{"id":"arxiv/2608.22554","url":"https://arxiv.org/abs/2608.22554","tags":["stat.ME","math.ST"],"score":0,"title":"Detection of Cognitive Diagnostic Model Misspecification using New Lancaster-Chesher Information Matrix Tests","domain":"stat.ME","source":"arxiv","authors":["Richard M. Golden","Reyhaneh Hosseinpourkhoshkbari"],"reviews":0,"subject":"stat.ME","summary":"Model specification tests play a crucial role in evaluating the appropriateness of probability models for estimation and inference. Existing methods for the detection of model misspecification such as the chi-square goodness-of-fit (GOF) tests and more recently the M2 statistic \\autocite{MaydeuOlivares2005,MaydeuJoe2014} tend to result in test statistics with excessive degrees of freedom for models with larger numbers of parameters. An alternative approach is based upon the Information Matrix (IM) equality. The IM equality asserts that if a probability model is correctly specified, the…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-08-23","reactions":{}}