In Michigan, all conventional public school teachers must be certified, and nearly 99 percent of certified teachers in the state earn their certification through traditional means. The No Child Left Behind Act also requires teachers to have state certification in order to earn the "highly qualified" teacher designation. Given that all teachers must meet this NCLB requirement, it is important to explore what is involved in earning certification in Michigan, to examine the rationale behind this quality-control mechanism, to determine whether certification affects student achievement and ultimately to explore whether there are ways to improve teacher quality by reforming the certification process.
Having a command of the subject and some experience as an instructor would not qualify a teacher to be certified in Michigan. Rather, a teacher must have at least a bachelor’s degree from an accredited college or university and must complete coursework at an approved teacher preparation program.[*] The hours of university coursework required to earn an endorsement in a given subject or grade level vary. According to the Michigan Teacher Certification Code, "‘Certificate endorsement’ means subject or subjects that a teacher is authorized to teach at specific grade levels based on completion of appropriate coursework and passage of the appropriate state teacher subject area examination." To be certified, all teachers must pass at least two of the Michigan Test for Teacher Certification exams.[†]
In Michigan, a teaching license is granted at completion of the certification process and is valid for five years. The terms "teaching license" and "teaching certificate" are commonly used interchangeably. According to the Michigan Department of Education: "The renewal of a Professional Education certificate requires the completion of 6 semester hours of credit (these credits may be completed at a 2-year or 4-year institution) or 18 State Board-Continuing Education Units (SB-CEUs) or a combination of the two. Three SB-CEUs equals one semester hour. The required credit hours or SB-CEUs must be completed after the issuance of the Professional Education certificate and within five years of the date of application for the renewal." For teachers who were trained or provisionally certified outside of Michigan, the procedures for earning or renewing a teaching certificate are also rather complicated.
The requirements for initial licensure vary somewhat based on the teacher preparation program. The state sets minimum standards for hours of coursework that constitute a major and minor, 30 and 20 hours respectively, but individual teacher preparation programs can choose to require additional training. To earn an elementary provisional teaching certificate, a teacher must complete no fewer than six semester credit hours in teaching reading; for the secondary provisional certificate, teachers must take at least three credit hours in teaching reading. Aside from passing the testing requirements outlined in the next section, a final requirement for certification is that a teacher must participate in student teaching that is coordinated and supervised through an approved program.
At the time of this writing, Michigan has 31 approved teacher preparation programs. Teacher preparation programs are approved by the State Board of Education when they meet the program standards articulated by the Michigan Department of Education and pass the review that is part of approval process. Preparation programs must also pass annual quality reviews by the MDE.
Pursuant to Title II of the U.S. Higher Education Act, the MDE recently rated these programs. The MDE uses a 70-point scale: 40 points are earned through participant completion rates, the degree to which programs prepare teachers to fill high-demand slots, surveys of teacher candidates regarding their programs and the percentage of students taking the licensure exams who belong to minority groups. The final 30 points are based on a three-year aggregate of the specialty content area licensing exams.[‡] The overall ratings are "exemplary," "satisfactory," "at-risk" or "low performing."
In the 2007 report, which reported program performance in the 2005-2006 academic year, 18 programs were rated "exemplary." The two highest-scoring programs were at Oakland University and Hope College; each earned 68 total points. Andrews University, Eastern Michigan University, Grand Valley State University, Michigan State University and the University of Michigan-Dearborn tied for third, with a rating of 66 points. Ten programs earned a "satisfactory" rating, and one — Olivet — was rated "at-risk." Finally, two of Michigan’s teacher preparation programs — Adrian College and Marygrove College — were deemed "low performing," the lowest possible category. According to the MDE: "Institutions identified as low performing have two years to improve their performance before state sanctions occur. Institutions identified as at-risk must progress to the satisfactory category within two years or move to the low performing category, even if their raw score is still in the at-risk level."[**]
The stated purpose of requiring all teachers to become certified is to establish minimum standards for teacher quality, so qualified teacher training experts have vouched that a new teacher has the requisite skills to be successful before that teacher takes charge of a classroom. Thus, teacher certification is theoretically akin to professional certification for doctors, lawyers and engineers.
Certification in these other fields, however, usually involves a much more rigorous screening and training process. As a result, the achievements of personnel who actually earn their certification in these other fields tend to be higher. If policymakers were to raise teacher certification requirements to sift out low-quality teaching candidates, they would need to change other features of the labor market to avoid negative unintended consequences. For example, simply raising the certification requirements could exacerbate shortages in high-needs areas, such as special education and secondary and middle school mathematics and science, since some candidates might not be willing (or able) to satisfy the expanded requirements. Moreover, since raising certification standards could require considerable legislative action, it is important to examine whether the quality of education is helped by extensive certification requirements.
