What IS the Best Way to Assess Students' Learning?

Standardized testing is all the rage today.  Thanks to No Child Left Behind, it is just about the only thing you hear about anymore when education comes up.  But what are we really measuring, and does it help student learning at all? 

I read an article called Measuring What Matters Least which lay out some of the worst factors about standardized testing.  First of all, multiple choice tests rarely tell us what students know, they tend to measure who is a good test taker.  This, however, is the mode most common in standardized tests, probably in large part because it is simple to score. However, open-ended or free-response items are equally poor as a measure of student achievement.  At first this seems odd, however, one has only to think about the dreaded "Extended Response" items found on the ISAT test to know why.  We spend an inordinate amount of time preparing for these extremely rigid types of responses.  I have even seen teachers develop basic "fill in the blank" type general responses that they make their students learn so that they can adjust their writing to the specific question come testing day.  Are these students learning? Of course.  They are learning to think that there is only one "right" answer in the world.  They are learning not to think on their own.  They are learning to write like robots.  None of these are the intended consequence of course, but with the pressure dealt to many schools to raise test scores- OR ELSE!- this is what we wind up with.  

Another standardized testing sin is the norm-referenced test.  A norm referenced test, like many commonly used today, measures student achievement against everyone else who took the test.  There is always a top and bottom 10% and always a middle 50%.  This does a horrible injustice to kids on the bottom.  How does someone in the lower 10% get out?  They have to progress past a certain percentage of other students.  This system does not honor students who show great personal growth and sets up an atmosphere of inappropriate competition.   

So what should we do?  In the article If Not Standardized Testing, Then What?, some measures of alternative assessment come up such as portfolios, written reports to parents, and parent-teacher conferencing as ways to help show the complexity of student learning. Interestingly enough, they also show some conditions to help make testing as viable as possible if in fact we cannot tear ourselves away from them.  For instance, they suggest no multiple choice tests, using criterion referenced tests, not using time limits, and testing less frequently.  This may be the most viable place to start as we think about changing the ways we assess students' learning.  As much as I would like to be able to use some other methods, I'm not sure the general public is ready for portfolios quite yet (and I don't think many teachers and school districts are ready for them yet either).  The No Child Left Behind era has made people think that a number can be the answer for determining student learning, effective teaching, and effective school management.  It will be a long time before people are ready to start addressing the variables involved in education and how to account for them.  

In the article Alternatives to Standardized Testing there are some descriptions of pros and cons of some of the alternatives mentioned earlier.  The article starts by saying that the primary role of assessment should be to improve student learning.  They go on to say that the general public needs to know that teachers assess students regularly as part of their teaching, so repeated testing throughout the year is a waste of time.  In fact, the goal of all the alternatives in this article value showing growth over a long period of time instead of the pressure of "showing what you know" in one sitting.  

First mentioned is portfolios.  A student portfolio in it's best use should be a collection of the student's work over time.  The work selected should be the student's choice of work to allow the student to reflect on their own learning.  At the end of a marking period, the teacher would use a scoring guide to score the portfolio.  Some would also have the student and/or peers score the work as well.  In order to ensure validity and reliability of scoring, schools could use "random sampling".  This means that an independent group could pick random portfolios from each classroom to compare scoring.  Furthermore, the random sampling could be used to show and compare trends across classrooms and school-wide.  This method is excellent for students because it honors their individual growth over time.  There are several negatives in portfolio assessments as well.  Portfolios work best with quality teachers who really know how to use them effectively.  There would need to be extensive professional development for anyone learning to implement them.  In addition, if schools mandate what has to be in a portfolio, then the curriculum can slant toward those particular activities.  Lastly, scoring is extremely subjective and a lot of work would need to go into addressing how to best deal with this.

I think it is this subjectivity that is the biggest obstacle to alternative assessments.  In the NCLB era, teachers are looked at as wanting to shirk accountability if they are against standardized testing.  However, I think most teachers are perfectly agreeable to being held accountable for doing their jobs as long as the method for evaluating this has anything to do with student learning and achievement.  Standardized testing has too many negative effects on students and is too unfair for us to say that it is a true measure of our students.  Are portfolios the answer?  Maybe, maybe not.  Perhaps it is a combination of approaches that is needed.  However, the most important thing to do now is to open up the debate and start objectively teaching parents and the public about alternatives to standardized testing and why they may provide a better, more fair way of looking at our students.