Big question. Maybe we need to clarify some intents of the assessment before saying what is 'good'. In other words, determine criteria for good assesment. I will open this debate with a humanist perspective. I want assessment to ONLY have positive impact on students' mindset about their ability to learn. To only leave them with positive self eficacy, to only speak to the of capacity to improve would be a number one priority.

Well, on the one hand, there are many types of evaluations-- formal, less-formal, progress evaluations, semester, so-called terminal, qualitative, more subjective, etc.

The problem here will be to have any evaluation handle the standard guidelines of testing including (1) validity, (2) reliability, (3) accuracy, etc.

More importantly, one has to look at the washback effect of the evaluation on the entire system of learning and teaching.

An ideal assessment would incorporate oral, visual, kinesthetic feedback on a topic that would utilize the learning style of that individual.

Effective assessment, I believe, is closely tied to motivation and learning. Positive reinforcement of assessment results on the learners, especially younger ones, is crucial...no one will always eventually fail...with correct dose of motivation, self-improvement and desire to learn will set in.

Designing effective assessments should be based on what your objective is for the course or subject matter. You want to test the students’ knowledge from multiple perspectives. I think the best way to assess students is to provide authentic learning experiences based on the desire outcome--authentic assessment will be much easier to design when you know exactly what you want students to learn, to able to do and at what level.

Starting from the purposes of pupil assessment; to support learning (formative) and to report achievement resulting from learning (summative).

The Dept for Education (UK) are moving away from levels and their descriptors, what might the next assessment system look like?

I do agree, to a large extent with Zainab Jaafar's thoughts as they are close to reality, on the other hand knowing and discovering yourself along with your students will also help you understand how and what to assess or what and how to assess. There is another technical difference to be understood between assessment and evaluation and then examination, replacing and misunderstanding their application may make you confused as well.

to complete my first post, an outstanding system would (reasonably) accurately measure learning in a multivariate format so that the learning path could be tailored to meet student needs. It would describe a paradigm where students recognised the link between effort and success and success is defined as progress. The paradigm would speak to students of the collaborative nature of learning, so that the teacher was recognised as the learning expert, and students recognised and valued their guidance. The best outcome of the system would be that students valued 'doing a good job' above getting a good mark. For this to happen a key parameter needs changing in the current paradigm - how success is reported. Reporting on performance comparative to others is anathema to good assessment. Reporting needs to be of position along a continuum of learning for all of the relevant variables. This focusses the student on progress, while still allowing comparison when needed.

From personal experience there seems to be confusion between assessment for learning or formative assessment which has been evidenced as having a positive impact on pupil performance and assessment of learning which involves mainly monitoring pupil progress which seems to have a a negative impact on specific pupil cohorts

True Bobby, i admit that i assumed we were discussing summative because I regard formative assessment as a pedagogical process. The distinction is important.

Mark, of course yes. I think the problem lies with the use of the term "assessment" in formative assessment or assessment for learning which causes confusion.

Kevin. I'm glad you said that washback is the most important consideration. I have been having trouble onvincing others of its position in the hierarchy of considerations. Having you state it so clearly is a boost.

Mark, washback is neglected almost everywhere in every country, especially on high stakes testing and even too often in teachers training if ever. 

Mark Kevin, I'm uncertain what "washback" is, I wonder if there are any references that you could signpost me to?

I will give a personal description. I hope it is not too superficial for your tastes Kevin. I see it most commonly along the lines of 'unintended consequences'. The construct of the assessment (and reporting style) positions stakeholders; parents, students, teachers etc to draw conclusions, rightly or wrongly, about what is valued and to act on those conclusions  often in innappropriate ways. A good example is the actions taken by teachers in response to the no child left behind' system.

Here are a few articles where washback or backwash in testing are discussed.  

http://tapchi.vnu.edu.vn/nn_4_09/b.8.pdf

https://www.ets.org/Media/Research/pdf/RM-99-04.pdf

http://ltj.sagepub.com/content/13/3/241.short?rss=1&ssource=mfc

Here is a simpler definition than provided by some:

"Tests can have positive and negative effects, or washback. Positive washback refers to expected test effects. For example, a test may encourage students to study more or may promote a connection between standards and instruction. Negative washback refers to the unexpected, harmful consequences of a test. For example, instruction may focus too heavily on test preparation at the expense of other activities. Washback from tests can involve individual teachers and students as well as whole classes and programs.

