Tuesday, November 10, 2009

How Do We Determine if Certification and Credentialing are Effective?

The importance (or lack thereof) of certification and credentialing seems to be a popular topic these days.  We have plenty of evidence that little to no difference exists (in terms of impact on student achievement scores) between teachers that are certified, have a master's degree, etc. versus those who are not.  So we can conclude that certification and credentialing are ineffective, right?  I heard that conclusion drawn at a presentation I recently attended, and I've read similar statements countless times in news reports and on blogs.

But hold on a minute.  To say something like "certification (or credentialing) is ineffective" is a strong statement.  Before you say something like this, or believe others who do, let's take a look at what we'd need to know in order to justify making such a statement.

1.) What are the purposes of certification and credentialing?  Historically, I believe certification has been designed to ensure that completely unqualified people don't enter a field.  For example, hiring a certified electrician means they probably won't burn your house down.  I'm sure there are plenty of self-taught electricians who would do the same, but by not hiring a certified electrician you also run the risk of hiring somebody like me who would create more problems than they would solve.  Many seem to assume that when somebody is certified it means they're better at it, so I think it's fair to assume -- at least based on the reaction to certification -- that it should also increase the mean quality (or at least indicate that certified workers are better than uncertified ones).

2.) How do those who completed certification and credentialing processes compare to those who didn't?  This is what most people try to measure.  In terms of education, they usually look at the mean value-added score for certified versus uncertified teachers, controlling for other factors.  We can also examine the mean change in performance as teachers complete various credentialing processes.  This partially answers the question, but we also need to know how the distribution of quality changed as a result of the process -- including whether there would be more really bad teachers if certification didn't exist.

3.) How do certification and credentialing processes change the make-up of the field?  Lastly, comparisons between certified and uncertified teachers, for example, may fall short if only certain types of people are allowed to enter the field without certification (e.g. high-achieving TFA and TNTP members).  In this sense, we're really answering whether certified teachers are measurably different from a select group of uncertified teachers -- not how the field would differ if we abolished certification.  In the case of the latter, we might see a completely different group of people who decide to enter and remain in the field.

So before we draw any hasty and uninformed conclusions about the efficacy of teacher certification, let's make sure that we're asking and answering the right questions.  We need to know both about changes in mean scores and changes in distribution of performance before we reach a conclusion.  And we need to make sure we don't assume that ending certification won't change the field in ways we don't anticipate.


Roger Sweeny said...

As a matter of simple math, if certification keeps out the lowest performers, it has to increase the average level of performance. Keeping out the bottom feeders will boost the average

The fact that, on average, uncertified teachers seem to do as well as certified teachers may well mean that uncertified teachers are coming from a different (and more promising) demographic.

Doing away with certification and credentialing might well increase the number of both good and bad teachers. A significant number of people who would make good teachers don't want to spend one of two years in the brain-bleeding fantasy world that is ed school (okay, perhaps I exaggerate). On the other hand, without any certification, some really poor performers are going to be applying for teaching jobs.

Corey Bunje Bower said...

One group can have both a higher average performance *and* more lower performers than another group.

And remember, looking at current teachers who are uncertified isn't a counterfactual to the existence of certification.

TFT said...

Teaching is more like acting that plumbing.

Some actors are naturals. Some take lots of acting classes and some don't, but both groups may produce good actors.

Good plumbers know how so solder, relieve pressure, understand crimping, cutting and measuring, all due to rather concrete processes spelled out over years of plumbers plumbing and mentoring. All sorts of plumbing innovations have made plumbing more efficient, but you need to know what these things are and how to do them. Hence apprenticing for years, spending lots of time at the hardware store learning about all the fittings, sizes, tools and the rest. None of that stuff comes naturally.

Teaching, less so. If you know your subject, that's just a start; unlike plumbing. To teach your subject you need more than the content knowledge. And what you need is personality and whatever it takes to get a bunch of kids to want to listen to you and do what you are asking them to do. No amount of training, like a bad but well-schooled actor, will make you a good teacher.

Very few people understand how to plumb so they need lots of training.

Many people know how to teach and don't need to go teacher school to learn about BICS, CALP and ZPD's.

