Response to the Review of French Second Language Programs and Services in NB



Dr. Diana J. Hamilton, assistant professor of biology at Mount Allison University in Sackville NB. She teaches advanced statistics to biology students, in addition to other courses.

Dr. Matthew K. Litvak, professor of biology at the University of New Brunswick, Saint John. He has taught advanced statistics to biology students, in addition to other courses.


This document has been reviewed by:

Dr. Jennifer Baltzer, Ms. Wendy Burnett, Dr. Amanda Cockshutt, Ms. Bernadine Conron, and Mr. Andrew Spring


The Croll and Lee FSL report has recommended the total elimination of Early French Immersion in New Brunswick schools. This report has stimulated an intense and emotional debate in the public. When we initially read their recommendations, we were surprised and curious as to why their results and conclusions were so different from those of previous reports prepared by education experts in NB and elsewhere in Canada. Consequently, we decided to read the report carefully.  In the process of doing this, we found that it is a deeply flawed document which fails to provide valid evidence in support of the sweeping policy changes that its authors propose. We also found the biased tone used, in what is presumably an objective report, to be troubling. Here, we systematically point out errors in both analysis and interpretation by copying sections of the report into the document and commenting on them. When possible, we provide alternative analyses to document our findings. We restrict most of our comments to quantitative aspects of the report, though at the end we also provide general comments and suggestions based on our analysis. We hope that the NB government will read this document and seriously consider the points we raise. Based on much of the public reaction to this document, we are very concerned that most people are simply accepting the conclusions of the report at face value, and that those who question it are criticized as having an agenda or belonging to a special interest group. The flaws in this report are significant and are of particular concern if the government is considering acting on its main recommendations.



We have copied relevant sections and page numbers from the FSL report and pasted them below. Comments on each of these sections follow directly after the pasted segments.


Review of Croll and Lee FSL report March 2008.pdf


To comment on this report please link to our blog

Executive Summary


Ø      The authors suggest that there is a downward trend in students in the Core program meeting intermediate proficiency, but appropriate statistical analysis shows no such trend. Numbers are not good, but they are not declining as indicated.


Ø      All attrition rates in the document are calculated incorrectly and are highly misleading. The authors contrasted registrations across grades within single years rather than track a cohort of students through each year. They concluded that attrition was much higher in EFI than LFI. When we correctly calculated rates in high school for both programs (by following cohorts of children through as many years as possible), we found absolutely no difference in attrition rates. The method they used was a gross misrepresentation of the data – claiming there is a huge difference when in fact there is none. Clearly one of the main reasons cited to dismantle EFI in favour of LFI is based on incorrect and inaccurate analysis.


Ø      The authors used oral proficiency assessment results to establish the number of students succeeding in the program, although this test is optional and not all students choose to take it. Many students may drop out in grade 10 or 11 to prepare for English university. Assuming that none of them have achieved the expected proficiency levels is not justified and produces a downward bias in results. The report chooses achievement of a certificate as the sole measure of program success rather than determining how much French the children who went through all or most of it actually learned.


Ø      We used the authors’ data to calculate how many students achieved the various levels of proficiency in both the EFI and LFI programs. We found that even though fewer students were tested in EFI than in LFI, more EFI students attained intermediate plus and advanced proficiency standards. Given the expectation that many who were not tested would also have met these standards, this difference may be even larger than measured.


Ø      Based on data presented in the report, we calculated how many children in total meet particular proficiency levels (i.e., combine LFI and EFI numbers), and compare that to what would have happened if all these children had gone thorough the LFI program (which is what is being proposed). We found that there was no significant difference between the current program and the all LFI approach up to intermediate proficiency. However, the number of children attaining intermediate plus and advanced levels of proficiency would drop significantly under the LFI only scheme. If, as they propose, we switch to an “LFI only model” 270 tested students will do worse than the current model each year. Given that more enroll in the program than are tested, that number is likely a substantial underestimate of the number of students who will be detrimentally affected by this change.


Ø      The authors suggest that eliminating EFI would dramatically reduce the number of exceptional students in each class. Based on the total number of exceptional students in grade 2 in 2005 (one example provided in the document), and estimating an average class size of 23, elimination of EFI would reduce the number of exceptional students in core classroom from a province-wide average of 5.4 to 4.25 students. Not the massive reduction that was implied in the report.


Ø      On many standardized tests, core students are doing as well or nearly as well as EFI students. This suggests that streaming may not be as big a problem as it is purported to be.


Ø      The authors suggest LFI is more cost-effective at meeting goals. However, based on stated percentages of students in each program and percentage of funds allocated to each program, EFI is 30% less expensive per student than LFI.  EFI is the better financial choice if the goal is to produce French-speaking graduates.


To summarize, EFI produces better French speakers, costs less on a per-student basis, and has essentially the same attrition rate as LFI. The logical choice is to retain EFI. Core French certainly needs to be fixed, but we have found no justification in this document for eliminating EFI in the process. We strongly suggest that the central recommendation of this report not be adopted; it will lead to a reduction in French competence of hundreds of graduates per year, and result in a general lowering of standards. Numerous education experts have stated that EFI is the best program, and based on our analysis, we fully agree.


We feel that many of the legitimate shortcomings in New Brunswick FSL programs that the Commissioners have identified can be addressed more effectively as follows:


Ø      by providing adequate resources to support a wider range of French-language course options and flexibility in Grades 10-12 in order to reduce early drop-out;

Ø      by providing adequate support for exceptional children in EFI and LFI classes so that FSL training becomes available to these children;

Ø      by actively promoting the benefits of learning French and encouraging all students equally.

Link to our update page posted March 15th, 2008.

Link to our Review of Croll and Lee FSL report March 2008.pdf

To comment on this report please link to our blog