Steve Strickland
-
The Basics
-
Steve on Patch
-
More Stuff
Comments
-
On the article Taxes May Rise 5% in Holmdel, Sandy Cleanup Costs Millions
-
On the article Principal: Holmdel High Slighted in Top High Schools Ranking
Steve Strickland
11:20 am on Tuesday, April 30, 2013
With a reported 386 New Jersey schools in the study, the difference between the expected 31% and the actual 22% is about 35 schools that washed out in Step 1. So certainly Holmdel is not alone.BTW, the state-by-state statistics and the details of the methodology are documented in http://static.usnews.com/documents/best-highschools/Identifying_Top_Performing_High_Schools_April2013.pdf where you can see the variation among states and also how a few points in the score can make a critical difference in NJ.
-
On the article Principal: Holmdel High Slighted in Top High Schools Ranking
Steve Strickland
11:20 am on Tuesday, April 30, 2013
ReplyWith a reported 386 New Jersey schools in the study, the difference between the expected 31% and the actual 22% is about 35 schools that washed out in Step 1. So certainly Holmdel is not alone.BTW, the state-by-state statistics and the details of the methodology are documented in http://static.usnews.com/documents/best-highschools/Identifying_Top_Performing_High_Schools_April2013.pdf where you can see the variation among states and also how a few points in the score can make a critical difference in NJ.
-
On the article Principal: Holmdel High Slighted in Top High Schools Ranking
Steve Strickland
11:59 pm on Monday, April 29, 2013
ReplyIt does appear that a calculation error was made. I compute Holmdel's "State Test Performance Index" as 118.75, not the 116.9 reported by US News. This may well have caused Holmdel to miss the threshold in Step 1 of the "competition", dropping them from further consideration, regardless of any other criteria. But beyond this apparent error, this study has so many arbitrary choices and methodological flaws as to make the results largely meaningless. A different set of choices or a different methodology would product substantially different results. To begin with, their (arbitrary) threshold of one-half standard deviation (of the regression residuals) would normally be expected to pass about 31% of schools (assuming a normal, "bell-shaped" distribution). In New Jersey, only 22% of schools made this gate---the third lowest among the 50 states. This is almost certainly because each state was analyzed separately and its large number of "magnet" schools distorted New Jersey's distribution. Magnet schools should be in a different category because, by design, they are skimming the best students. I suspect that NJ doesn't do a worse job than Mississippi (31% of schools passing) or Alabama (27%). And the up-or-out system is like running the Olympic decathlon by dropping the bottom half of the athletes after the first event. I could go on (and on). Bottom line, this is shoddy science, and no basis for judging the school administration.
-
On the article Principal: Holmdel High Slighted in Top High Schools Ranking
Steve Strickland
11:59 pm on Monday, April 29, 2013
ReplyIt does appear that a calculation error was made. I compute Holmdel's "State Test Performance Index" as 118.75, not the 116.9 reported by US News. This may well have caused Holmdel to miss the threshold in Step 1 of the "competition", dropping them from further consideration, regardless of any other criteria. But beyond this apparent error, this study has so many arbitrary choices and methodological flaws as to make the results largely meaningless. A different set of choices or a different methodology would product substantially different results. To begin with, their (arbitrary) threshold of one-half standard deviation (of the regression residuals) would normally be expected to pass about 31% of schools (assuming a normal, "bell-shaped" distribution). In New Jersey, only 22% of schools made this gate---the third lowest among the 50 states. This is almost certainly because each state was analyzed separately and its large number of "magnet" schools distorted New Jersey's distribution. Magnet schools should be in a different category because, by design, they are skimming the best students. I suspect that NJ doesn't do a worse job than Mississippi (31% of schools passing) or Alabama (27%). And the up-or-out system is like running the Olympic decathlon by dropping the bottom half of the athletes after the first event. I could go on (and on). Bottom line, this is shoddy science and no basis for judging the school administration.
-
On the article Principal: Holmdel High Slighted in Top High Schools Ranking
Steve Strickland
11:57 pm on Monday, April 29, 2013
ReplyIt does appear that a calculation error was made. I compute Holmdel's "State Test Performance Index" as 118.75, not the 116.9 reported by US News. This may well have caused Holmdel to miss the threshold in Step 1 of the "competition", dropping them from further consideration, regardless of any other criteria. But beyond this apparent error, this study has so many arbitrary choices and methodological flaws as to make the results largely meaningless. A different set of choices or a different methodology would product substantially different results. To begin with, their (arbitrary) threshold of one-half standard deviation (of the regression residuals) would normally be expected to pass about 31% of schools (assuming a normal, "bell-shaped" distribution). In New Jersey, only 22% of schools made this gate---the third lowest among the 50 states. This is almost certainly because each state was analyzed separately and the large number of "magnet" schools included distorted New Jersey's distribution. Magnet schools should be in a different category because, by design, they are skimming the best students. I suspect that NJ doesn't do a worse job than Mississippi (31% of schools passing) or Alabama (21%). And the up-or-out system is like running the Olympic decathlon by dropping the bottom half of the athletes after the first event. I could go on (and on). Bottom line, this is shoddy science.
-
On the article How Did Holmdel and Hazlet Schools Fare on State Performance Results?
Steve Strickland
8:44 pm on Wednesday, April 24, 2013
ReplyA little independent analysis would provide context & insight, and (among other things) highlight the numbers that are meaningful and not those that are meaningless. For example, take the 6% number. Village School at the 6% level in college readiness? Shocking! Tax dollars wasted! Administrative failure! But spend a few minutes digging into the report and you find that this number is *completely* determined by the % of students absent for more than 10% of the days---regardless if the absence is excused. Something is likely wrong with this number (which is probably designed to capture truancy---not likely a problem in Holmdel), but regardless, a parent thinking of sending their child to Village school can ignore this 6% number as long as they make sure their kid actually goes to school.
-
On the article Principal: Holmdel High Slighted in Top High Schools Ranking
Steve Strickland
5:45 pm on Wednesday, April 24, 2013
ReplyLooks like Holmdel is being dragged down by its low "college readiness" score that seems to focus on absenteeism. I'm not sure how relevant this is if the academic scores are high. Presumably this is intended to capture truancy---not likely a problem in Holmdel.
-
On the article How Did Holmdel and Hazlet Schools Fare on State Performance Results?
Steve Strickland
5:30 pm on Wednesday, April 24, 2013
ReplyAh, I see what you're doing Patch. Publish obviously misleading numbers with no analysis in order to attract the eyeballs of outraged parents wondering why their school is at the 6% level. Well played!
-
On the article Budgets Not Easily Available on Holmdel and Hazlet Websites
Steve Strickland
12:55 pm on Sunday, April 7, 2013
"As I recall, the 2012 budget was balanced without selling open space, while the property tax increase was under the state mandated 2% cap..."
The actual increase (if you look at your bill) was a bit over 12%. It is only after some budgetary/accounting maneuvers leading to a page of exclusions---partly documented on Sheet 3b(la)---that the 2.5% cap was met. But by all means, don't believe me; take a look at the numbers on your own bill. Make sure you include the "County Health Tax" that represents the Health Dept that was outsourced in 2012.
Steve Strickland
3:49 pm on Wednesday, May 8, 2013
As I have repeatedly pointed out, last year's increase was actually 12.5% since you need to include the new "County Health Tax" that represents the outsourcing of the Health Dept. to the county. Outsourcing only lowers costs if what you have to pay is less than what it was costing to do it yourself. By my calculation, we are actually paying more. Regardless, you can still look at the total and compare with the previous year. And our total went up 12.5%. Don't believe me? Look at your own tax bill.