Upon a first reading of the link you might find it hard to believe that IMP was a complete and total failure when used as part of the University of Washington NSF funded 5 year professional development "PD Cubed" at Cleveland High School in Seattle, WA. ... from 2004-2005 through 2008-2009. IMP was used for three years as the school based project. [06-07; 07-08; 08-09.] See this link for Cleveland HS data.
As incredibly poor as these Cleveland math results were, these types of results are far from unusual when the "experts" produce or rate programs as worthy of use. The 1999 Exemplary and Promising Mathematics Programs were largely a complete bust. See this link as to how these programs gained the "Exemplary or Promising" rating. Reading the beginning one-paragraph abstract from the 71 page download is quite revealing and demonstrates how wrong "supposed math experts" can be.
Exemplary = Cognitive Tutor Algebra; College Preparatory Math; Connected Mathematics; Core-Plus Mathematics Project; Interactive Mathematics Program (IMP)
Promising = Everyday Mathematics; MathLand; Middle-school Mathematics Through Applications Project (MMAP); Number Power; University of Chicago School Mathematics Project (UCSMP)
In 1995 the first TIMSS International Mathematics Testing occurred and every four years TIMSS is given. The most recent TIMSS occurred in 2015. The highest scoring TIMSS countries are in East Asia and each of these far surpasses the USA. It is particularly interesting to examine what takes place in Singapore and contrast it with the bizarre recommendations that occur in the USA.
Singapore bases its instructional program not on “inquiry” or “student-centered design” but on a combination of teacher-centered Traditional Instruction with Direct Instruction and then builds from this to Teaching for Understanding. Why is the USA not doing the same?