Comments on: Method Madness http://www.eduwonk.com/2007/11/method-madness.html Education News, Analysis, and Commentary Sat, 11 Feb 2012 21:00:28 -0500 http://wordpress.org/?v=2.8.2 hourly 1 By: Anonymous http://www.eduwonk.com/2007/11/method-madness.html/comment-page-1#comment-178 Anonymous Thu, 29 Nov 2007 16:56:00 +0000 http://www.eduwonk.com/wordpress/2007/11/method-madness.html#comment-178 FYI, there is a now an amendment attached to the HEA reauthorization that cancels both the Upward Bound evaluation and the Absolute Priority that justifies it, AND says "The Secretary shall not require an eligible entity desiring to receive assistance under this chapter to recruit students to serve as a control group for purposes of evaluating any program or project assisted under this chapter." Very disturbing precedent! FYI, there is a now an amendment attached to the HEA reauthorization that cancels both the Upward Bound evaluation and the Absolute Priority that justifies it, AND says “The Secretary shall not require an eligible entity desiring to receive assistance under this chapter to recruit students to serve as a control group for purposes of evaluating any program or project assisted under this chapter.” Very disturbing precedent!

]]> By: age group mom http://www.eduwonk.com/2007/11/method-madness.html/comment-page-1#comment-130 age group mom Thu, 15 Nov 2007 13:13:00 +0000 http://www.eduwonk.com/wordpress/2007/11/method-madness.html#comment-130 No, not optimal design. Researchers should use the most valid design. The design with the most scientific power to reveal program effects. It shouldn't even be up for debate, actually.<br/><br/>Randomized trials are the most powerful research design. Followed by observational studies that use methods such as propensity modeling to improve the chance to for causal claims. <br/><br/>This is not some arbitrary conclusion that Congress can change with a wave of a wand. No, not optimal design. Researchers should use the most valid design. The design with the most scientific power to reveal program effects. It shouldn’t even be up for debate, actually.

Randomized trials are the most powerful research design. Followed by observational studies that use methods such as propensity modeling to improve the chance to for causal claims.

This is not some arbitrary conclusion that Congress can change with a wave of a wand.

]]>