Teachers’ Unions II: Protecting Bad Teachers

One stock conservative talking point is teachers’ unions are a primary cause of educational woes.  If only unions could be eliminated or significantly changed, then education would improve significantly. Those defending unions argue that education would be worse without unions and some contend the effort to eliminate teachers’ unions is part of a plan to transform public education into a for profit-system to benefit a few well-connected elites.

Since the debate is so politically charged, it is difficult to objectively address the issue of whether teachers’ unions harm education or not. However, I will endeavor to address the matter as objectively as possible and acknowledge that as an educator and union member I am biased. As such, my arguments should be reviewed with due caution. Now, to the matter at hand.

One standard criticism of teachers’ unions is that they harm students by protecting bad teachers from being fired. If unions could be changed or eliminated, then bad teachers could be replaced with good teachers and the students would benefit. One variation of this criticism is focused on the practice of last-in first-out: those hired last are the first fired, should firings occurs. The concern is that teachers are retained based on seniority rather than ability, which can result in bad teachers remaining employed and good teachers being fired. Retaining bad teachers and getting rid of good teachers would clearly be bad for the students.

On the face of it, this criticism does match a plausible narrative about unions: since they exist to protect dues paying members, the leadership is not overly concerned about the quality of these members. As such, they do their best to see to it that no one is fired and thus bad teachers remain in the system. These bad teachers, obviously enough, do a bad job at teaching students and this harm can impact them throughout their entire life. Being able to fire these bad teachers would open positions for good teachers. The good teachers would do a good job, thus benefiting the students. From this it follows that eliminating unions would be good for students.

In the case of the policy of firing the last hired, the claim is that eliminating unions would result in merit based hiring and firing, so that when there was a need to fire teachers, the bad teachers would be eliminated regardless of seniority. As such, being rid of unions would improve things for students.

One easy and obvious reply to these criticisms is that they are not criticisms of unions as such. Rather, they are criticisms of specific practices: retaining bad teachers and retaining based on seniority rather than quality. There is nothing essential to a teacher’s union that requires that it mandate the retention of bad teachers nor that it mandate a seniority based retention system. To use an obvious analogy, there are countless examples of bad policies followed by corporations that do not arise simply because a corporation is a corporation. Roughly put, the bad policies are bad not because they are policies of corporations but because they are bad policies. As such, they do not provide grounds for the elimination of corporations. Rather, the badness of a corporation’s policy provides grounds for changing that policy. The same applies to teachers’ unions: the badness of a union policy serves as grounds for changing that policy, not elimination unions.

It could, of course, be argued that by their very nature unions must protect bad teachers and that it is impossible for them to do otherwise. Likewise, it could be argued that corporations by their very nature must have various terrible policies that harm the public. If so, then solving these problems would require eliminating unions and corporations. However, this view seems implausible; although people’s ideologies do often compel them to see things this way.

A second reply to these criticisms involves considering the facts of the matter. If unions protect bad teachers, then highly unionized districts should retain more bad teachers than districts that are less unionized. But, if unions do not protect bad teachers, then districts should have comparable percentages of bad teachers (adjusting for other factors, of course).

As should not be surprising, the debate over this factual matter tends to involves anecdotes about bad teachers and intuitions about unions. While anecdotes can provide some illustrative examples, they do not provide a foundation for general conclusions. There is, after all, the classic fallacy of anecdotal evidence which involves doing just that. Intuitions can provide some guidance, but by their very nature they are feelings and thoughts one has prior to considering the evidence. As such, anecdotes and intuitions do not suffice to show whether unions are good or bad in regards to the retention of bad teachers.

Fortunately, Professor Eunice Han has conducted a study of the claim that unions overprotect bad teachers. While it runs contrary to the anecdotes about bad teachers that cannot be fired and intuitions about overprotective unions, the evidence shows that “highly unionized districts actually fire more bad teachers.” Somewhat ironically, districts with weak or no unions retain more lower quality teachers than highly unionized districts.

As Han notes, stronger unions reduced the attrition rate of teachers and increase teacher wages. Because of the higher salaries, there is greater incentive to remove bad teachers and good teachers have a greater incentive to remain. This nicely fits the conservative mantra that top talent can only be kept by paying top salaries, although this mantra is usually just applied to people like CEOs and not workers.

In contrast, weak unions (and the absence of unions) increase the attrition rate of teachers and decrease teacher wages. As such, good teachers will tend to leave for areas with strong unions while bad teachers will often end up in areas with weaker unions or those that lack unions. The statistics show that unions have a positive impact on teacher quality and that the myths of the overprotective union and the irremovable bad teacher are just that, myths unsupported by facts. This also nicely matches the conservative mantra about compensation: lesser talent will settle for lower salaries.

