's a touch of behaviourism lurking behind the idea: can the student perform as expected? There are other things that matter. For instance, I think that it is incredibly important to reflect on any learning activity, even though that might not mirror what is typically done in an authentic context. It can significantly contribute to learning but it can also reveal things that may not be obvious when we judge what is done in an authentic context, such as why people did what they did or whether they would do it the same way again. There may also be stages along the way that are not particularly authentic, but that contribute to learning the hard skills needed in order to perform effectively in the authentic context: learning a vocabulary, for example, or doing something dangerous in a cut-down, safe environment. We should probably not summatively assess such things (they should rarely contribute to a credential because they do not demonstrate applied capabilityre), but formative assessment - including of this kind of activity - is part of all learning.

Principle 2: accessible and inclusive assessment. Well, duh. Of course this should be how it is done. Not so much a principle as plain common decency. Was this not ever so? Yes it was. Only an issue when careless people forget that some media are less inclusive than others, or that not everyone knows or cares about golf. Nothing new here.

Principle 3: appropriately automated assessment. This is a reaction to bad assessment, not a principle for good assessment. There is a principle that really matters here but it is not appropriate automation: it is that assessment should enhance and improve the student experience. Automation can sometimes do that. It is appropriate for some kinds of formative feedback (see examples of non-authentic learning above)  but very little else which, in the context of this article (that implicitly focuses on the final judgment), means it is a bad idea to use it at all.

Principle 4: continuous assessment. I don't mind this one at all. Again, the principle is not what the label claims, though. The principle here is that assessment should be designed to improve learning. For sure, if it is used as a filter to sort the great from the not great, then the filter should be authentic which, for the most part, means no high stakes, high stress, one-chance tests, and that overall behaviours and performance over time are what matters. However, there is a huge risk of therefore assessing learning in progress rather than capability once a course is done. If we are interested in assessing competence for credentials, then I'd rather do it at the end, once learning has been accomplished (ignoring the inconvenient detail that this is not a terminal state and that learning must always undergo ever-dynamic renewal and transformation until the day we die). Of course, the work done along the way will make up the bulk of the evidence for that final judgment but it allows for the fact that learning changes people, and that what we did early on in the journey seldom represents what we are able to do in the light of later learning. 

Principle 5: secure assessment. Why is this mentioned in an article about assessment in the digital age? Is cheating a new invention? Was it (intentionally) insecure before? This is just a description of how some people have noticed that traditional forms of assessment are really dumb in a context that includes Wikipedia, Google, and communications devices the size of a peanut. Pointless, and certainly not a new principle for the Digital Age. In fairness, if the principles above are followed in spirit as well as in letter, it is not likely to be a huge issue but, then, why make it a principle? It's more a report on what teachers are thinking and talking about.

The summary is motherhood and apple pie, albeit that it doesn't entirely fall out from the principles (choice over when to be assessed, or peer assessment, for instance, are not really covered in the principles, though they are very good ideas).

I'm glad that people are sharing ideas about this but I think that there are more really important principles than these: that students should have control over their own assessment, that it should never reward or punish, that it should always support learning, and so on. I wrote a bit about this the other day, and, though that is a work in progress, I think it gets a little closer to what actually matters than this.