testing fixed effects in binomial lmer...again?
On 09/01/2008, at 10:45 AM, Reinhold Kliegl wrote:
Not sure it applies exactly to the cases discussed so far here, but there is analytical and simulation evidence against using BLUP-based residuals of mixed models for bootstrapping. See:Morris, J.S. (2002). The BLUPs are not "best" when it comes to bootstrapping. Statistics & Probability Letters 56, 425-450. A PDF is available on his website. Reinhold
A solution to this problem is Carpenter, JR; Goldstein, H; Rasbash, J "A novel bootstrap procedure for assessing the relationship between class size and achievement", JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES C-APPLIED STATISTICS, 52: 431-443 Part 4 2003 but I expect that parametric bootstrapping is enough for most problems. Ken