So, education programs seem to be pretty bad, yet good teachers teach well in spite of them. I do not know of many teachers who had good things to say about their programs, and I have never heard anyone say their program helped them with the "craft" of teaching which requires a certain knack. No knack, don't even try. Got the knack, maybe you don't need teacher school.

My program was an embarrassment.

TFT said...

Teaching is more like acting than plumbing.


Roger Sweeny said...

One group can have both a higher average performance *and* more lower performers than another group.

Of course. But very unlikely if both groups are drawn in the same way from the same pool.

Let's say you could rank every adult in the United States on their raw "teaching ability." If you randomly picked people and assigned them to be teachers, you would get a group whose performance would be very, um, diverse. If somebody else randomly picked the same number of people to be teachers, their performance would also "have a high variance" but on average would be about the same.

If you screened out the people who ranked in the bottom half of "teaching ability", the average performance in both groups would go up. It would go up by about the same amount, so neither group would end up noticeably better than the other.

If credentialed teachers and uncredentialed teachers were randomly drawn from the same group, they would start out with the same average performance. If you then screened out the worst of the credentialed group, their average performance would be substantially higher than the non-credentialed group

Since the performances don’t seem to be substantially different, I am led to believe that the uncredentialed are drawn from a more promising group, or they are subject to different effective screens (that keep out the worst) or both.

(Another logical possibility is that both groups start with the same potential but ed school actually makes you a worse teacher. However, the process of actually passing classes and getting the certification washes out enough bottom feeders to bring the average up to the average of the uncertified and uncorrupted. But I’m not willing to go that far :) )

Corey Bunje Bower said...

they're not drawn from the same pool

Brian Rude said...

If we had only statistics and research to go by I would have to agree that the issue is not settled. But that is not all I have to go by. I went to ed school, admittedly many years ago, but there seems little reason to believe much has changed. I personally experienced the frustration of finding out first hand when I started teaching that ed school had totally ignored some important areas (especially discipline) and had been dead wrong in other areas (like "children really want to learn", a gross approximation at best). This did not come as a surprise to me. I was well aware before I started teaching that conventional wisdom was that you have to take the ed classes, but they won't do you any good. So my long held conviction that we would be better off without credentialing controlled by ed school will not be easily swayed.

Your post seems to have the premise that so long as the issued is not firmly settled we should go with credentialing as the default position, giving the benefit of the doubt to ed school, and putting the burden of proof on doubters. I see no reason for this premise. It's not like "innocent until proven guilty". I'm not charging ed school with a crime, just incompetence. So my view is that it is sensible to take as the default position that credentialing is not effective. The burden of proof is firmly and squarely on those who would argue that credentialing is effective.

Brian Rude said...

(continuing) You point out that to say that credentialing is ineffective is a strong statement. Sure, but I would point out that to say that credentialing is effective is also a strong statement, a statement for which the evidence is not conclusive, not even close. I would start with the premise that credentialing by ed school should not be required until it is proven definitely that it works. If someone promised to make it rain I would consider that a strong statement, and I would see no reason to have the default assumption of competence. I would take the default assumption of incompetence until the evidence is overwhelming, or at least convincing, in the other way.

To say that without credentialing a lot of poorly prepared people would apply for teaching jobs is true enough I suppose, but that is quite a different thing from saying they would be hired. Presumably school administrators know something about who they want to hire and who they don't. I think it's a matter of simple logic that credentialing limits the field of candidates. To think this is a good thing, one would have to something more than agnostic about the benefits of credentialing. Since I am something less than agnostic on the issue my view is that limiting the field of candidates is a bad thing. Doesn't the existence and relative success of TFA argue that it is a lot more than the poorly qualified people who would like to teach if given the chance?

I am not one to say we should ignore studies and statistics. But I am one to discount them considerably and quickly. I apply this skepticism broadly in all the social sciences. Every study has its critics who point out flaws, sometimes obvious and sometimes only potential. The social sciences cannot possibly have the rigor of physics, or even biology. Results will always be muddled and compromised to one degree or another. Thus it becomes very important to which side you give the default position, to which side you give the benefit of the doubt, which side you accept as the conservative interpretation, and on which side you impose the burden of proof. I do not consider accepting the value of credentialing controlled by ed school as the default position. Ed school has not earned it. And we have about a century of ed school experience to draw on.