It must be noted that since this issue is so ideologically charged, those who oppose unions will tend to regard the study as biased and might offer “alternative facts” of their own on the grounds that what they believe must be true. Likewise, those who favor unions can be accused of accepting “facts” that match their views. This is, of course, a much larger problem than the debate over unions: if there is not a shared set of facts and methods, then no rational discussion is possible. Only the howling of ideological stances driven by desire for profit and power.

My Amazon Author Page

My Paizo Page

My DriveThru RPG Page

Follow Me on Twitter

  1. Kevin Henderson

    Unions have an advantage now: social media and networking. Students can, overtime, undermine a teacher’s inabilities and put a teacher’s performance records on track to be lowest for pay raises, etc.

    Of course, students can lie, but integrated over time and averaging the rated reviews, teachers can have some accountability from students other teachers and possibly parents. Even if still employed, the teachers will have all respect dragged from them. Unabridged humiliation can be a motivation to improve.

  2. I love the phrase “alternative facts”, complete with scare quotes, as a sly synonym for a lie. I hope it becomes a regular part of discourse. I hope newspapers devote a small box on their political page to ‘”Alternative facts” of the day’. Even if the White House doesn’t reliably serve up a deserving entry, surely there must be one to be found daily somewhere in US politics.

    However, we should surely first agree on what “facts” are. For a start, vaguely defined generalisations from statistical inferences are not “facts”. “The height to the tip of the Eiffel Tower is 324m” is a fact, or at least a factual claim. So is “Sodium reacts exothermically with liquid water”, or “More people attended Obama’s 2009 inauguration than Trump’s”. These claims are specific and observable, and independent observers can measure and test such claims.

    However, in this case, we don’t even have an operational definition of “bad teacher” to start with. How can we claim either that “unions protect bad teachers” or “unionized districts dismiss more bad teachers” when we haven’t agreed what a bad teacher is? We live with these vague generalisations in our heads; I would hope that academics would turn their skills to clarifying them rather than engaging in more trench warfare in the fog.

    When we come to “teachers’ unions harm students by protecting bad teachers from being fired” or “districts with weak or no unions retain more lower quality teachers than highly unionized districts”, there are so many unsupported steps and unquantifiable assumptions that people could argue about them, in good faith, until union conditions and educational outcomes have changed enough to make the original statistics obsolete. Depending how you filter the statistics, you get different results, and you can then use your the ones that support your narrative. Dr Han herself says “This is where my study differs from some earlier ones that found that unionism either had no impact or had a negative effect on the dropout rate. I define unionism more broadly than those earlier studies. It’s not just collective bargaining that matters; it’s the union density of teachers in a district that’s important.” In this, it’s Dr Han’s specific definition of unionism that matters. Somebody else’s definition gives different results.

    My personal bias is that unions probably make little direct difference to teacher quality, and even less to pupil outcomes – maybe a little good here, and a little bad there. If they made a huge difference either way, that would be obvious to everyone, and we wouldn’t need duelling statistical studies. “If your experiment needs statistics, you ought to have done a better experiment.”

  3. Coffee Time Rebuttal

    While I agree in theory with some of the things the Coffee Time says, he needs to step outside and realize that life is not an experiment. Yes, experiments must be have strict controls and procedures to avoid contaminations and human error / bias. Unfortunately, not everything in this world can be studied using an experiment. There are so many things out there that are not ‘set’, including things like the definition of life and a planet – you can’t get everyone to agree on these because they are difficult questions.

    Specifically defining a bad teacher is similar. There are things that we would all agree are BAD (molesting students, teaching their own political views and not just the facts, etc), but there are other things (methodologies and level of excitement) that matter to some and don’t matter to others.

    The only way to get to the bottom of these are to have intelligent conversation with the best possible information (leaving out far left or far right attitudes and snarky comments is a big part of that ‘intelligent conversation’.) Just because you disagree with something doesn’t mean its wrong – that’s what opinions are. What might be right for my sons, may not work for other children – learning to learn is a lost art (just read comments on the internet).

    I would like to thank Mr. LaBossiere for this article. I am for the concept of unions, as they once saved this country. However, most unions have become corrupt because those in power just want to support their power base – just like lifetime politicians of any party. I have walked through factories with employees reading the paper at 10am – because they met their ‘quota’. I have a huge problem with this.

    This article pointed out some ideas to think about. He stated his possible bias up front (journalists take note), and then tried to remove that bias in his report – and THAT I appreciate.