Stanford’s Linda Darling-Hammond, perhaps the most prominent advocate of teacher certification, released a study in 2005 with her colleagues in which she explored whether having traditional certification makes teachers more effective. Darling-Hammond et al. analyzed fourth- and fifth-grade student achievement data from the Houston public schools from 1995 through 2002. They found that teachers with traditional certification were generally more effective at producing student achievement gains than those who were teaching without certification or who had become certified through alternative means. This finding included teachers from Teach for America, a program that offers limited teacher training to uncertified, academically able recent college graduates before placing them as teachers for two years in economically disadvantaged schools. Other studies of Teach for America are discussed at length below.
In reporting her results, Darling-Hammond cited numerous studies with concurring opinions. In a more recent study of North Carolina elementary school student achievement in 2007,[††] Goldhaber finds, "[S]tudents of teachers who graduate from a North Carolina-approved training program outperform those whose teachers do not (that is, those who get a degree from an alternative state program or a program from outside the state) by about 1 percent of a standard deviation. ..." Although this finding was statistically significant, it is difficult to argue that such a small difference has practical significance for policymakers.
The other side of this dispute is represented by equally prominent researchers — for example, Eric Hanushek — whose analyses show that certified teachers perform no better or worse than their uncertified colleagues. Opponents of certification argue that certification erects unnecessary barriers to entry into the teaching profession, especially for career-changers.
Considerable evidence exists to support their claims. For example, in a recent, detailed study of New York City public school teachers, Thomas Kane of Harvard University, Jonah Rockoff of Columbia University and Douglas Staiger of Dartmouth College explored the relative effectiveness of teachers with traditional certification, alternative certification and no certification in New York City public schools. As the authors explained, New York City schools are particularly interesting for the study of teacher certification because large numbers of the city’s teachers fall into each certification category. Of the more than 50,000 new teachers hired in the district from 1999 through 2005, 46 percent were certified; 34 percent were uncertified; and 20 percent were alternatively certified. Most of the alternatively certified teachers participate in the New York City Teaching Fellows program, in which they are given provisional certificates and intensive preservice teacher training. They also enroll in a master’s degree program that will allow them to earn full certification upon completion of three successful years of teaching in the district. Uncertified Teach for America teachers are also well represented, although they form a small percentage of New York’s total new teacher work force.
In examining the math and reading learning gains of New York City students in grades three through eight, Kane and colleagues found that the differing teaching credentials produced minimal differences in student performance. The researchers found: "On average, the students assigned to [the alternatively certified] teaching fellows performed similarly to students assigned to certified teachers in math, and slightly lower (-.01 standard deviations) in reading. ... We find evidence that Teach for America corps members have slightly higher value-added (.02 standard deviations) for math test scores than traditionally certified teachers, but we find no difference in reading."
Kane, Rockoff and Staiger’s findings are similar to those of the Mathematica Policy Research’s study of Teach for America teachers. In this study, Mathematica used a random assignment experimental research design to measure the effectiveness of TFA participants compared to teachers with traditional credentials. Randomized design is the gold standard of social science research, and the findings of such experiments are usually given extra weight.[‡‡] 
In the Mathematica study, students in 17 schools representing six geographically diverse regions of the country were randomly assigned to certified teachers or uncertified TFA teachers. The researchers found that students of TFA teachers demonstrated more growth in math than those of their certified peers, while students of TFA teachers did not routinely score any differently from other students in reading. As the authors of this study point out, their findings help to settle the dispute that has resulted from the mixed findings on the effectiveness of TFA in quasi-experimental research designs.[***]
Two other reforms related to raising certification requirements could also be considered. First, compensation levels and methods could be enhanced to motivate potential teachers to accept the opportunity costs involved in earning a more rigorous certification. Alternatively, since research suggests that traditional certification is not a guarantee of teacher quality, policymakers could remove some certification requirements that may act more as a deterrent to potential teachers than as a guarantee of quality. University coursework in pedagogy, for instance, has not been shown to improve teacher performance. By making such coursework elective, rather than required, Michigan might be able to attract new teachers who would be at least as effective as those currently in the profession. [†††]
[*] Information about programs at approved teacher preparation institutions can be found at the Michigan Department of Education Office of Professional Preparation Services Web page on "Approved Teacher Preparation Programs," https://mdoe.state.mi.us/proprep/.
[‡] While federal law mandates that all states rate their teacher preparation programs annually, each state has discretion over the rating criteria it will use.