I have developed my thoughts on ed school extensively in my article "A Personal Indictment Of Ed School" on my website. It's at http://www.brianrude.com/indict-ed.htm . I don't consider my position "hasty and uninformed". I consider it the result of a lifetime of experience, learning, observation, and careful reflection.

Corey Bunje Bower said...

I've never said that current certification and credentialing processes are effective (as a side note, my experience with teacher training much left to be desired). I simply wanted to think through what we'd need to know to determine that it was effective -- or ineffective.

As to what TFA proves, it does not prove that certification is ineffective. It proves that there are some talented and motivated people out there who will go into teaching, at least for a little while, if given the chance to cut through some red tape and told that they will change the world. And it proves that when a well-organized group recruits a large pool of applicants, carefully selects very few from this pool, and then trains a small number over the summer that they will perform at least adequately in the classroom.

Brian Rude said...

In the last year or so I have been coming to the conclusion, which I have stated now an then, that the basic resources needed for teaching are general intelligence, specialized subject matter knowledge (depending on the grade level), communication (mostly meaning verbal) skills, general social skills, knowledge of cultural expectations of the culture in which the teacher teaches, and that's about it. If this is the case it would seem to mean that most anyone can teach. This fits my general impression of the world I live in and observe. Yes, most anyone can teach. Indeed we see it everyday. The average burger flipper at a fast food place is quite capable of teaching me how to flip burgers at a fast food place. The average person on the street is quite capable of teaching me the way to the post office if I ask them and they know the answer. The average Dad who can deal with a leaky pipe, or the average Mom who can deal with a Thanksgiving dinner, can teach their children to do the same. And they do, everyday. Teaching of one sort or another is a part of everyday life for almost everyone.

Certainly it can be argued that these examples of teaching are far removed from what teachers do in the classroom. It can be argued that they are trivial examples. Well, they certainly may be trivial, but that does not mean they are irrelevant. And it can be argued that they are examples of very inefficient teaching. Efficiency often doesn't matter much. If I ask for directions to the post office it usually doesn't matter if I get my information from a stranger in 10 seconds, or two minutes, or even more. If a mother takes two weeks, or two years, or twenty years to teach her daughter to cook, it doesn't matter. But in the classroom efficiency does matter. So it certainly can be argued that teaching in a school is something apart from the incidental teaching that people do every day. Does that mean we need pedagogical theory from ed school?

I have an opinion on that. Yes, I think we need pedagogical theory. Pedagogical theory could do wonders, or at least a lot, to make good teachers out of ordinary people - ordinary people who have reasonably good general intelligence, appropriate subject matter knowledge, reasonably good communication skills, and reasonably good social skills. We need pedagogical theory. It's too bad ed school doesn't have any.

I think pedagogical theory can be valuable, but only if it's good theory. I have no evidence that ed schools are in possession of anything deserving of the label "pedagogical theory". What they have instead, as I have argued in my article I mentioned in my previous post, is ideology. They firmly believe a lot of things. They have a lot of good intentions. What they do not have, at least so far as I am aware, is a simple understanding of the mundane teaching and learning that goes on in classrooms, and many other places, everyday.

Or do they? Am I missing something? Let me know. I'm all ears, but I'm not holding my breath.

Attorney DC said...

I agree with TFT and Brian Rude that certification is not effective. My certification program was almost completely useless. In addition, it cost me several thousand dollars in tuition plus lost earnings for the year I attended the full-time program.

Interestingly, I taught private school prior to enrolling in my certification program, and taught public school after "graduating" from my program (and being officially "certified" in the eyes of the state), and I saw little, if any, difference in my teaching skills. I did learn some on-the-job skills from the student teaching component, but would have learned more if I'd been permitted simply to TEACH that year (or at least be paid as a teaching assistant).

What really rankled about my credentialing program (in California) was that I had to delay entry into the program for a full year in order to take a mandatory 3 credit class (only offered in California colleges) on "multiculturalism." Having attended college in another state, I had not taken this course.

Of course, none of this had any impact on teaching skills. Not to say that there should be NO standards (such as a basic subject competency test or a college diploma), but these endless ed school courses and requirements are nuts. I say, if we can get rid of credentialing programs, good riddance!

Corey Bunje Bower said...