  4. Thank you. I’m flattered to have a rebuttal!

    While Rutherford was talking about experimental science, his point could be applied to non-experimental studies as: “If your measured effect is so small and so uncertain that it is not easy to measure, and cannot be relied upon to appear every time, you should rethink your study.”

    If all unions always made a noticeable difference to the quality of education in a region, and in the same way, we would not be having this discussion, because it would be clear. However, we can’t possibly expect that different unions, with different penetration rates, with different policies, in different areas, serving different teacher populations, in different decades, under different laws and guidelines, in schools with different average attainments, would all have the same effect.

    I favour restricting the term “fake news” to stories entirely invented for clickbait, trolling, or political skirmishing. However, if we wanted to apply a wider definition, one of my first candidates would be the recounting of unfounded conclusions from single, untested studies. When I see a newspaper article begin with the words “A new study found …” or “Scientists today said …” I either stop reading or become fascinated by the inevitable trainwreck.

    Before commenting, I actually read the study paper.

    In the second sentence, I come across “Yet, researchers find that the quality of the teaching workforce in the US has been decreasing for the past several decades (Murnane et al., 1991; Lakdawalla, 2001; Bacolod, 2003; Corcoran, Evans, and Schwab, 2004; Hoxby and Leigh, 2004).” So now I’m interested: how do we determine the quality of teaching? I could not locate the first three studies on the net. The fourth discusses specifically the decrease in academic achievements of women entering teaching, as more academically achieving women choose higher-paying jobs. The fifth also discusses sex differentials, but also deals with teaching quality as a function of academic achievement and “the mean combined SAT scores of their college”.

    Already we are conflating the quality of teaching with the quality of academic attainment of teachers. This doesn’t pass the sniff test. While a teacher certainly needs to know the material, further knowledge is no guarantee of better teaching. We have all met the teacher or lecturer who understands the material inside out, but lacks the capacity to communicate it. A deep understanding of Hodge theory is no predictor of the ability to communicate basic algebra to 13-year-olds. In order to respect a link between academic attainment and teaching quality, one would first have to be demonstrated, and no-one has done that. I can not blame these researchers for not having a reliable metric for teaching quality, and Dr. Han herself regrets its lack, but without one, their results cannot say anything about teaching quality. All references to “bad teachers” are undefined.

    However, carrying on, we can instead consider the term Highly Qualified Teacher (HQT), used both in NCLB and Dr Han’s study. Dr Han summarises this as teachers who “1) have a bachelors’ degree; 2) hold full state certification or licensure, including alternative certification; and 3) demonstrate competency in the subject area they teach, such as passing a subject area test administered by the state.” I would have thought this would be the basic required qualification set of any teacher, but let us be clear that this HQT definition is at least sometimes being conflated with “good teacher” for the purpose of these studies.

    If I may further simplify, Dr Han extracts three conclusions from the data:
    1. Unions drive salaries up, and increase the discrepancy in salaries based on years of experience.
    2. Higher salaries attract and retain better people.
    3. Higher salary costs motivate districts to weed out “bad teachers” (we seem to be back to “bad teachers” without a definition again) in the first five years, before they get tenure.

    Reading through the study, many questions spring to mind, but two obvious questions about the title conclusion would be:
    1. How do we know that the weeded-out were “bad teachers”? What definition and test was used?
    2. A district that has higher salaries, and a greater disparity in pay based on length of service, would be motivated to put in place policies that would attract and use junior teachers (while they’re cheap) and dismiss them before they get both tenure and expensive. How do we distinguish between this possibility and the hypothesis that it is “bad teachers” that are being dismissed?

    I took a quick look to see if the data were available anywhere, so that I could see for myself whether these questions were susceptible to quick answers, but the dataset is not online. I might be able to get it from the author, but the process of giving out data in a usable form is not a trivial job, if it hasn’t already been archived in a suitable way. Even if Dr Han were willing to do this, I am not going to spend days or weeks reconstituting it.

    I am not implying any error or oversight, but all the paper tells me is that one author did an impressive amount of work collecting bodies of statistics, looked at them using one model, and published some personal conclusions – but without laying out the workings that would enable others to duplicate them, or considering other models.

    All of that is in line with normal academic practice, but to have the conclusions repeated unquestioningly by people who haven’t considered the limitations and alternatives is what causes the fog around the trench warfare. Next month, someone will lob a volley from some other trench with contradictory conclusions, and some other newspaper or blog will report its conclusions, and those who want to believe will continue believing whatever they want to believe.

Leave a Comment


NOTE - You can use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>