[**] According to the most recent data in a U.S. Secretary of Education report, 17 programs nationally fell into the remediation categories of "low performing" or "at-risk" in 2005. These national data on preparation programs show that 11 states had failing programs. Three states, Illinois, Kansas and South Carolina, each had three programs on the list. No Michigan colleges of education were rated in either category at that time (2002-2005), though the three listed above fall into those categories now. It is difficult to assess whether Michigan’s relatively high rate of lower-performing teacher preparation programs is the result of truly underperforming institutions or of a tougher evaluation system. (See "The Secretary’s Fifth Annual Report on Teacher Quality: A Highly Qualified Teacher in Every School Classroom," 141.) The same federal report also stated that at least 90 percent of Michigan’s public school teachers graduated from Michigan colleges of education.
[‡‡] Randomized experimental designs, also known as randomized controlled trials or randomized field trials, are experiments in which participants subjected to a given treatment are selected for that treatment essentially by a flip of the coin. In these experiments, those who are not selected for the treatment are placed into a control group for the sake of comparison. Because participants in the experiment are placed in the treatment and control groups simply by chance, this randomization distributes variation in study participants’ characteristics equally between the two groups. As a result, researchers can be confident that any differences in the groups after the treatment occurs are due to the treatment and not to any pre-existing differences in the two groups. The primary alternative social science methodology to random assignment studies is called "quasi-experimental research," which involves significant statistical controls. This quasi-experimental research predominates in education studies for many reasons, while random assignment experimental research is exceedingly rare. (See Thomas D. Cook, "Considering the Major Arguments Against Random Assignment: An Analysis of the Intellectual Culture Surrounding Evaluation in American Schools of Education," Harvard Faculty Seminar on Experiments in Education (Cambridge, Mass.: 1999).)
[***] Summarizing the mixed research literature on Teach for America instructors prior to the Mathematica study, Decker, Mayer and Glazerman wrote: "Despite TFA’s rapid expansion, there is little evidence whether teachers with strong academic backgrounds, but limited exposure to teaching practice, can be effective. Some critics argue that programs such as TFA are ‘loopholes’ that permit unlicensed and under-trained teachers into the classroom simply as a way to address teacher shortages. Darling-Hammond (1994, 1996) has argued that TFA teachers ‘often have difficulty with curriculum development, pedagogical content knowledge, students’ different learning styles, classroom management, and student motivation.’ Other researchers are more optimistic about the potential benefits of hiring teachers through programs such as TFA. Ballou and Podgursky (1998) argue that there is no evidence that formal teacher certification produces more qualified teachers and that certification policies may discourage talented individuals from entering the profession. Two recent studies (Raymond et al. 2001; and Laczko-Kerr and Berliner 2002) attempted to assess the impact of TFA using nonexperimental methods on samples drawn from single regions, and generated mixed findings regarding the effectiveness of TFA teachers." (See Decker, Mayer and Glazerman, "The Effects of Teach for America on Students." See also Raymond, Fletcher and Luque, "Teach for America: An Evaluation of Teacher Differences and Student Outcomes in Houston, Texas" and Ildiko Laczko-Kerr and David C. Berliner, "The Effectiveness of ‘Teach for America’ and Other Under-certified Teachers on Student Academic Achievement: A Case of Harmful Public Policy," Education Policy Analysis Archives, 10, no. 37 (2002).)
Note that the study by Darling-Hammond et al. of the Houston public schools used a quasi-experimental design. In that report, Darling-Hammond criticized this Mathematica study, arguing that it did not choose an appropriate control group. Even bearing her criticism in mind, the Mathematica study is a particularly strong data point questioning the contention that traditional teacher certification is necessary for student success.
One other objection might be raised against the studies cited in this section: They do not involve certified teachers in Michigan. Since states have different certification requirements, it might be argued that Michigan has a superior certification process that would in fact be linked to teacher effectiveness in an unbiased study. Such an outcome is unlikely, however. The studies cited in this section are among the highest-quality in the field, and there is no reason to believe that Michigan’s teaching corps and teacher preparation programs are dramatically different from the other states studied.
[†††] As Goldhaber has written: “If the required credentials are only weakly correlated with student achievement, it will result in significant numbers of ‘false positives’ and ‘false negatives’ — that is, many applicants who satisfy the criteria for employment eligibility turn out to be ineffective teachers (false positives), and many who do not satisfy the criteria but who would have been effective in the classroom had they been allowed into the teacher workforce (false negatives). The false negatives may never persevere to become teachers — a loss to the profession — and the false positives may be difficult to remove from the classroom once they have attained the job security, via tenure, which typically exists in public schools.” See Goldhaber, “Teacher Pay Reforms,” 6.