1. 21st century inequality – the declining significance of discrimination [Roland Fryer]
“TODAY I want to talk about inequality in the 21st century, in particular on the decline in the significance of discrimination and the increase in the significance of human capital.”
Note: so much to consider here; much I agree with, much I’m skeptical about. But worth reading!
2. Arkansas should drop PARCC for a NRT
“Norm-referenced tests are not aligned with any particular set of standards, but can still provide general measures of how well students are performing academically. They meet our reasonable goal of wanting transparency about how students are progressing in school. But because they are based on a generic curriculum rather than a particular set of standards, it really isn’t possible for schools to game them by focusing exclusively on a narrow set of content.”
Note: in my post, the Voucher State, I noted that NRT would likely be an option.
3. Recovery School District expulsions down
“Recovery School District charters in New Orleans are on track to cut expulsions by one third in the 2014-15 academic year, and officials say it’s due to a new policy that lets administrators counsel students instead of kicking them out.”
4. National high school graduation up 1 percentage point
“The nation’s high school graduation rate ticked up for the second year in a row, according to new federal data released Thursday, which shows that 81 percent of the Class of 2013 graduated within four years. That’s an increase of one percentage point since 2012 and two percentage points since 2011, making it the highest rate since states began calculating them in a uniform way in 2010.”
5. Tech starts up raise more money in USA than Europe
“European start-ups raised about $7.6 billion last year, a 41 percent leap over 2013, according to the data provider Dow Jones VentureSource. But that was only about one-fifth the amount raised by American technology companies, which secured a combined $37.9 billion in 2014, up more than 30 percent from the previous year.”
Note: the article did not answer the question “Why?”
6.
“A big question for MOOCs, the free online courses that hundreds of colleges now offer, is whether employers will take them seriously as credentials. But some of the biggest MOOC producers may have figured out how to jump-start employer buy-in: Get big-name companies to help design them.”
“Listen — you are going to go through hell. Your mind is going to be loud. You’ll scream into a towel and talk to the walls and maybe even smash a glass or two, kick a door, or worse. You’ll see couples on Facebook congratulating each other on 10, 20, 30 years together and you’ll hate them. Sometimes you’ll feel as if you deserve this, to be all alone, away from your family. Then one day you’ll come out the other side.”
Neerav, NRTs are not good. How much do you know about how the “norm” is set?
Sent from my iPad
LikeLike
I know very little. Not advocating for them specifically here – but saying that there is a tension between autonomy and testing, and if we want to mover further down the autonomy axis, we’ll likely open up testing from singular state tests to models where school can choose one of many options (I think this how Florida does its voucher program).
LikeLike
I have to agree with Leslie on this one. Forgive the rant, but I think a brief tutorial is in order:
I am astounded that any professor would advocate for the use of norm-referenced tests. It shows a gross disconnect with the actual work of schools, with how student data is used in schools, and with current school accountability practices. I’d add, with a lot of love and only a little shade, that ed reform talking heads, bloggers, and policy wonks who opine about such things could do everyone a huge service by learning more about quality pedagogy and instructional practices – for example the difference between these types of tests. I’m not saying everyone needs to be a teaching and learning expert, but if we were physicians I’d not want someone to confuse an X-Ray, Electrocardiograph, and MRI when establishing or advocating policy.
Norm-referenced testing can really only help us understand the individual relative to a previously tested whole. It cannot tell us anything about progress toward an absolute standard or progress by a group. Moreover, it may not reflect the actual performance of the current population and until recently the groups who have been tested to establish the norm have not represented adequately historically disenfranchised minorities or the poor. Tests such as the CAT, Iowa, and TerraNOVA often do reflect more heavily a relatively homogenous group or sometimes have test design reflecting the wants and needs of the designer’s geographic area. Thus these tests often reflect some degree of bias that makes their every nature as tests that place you on a relative performance scale problematic.
Further, when establishing a standard for performance at a grade level it poses significant issues. For example, we know that we want all third graders reading at or above grade level by the end of the grade in order to maximize the chances of later academic success. A norm referenced test redistributes children’s results along a bell curve meaning that at the very least we would need to supplement testing with another test such as STEP in order to have a sense of how they are performing relative to a grade level standard. Thus it is not helpful to the classroom teacher, except for perhaps establishing special education status.
Norm-referenced tests are often used to determine if students deviate significantly from the mean. Using a normal distribution, we can see (with the WISC or Woodcock Johnson) if a student may qualify as needing special education support or gifted services. Historically they have also been used to make determinations about who went into which track a student was placed. The possible prejudiced implications of this aside, it can do nothing to provide for the determination if schools are successfully teaching students what is needed for them to be successful.
In terms of teacher and school evaluations, norm referenced testing is useless and therefore undermine the very accountability engines that are necessary in a decentralized system. Because the results shared based on a national norm comparison, any school that is not selective should show a distribution of student performance that reflects and approximates the national norm. That is, the larger the group of students the more likely the school will reflect the national bell curve. Any significant deviation from that bell curve is unlikely, or is perhaps the result of tampering or secret selectivity (or de-selectivity in the case of schools to which students and families default).
This professor is just wrong. If you’d like to discuss, I’m headed to the neutral ground now to prep for Endymion. You’re welcome to join @ Lopez and Canal – Neutral Ground CBD Side of intersection.
LikeLike
Bill, thanks for note! Will review. Hope paths cross over Mardis Gras :)
LikeLike