A State of Suspended Belief

This letter to the editor was submitted by Louisiana teacher Herb Bassett who has the mathematical ability and determination to actually check some of the preposterous assertions of Superintendent John White.  White's latest public gaff was a claim that the teacher attrition rate in Louisiana is status quo and furthermore that those who are leaving the profession are the LEAST EFFECTIVE TEACHERS.  BELIEF (New World Dictionary) unquestioning acceptance of something even in the absence of proof.


LOUISIANA BELIEVES - I VERIFY 

In his recent press release, superintendent John White claimed that teachers are not leaving the classroom in greater numbers than before, and he used flawed data to overstate the case that more "ineffective" teachers than "highly effective" teachers left.  Here are his own numbers without his bizarre spin: 

"Ineffective" teachers are by definition (White's)  the lowest ranked 10% of the teaching force.  12% of teachers who left the classroom were "ineffective."  "Highly Effective" teachers are the top 20%.  According to White, 16% of teachers who left the classroom were "highly effective." 

The other 72% of teachers leaving the classroom were classified as "Effective."  So his own numbers show that 88% of teachers who left the classroom were effective teachers.  

White's conclusion?  "The teachers who are leaving are more likely to be ineffective." 

That statement is unjustified and an insult to my colleagues -- all dedicated teachers -- who retired this year. 

Furthermore, his data set is flawed; it lost track of 300-400 "highly effective" teachers over three years. White claims that 19% of the teachers who stayed and 16% who left were "highly effective."  This is mathematically impossible since LDOE set a quota of 20% for "highly effective" teachers. 

Based on the LDOE data and its quota, there were 7725 "highly effective" teachers. Using White's figures and accounting for the maximum possible rounding error, a (maximum) total of 7419 "highly effective" teachers either left the classroom or stayed. 

What happened to the missing 306 "highly effective" teachers? They were lost in data error. 

They probably were mis-categorized.  This kind of error could affect the careers of the very teachers the new evaluation system is supposed to identify and reward.  Furthermore, this skews the data;  the percentage of "highly effective" teachers that left is likely 2% higher than White stated. This weakens the his assertion even further. 

This data was likely hurriedly compiled because LDOE recently redefined "highly effective."  (Teachers have become frustrated with the continual changes to the new evaluation system which was implemented incomplete and untested.)  In his haste to create spin, John White sacrificed the truth of accuracy in order to muddy the waters. 

How many teachers have left this year?  White's data sample shows that 1204 left last school year and that 1157 have already left this year, which is barely half over.   

When the number doubles to over 2000 by the school year's end, I will call that a mass exodus.  

Is it  a wonder why? 

Herb Bassett, Grayson, LA

 

Further Explanation and Documentation
 

In his LDOE website press release, White tortured some data to overstate the case that more "ineffective" teachers than "highly effective" teachers left.  Here are his own numbers in table form without his bizarre presentation:

 

Classification              Percent of Teachers                 Percent of Teachers who left (Overall)

Ineffective                              10                                            12

Effective: Emerging                40                                            42

Effective: Proficient               30                                            30

Highly Effective                     20                                            16

 

These numbers fail to justify White's insult to the dignity of the 88% of teachers leaving the profession who were effective. 

* * * * * 

White claims that 19% of the teachers who stayed were "highly effective;" only 16% of the teachers who left were "highly effective."  This is mathematically impossible since LDOE set a quota of 20% for "highly effective" teachers.
 
Let's account for all those "highly effective" teachers using his (three-year) "Overall" data (the year by year data works out similarly).  To be completely accurate in disclosure, since this is three-year data, the figure 34890 teachers who stayed is technically 34890 teacher-years. Many teachers in the 34890 figure are counted three times for the three different years in the data. Still its use in the computation is correct and appropriate.  It is the same method used by LDOE in White's interpretation of the data. 

3734 teachers left, of which 16% were highly effective.  .16499 is the highest number that rounds to 16%; we will use it to compute the number of "highly effective" teachers who left (results rounded to the nearest whole teacher).
 

3734 x .16499 = 616 

34890 teachers stayed, 19% were "highly effective." Again we use the highest number that rounds to 19%. 

34890 x .19499 = 6803
 

We have accounted for a maximum of 616 + 6803 = 7419 "highly effective" teachers. 

There were 3734 + 34890 = 38624 teachers total in the data.  By LDOE's definition, exactly 20.0% of those were "highly effective." 

38624 x .20 = 7725
 

Over the three years, there were 7725 "highly qualified" teachers of which a 7419 either left or stayed. 

What happened to the other 306? They were lost in data error by LDOE.   

* * * * * 

How do we establish the assertion that the percent of "highly effective" teachers leaving the classroom was likely 2% higher than the 16%? 

Repeating the calculations above with .16000 and .19000 instead of the higher figures, we find that instead of 306 (with is the minimum possible number of teachers lost in the data), 498 are lost in the data.  This figure would be in the middle instead of the minimum or maximum. 

Using the figure of 12% of teachers leaving the workforce, we compute the increased number of "highly effective teachers leaving. 

498 x .12 = 59.76 

Now 59.76 is 1.6% of the total number of teachers leaving. 

59.76 / 3724 = .016004 or 1.6004%
 

That rounds to 2%, and therefore is likely to increase the 16% to 18%. 

Since LDOE made that miscategorization error, that 2% had to come from one of the other categories, most likely from Effective: Proficient, so his data table probably should have read:

 

Classification              Percent of Teachers                 Percent of Teachers who left (Overall)

Ineffective                              10                                            12

Effective: Emerging                40                                            42

Effective: Proficient               30                                            28

Highly Effective                     20                                            18

 

White's assertion that a higher proportion of "ineffective" teachers left than the proportion of "highly effective" teachers is supported by the data to some degree.  However, common sense says that poeple who struggle at something are going to seek something else to do, and those who are good at what they do are likely to be comfortable and continue doing it.  While this might show up in large group data, I do not concede that VAM is an accurate tool for measuring individual teachers.  His assertion that  "The teachers who are leaving are more likely to be ineffective" is misleading because when other categories of "effective" teachers are considered, it is very weakly supported in that context.  

* * * * * 

Finally, the LDOE data states that 8% of of teachers staying and 12% of teachers leaving are "ineffective".  My mental arithmetic says this is not within rounding error of the 10% quota for that category, but I have not explored that problem with the data.

No comments:

Post a Comment

Your Comments Welcome