Click here for an example of positive washback in action. http://www.cal.org/flad/tutorial/impact/5exwashback.html

One way to ensure positive washback is through instructional planning that links teaching and testing. By selecting a test that reflects your instructional and program goals, you can more closely align testing with instruction. The principles of backward design (Wiggins & McTighe, 2005) provide a helpful model for integrating teaching and testing." from: http://www.cal.org/flad/tutorial/impact/5washbackinstruction.html

Backwash (or washback) can also refer to design elements of the system within which the test is constructed. Eg My particular focus is the impact of reporting styles on teachers and students, independant of the quality and reliability of testing, which can have profound impacts for young students.

What Mark focuses on is the crux of backwash.  Negative backwash on the systems of teaching and learning should be avoided at all cost in testing design.  Alas, with high stakes exams, this is often de-emphasized--leading to dissatisfaction for teachers and entire system of delivery and acquisition. 

In other words, backwash observers note or ask that if a test is going to have a negative backwash on the system of learning and acquiring, you had better find another type of evaluation that does not offer negative backwash. 

Many thanks both, I'm clearer now.

Mark Gould put up this related question, you might try and follow through on with--and yes its related to our own test design based upon our own definitions of success and learning/acquisition. 

Question

Does the way we describe success in education affect students and teachers perception of their roles?

The notion of 'personal best' drives most athletes to keep training despite the knowledge that they may never be first or best in their field. In other words, they actually maintain their effort by seeking personal improvement and thus competing with themselves. In education the description of success is nearly always in terms of comparison with others. My contention is that this view of success is counterproductive for all but the top half of the cohort and we would be better off if we described success as 'improvement through persistent effort' as an athlete does. I also contend that teachers would benefit from this paradigm shift because it better describes what we should be doing, collaborating with each individual to maximise their potential and that the current paradigm of success confuses the role by positioning us to see lower performing students as having a deficit to be repaired.

https://www.researchgate.net/post/Does_the_way_we_describe_success_in_education_affect_students_and_teachers_perception_of_their_roles?_tpcectx=qa_overview_following&_trid=541743d5d4c1188a3b8b465f_

One must first of all make the distinction between assessment and measurement. Measurement is concerned with the development, administration, and scoring procedures use to gather data about students.   Assessment is the process of using the data gathered to make inferences about students/teachers/schools etc.

What should be measured is the level of mastery a student has achieved of what is set out by the course syllabus. An "outstanding" measurement instrument gathers data that has a high level of validity and reliability. If, for example, student scores 90% on a test. The score would indicate that the student has achieved mastery of 90% of the syllabus content.  

If the teacher had opted to teach other than what is on the syllabus, and set the test to measure what was taught, rather than what is on the syllabus, then the score would not be valid.

Testing agencies have procedures in place to ensure validity and reliability in their scores.  Classroom teachers may not.  One good place to start may be to ensure content validity through the use of a properly constructed table of specification.

I hope these may be useful:

http://www.aaia.org.uk/content/uploads/2010/07/The-Misrepresentation-of-Assessment-for-Learning.pdf

Also the following has comment, leads through other research in the references and has produced other work related.

Instructional Science

June 1989, Volume 18, Issue 2, pp 119-144

Formative assessment and the design of instructional systems

D. Royce Sadler

There has been a revolt against high stakes testing in my homeland.  What about yours?  

http://www.commondreams.org/views/2014/12/31/2014-greatest-year-revolt-against-high-stakes-testing-us-history

Kevin, I wouldn't describe it as a revolt yet in Australia. Many educationalists have complained for a while without effect, but there are signs of increased understanding that the many negatives outweigh any positives. Even a politician in power recently commented that he wanted to reduce the impact of high stakes testing, so a change must definitely be on the horizon.

I hasten to add that high stakes testing in Australia is NOTHING like that in the USA. Hopefully we can implement a change BEFORE actually going down the extreme high stakes route.

Not quite a revolt in the UK, however, we are British of course ;)