NM Leon
New member
Quote:No, we just have to know things that actually matter. Stats are so misleading that you can "prove" anything.
It's truly amazing that in an industry that's driven by statistics, so many of it's practitioners are so totally ignorant of the science of statistics. It's also illuminating that the folks that claim statistics can be manipulated to show anything are the ones that know absolutely nothing about statistical science.
Numbers are often mislabeled as statistics and used to mislead the gullible and to try to "prove anything".
Statistics are not misleading at all. Statistics are what numbers turn into after having been processed through the rigorous filter of statistical science.
Quote:I've given you stats - you shot them down and said they were invalid or something to that effect.
You've given us numbers that had in no way been processed through statistical mathematical testing. Your "numbers" were #1 incomplete (used 5 instead of 15 states), #2 used invalid sample rates (comparing the top 4% of one state to the top 70% of another state), #3 measured test scores of only college bound students, and #4 then claimed to show causation between the statistically invalid test scores of college bound seniors and the false set (5) of states, to the quality of public education in collective bargaining and non-collective bargaining states.
I've only been able to "shoot down" your "stats" because they weren't "stats" at all, they were just numbers designed to fool the gullible.
You (or anyone else) haven't been able to "shoot down" my stats because the stats I quote come from sources that rigorously use statistical science on the raw data (numbers) in order to come up with those "stats". If you followed the links you would find that my source is in fact usually the U.S. Department of Education Institute of Education Sciences' National Center For Education Statistics, the Programme for International Student Assessment (PISA), the U.S. Census, etc. If you follow the links you will see that in each case they provide not only the analysis, but the raw data, methodologies, and controlling factors, that they used to reach their conclusions. In other words they basically invite anybody to check their work.
Using statistics from unimpeachable sources who invite anybody to check their results pretty much guarantees that my statistics can't be shot down.
That's as opposed to your numbers where the sources of your data say that the analysis you cited is invalid.
Here's one for you jeffo, in examining the NAEP scores for 8th graders nationwide, the average scores for math/science/reading/writing in the 15 states that don't have union collective bargaining do better than the 35 states that do have collective bargaining.
Does that mean that those numbers show that curtailing collective bargaining will improve your kids scores? Absolutely NOT. They show absolutely NOTHING about collective bargaining one way or another because #1 averaging those raw scores is not a statistically valid methodology, #2 there is no causal link shown between collective bargaining and quality.
Until that raw data is assessed using scientifically valid mathematical formulae and methodology, it's just numbers, not statistics.
Quote:There are LOTS of very good teachers in the public schools who bust their asses every day for their students.
Absolutely true, but if you take the worlds best auto workers and put them on the assembly line of a Yugo factory, the result will still be a poorly designed car.
Would you advocate forcing everybody to buy that car?
Would you advocate using taxpayer money to keep that factory open?
How would you feel about the union who had enormous influence in ensuring that everyone had to buy their car, that the car's design worsened instead of improving, and that the factory remained open? Would you be alright with the union and it's influence because it represented those excellent auto workers?
Of course the reality is that some of the workers will be excellent, some average, and some poor quality, and that unionization generally produces a higher cost and/or a lower quality product (and yes I can provide valid statistics to show that).
It's truly amazing that in an industry that's driven by statistics, so many of it's practitioners are so totally ignorant of the science of statistics. It's also illuminating that the folks that claim statistics can be manipulated to show anything are the ones that know absolutely nothing about statistical science.
Numbers are often mislabeled as statistics and used to mislead the gullible and to try to "prove anything".
Statistics are not misleading at all. Statistics are what numbers turn into after having been processed through the rigorous filter of statistical science.
Quote:I've given you stats - you shot them down and said they were invalid or something to that effect.
You've given us numbers that had in no way been processed through statistical mathematical testing. Your "numbers" were #1 incomplete (used 5 instead of 15 states), #2 used invalid sample rates (comparing the top 4% of one state to the top 70% of another state), #3 measured test scores of only college bound students, and #4 then claimed to show causation between the statistically invalid test scores of college bound seniors and the false set (5) of states, to the quality of public education in collective bargaining and non-collective bargaining states.
I've only been able to "shoot down" your "stats" because they weren't "stats" at all, they were just numbers designed to fool the gullible.
You (or anyone else) haven't been able to "shoot down" my stats because the stats I quote come from sources that rigorously use statistical science on the raw data (numbers) in order to come up with those "stats". If you followed the links you would find that my source is in fact usually the U.S. Department of Education Institute of Education Sciences' National Center For Education Statistics, the Programme for International Student Assessment (PISA), the U.S. Census, etc. If you follow the links you will see that in each case they provide not only the analysis, but the raw data, methodologies, and controlling factors, that they used to reach their conclusions. In other words they basically invite anybody to check their work.
Using statistics from unimpeachable sources who invite anybody to check their results pretty much guarantees that my statistics can't be shot down.
That's as opposed to your numbers where the sources of your data say that the analysis you cited is invalid.
Here's one for you jeffo, in examining the NAEP scores for 8th graders nationwide, the average scores for math/science/reading/writing in the 15 states that don't have union collective bargaining do better than the 35 states that do have collective bargaining.
Does that mean that those numbers show that curtailing collective bargaining will improve your kids scores? Absolutely NOT. They show absolutely NOTHING about collective bargaining one way or another because #1 averaging those raw scores is not a statistically valid methodology, #2 there is no causal link shown between collective bargaining and quality.
Until that raw data is assessed using scientifically valid mathematical formulae and methodology, it's just numbers, not statistics.
Quote:There are LOTS of very good teachers in the public schools who bust their asses every day for their students.
Absolutely true, but if you take the worlds best auto workers and put them on the assembly line of a Yugo factory, the result will still be a poorly designed car.
Would you advocate forcing everybody to buy that car?
Would you advocate using taxpayer money to keep that factory open?
How would you feel about the union who had enormous influence in ensuring that everyone had to buy their car, that the car's design worsened instead of improving, and that the factory remained open? Would you be alright with the union and it's influence because it represented those excellent auto workers?
Of course the reality is that some of the workers will be excellent, some average, and some poor quality, and that unionization generally produces a higher cost and/or a lower quality product (and yes I can provide valid statistics to show that).