I think we're confusing the educational program we went through before becoming certified and the actual requirement that teachers must be certified before entering the classroom. Certification usually requires a number of things, including, but not limited to, having taken some relevant coursework, passing some sort of test, and possessing at least a bachelor's degree. It's possible for the certification process to serve a useful function (for example, keeping the least qualified people out of the profession) while the teacher training programs people attend are simultaneously not very helpful.

Corey Bunje Bower said...

And I have to reiterate again that I was asking what we'd need to know in order to determine if certification and credentialing are effective or not -- not arguing whether or not they are. In other words, it was a piece exploring research methods rather than opinion.

Attorney DC said...

Corey: Sorry, guess I got a little off track. I do think you have a good point: How to determine if certification requirements are effective. An interesting point of view may be to look at private schools, which don't usually require certification, and see who they hire, given their ability to pick and choose certified vs. uncertified teachers. Of course, this isn't a random sample.

It does seem that certification will act as a barrier to some qualified people entering the profession, as well as a barrier to keep some unqualified people out of the profession. I think the problem with today's certification scheme is that the requirements don't tend to be very difficult, yet they are pretty cumbersome. Thus, people who are willing to go through the red tape will be able to teach (even if they aren't particularly good) and people who aren't willing to go through the red tape can't teach (even if they'd be great).

TFT said...

I think a BA shows that a prospective teacher has what it takes--cognitively. Of course the ability to graduate from college is not ALL it takes, and that other component cannot be discovered in ed school because that other component has little to do with academic success.

Rather, it has to do with an ability to excite students, hence my previous comment about acting; teaching is more of a social, dynamic thing that requires something you can't learn in ed school.

Maybe you can learn it at summer camp, but not in ed school.

We have all probably had a professor or two who was a brilliant scholar but a terrible teacher. I remember these discussions in graduate school. Those professors who just couldn't teach were the bane of us students. I think this clearly applies to k-12 as well.

There is no substitute for mentoring and on the job training.

Attorney DC said...

TFT: Well said.

Brian Rude said...

What would we need to know in order to determine if credentialing is effective or not? I'm not sure. But I think my perspective might be different from yours, Corey. Correct me if I'm wrong but I can't help reading in to your words that you have in mind the idea of designing and carrying out some sort of study, of collecting data, and doing a statistical analysis of that data. I'll call that the "count something and do stats" approach. I don't wish to disparage this approach. It can be a very powerful tool in some cases, but ineffective in others. A contrasting approach can be characterized as "look closer and you'll see more". This approach starts with a lot of careful observation, followed by careful reflection and analysis of what is observed, followed by making hypotheses and observing more to prove or disprove them. And this approach assumes that knowledge accumulates, often from unanticipated sources and in unanticipated ways. I would argue that this approach produced the theory of evolution, as one example. Like the "count something" approach, this approach can be powerful in some situations, and ineffective in others. Both approaches have their strengths and weaknesses. In the first approach opportunism has little place. In the second approach opportunism is everything. In the first approach anecdotes have no value. In the second approach the value of anecdotes vary widely, anywhere from zero to earth moving.

And I would argue that both approaches have some irreducible philosophical base. There has to be some consideration to values. Our ideas of good and bad will influence what we wish to count and do stats on in the one approach, and what we want to focus on and look for in the other.

So you've asked a very good question, Corey. But I don't expect anything like a definitive answer to emerge anytime soon. I think the issue will always be muddled technically and philosophically, and perhaps in yet other ways.

So I tend to be more concerned with the simpler question - is credentialing effective? By the first approach the answer is inconclusive, and I think it always will be. By the second approach it's a slam dunk decision.

At least that's the way it looks to me.

Corey Bunje Bower said...

I'm not only advocating counting something and doing stats, but also thinking through what the comparison group and counterfactual should be.

Roger Sweeny said...

Alas, real life makes it impossible to get perfect comparison groups and our ignorance (and the limitations of the data) makes it impossible to "statistically control" for that imperfection.

One of my favorite "excluded variables" stories (though I can't vouch for its fairness):
If you don't want the long motivation, you can scroll down to the paragraph beginning, "Almost thirty years a study was published ..."

Roger Sweeny said...

It's interesting how, when you start thinking about something, you notice it all over. On the question of "what do we know and how do we know it?":



Roger Sweeny said...

And this (especially the question, since being overweight is correlated with breast cancer, will losing weight reduce the risk of getting breast cancer?):