PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 15077743

Asymptotic log-linear analysis: some cautions concerning sparse frequency tables.

Paul W Mielke·Kenneth J Berry·Janis E Johnston
View original source ↗Export metadata ↓

Abstract

Traditional asymptotic probability values resulting from log-linear analyses of sparse frequency tables are often much too large. Asymptotic probability values for chi-squared and likelihood-ratio statistics are compared to nonasymptotic and exact probability values for selected log-linear models. The asymptotic probability values are all too often substantially larger than the exact probability values for the analysis of sparse frequency tables. An exact nondirectional permutation method is presented to analyze combined independent multinomial distributions. Exact nondirectional permutation methods to analyze hypergeometric distributions associated with r-way frequency tables are confined to r = 2.

Explore related subjects

Chi-Square DistributionHumansLikelihood FunctionsLinear ModelsMathematical ComputingProbabilityPsychology, ExperimentalReproducibility of ResultsJournal Article

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Paul W Mielke, Kenneth J Berry, Janis E Johnston. 2004. Asymptotic log-linear analysis: some cautions concerning sparse frequency tables.. https://doi.org/10.2466/pr0.94.1.19-32

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Incidence and predictors of recurrent restenosis following implantation of drug-eluting stents for in-stent restenosis.

OBJECTIVES: We investigated the incidence and predictors of recurrent restenosis after drug-eluting stent (DES) implantation for in-stent restenosis (ISR) in routine clinical practice. BACKGROUND: Although DESs have been increasingly used for treatment of ISR, little is known about the predictors of DES failure. METHODS: We determined the incidence of recurrent restenosis and major adverse cardiac events (MACE) in 224 consecutive patients with 239 lesions treated with sirolimus-eluting (n=217 lesions) or paclitaxel-eluting (n=22 lesions) stents for the first episode of ISR. RESULTS: The procedural success rate was 99.2%, and in-hospital complications did not occur in any patient. Follow-up angiography at 6 months was obtained in 73.7% of patients. Angiographic re-restenosis rate was 12.6%, and target lesion revascularization was required in 7.6% of patients. Of the 22 incidents of re-restenosis, 15 were focal (68.2%), 5 were diffuse (22.7%), and 2 were total (9.1%) restenosis. Univariate analysis showed that lesion length, use of paclitaxel-eluting stent, and number of stents per lesion were significant predictors of re-restenosis. In multivariate analysis, however, lesion length and use of paclitaxel-eluting stent were independent predictors of re-restenosis. During the follow-up (mean, 18.3+/-8.1 months), there were 4 deaths (1 cardiac, 3 noncardiac), but no nonfatal myocardial infarctions (MIs). MACE occurred in 18 patients. The cumulative probability of MACE-free survival was 92.9+/-1.8% at 1 year and 90.5+/-2.4% at 2 years. CONCLUSIONS: DESs are highly effective for treatment of ISR, with recurrent restenosis related to lesion length and type of DES.

Chi-Square Distribution↗
PubMed2007

Differences in yeast intolerance between patients with Crohn's disease and ulcerative colitis.

PURPOSE: Alimentary factors, especially those modifying the intestinal flora, may influence the course of inflammatory bowel disease. It is known that T and B cells of patients with Crohn's disease can be stimulated with the yeast antigen, mannan. We evaluated the impact of eating habits with special respect to food containing yeast on the course of inflammatory bowel disease. METHODS: Questionnaires were sent to 180 German-speaking patients of the Inflammatory Bowel Disease Outpatient Clinic at the University Hospital Bern, Switzerland. The following information was obtained by the questionnaires: (1) course of disease, (2) eating habits, (3) environmental data, and (4) inflammatory bowel disease questionnaire. The survey was anonymous. RESULTS: A total of 145 patients (80.5 percent 95 with Crohn's disease, and 50 with ulcerative colitis) responded. Food items containing yeast were better tolerated by patients with ulcerative colitis than by patients with Crohn's disease. A significant difference between the two groups was observed concerning food containing raw yeast (dough, P = 0.04; and pastry, P = 0.001). CONCLUSIONS: Food items containing raw yeast led to more frequent problems for patients with Crohn's disease than for patients with ulcerative colitis. This observation supports our previous data, which showed the stimulatory effect of the yeast antigen, mannan, on B and T cells of patients with Crohn's disease but not of controls.

Chi-Square Distribution↗
PubMed2007

Short-term outcome after laparoscopic or open restorative mesorectal excision for rectal cancer: a comparative cohort study.

PURPOSE: The laparoscopic approach to rectal cancer is still a controversial procedure. A comparative cohort study was conducted to assess short-term results of laparoscopic restorative mesorectal excision. METHODS: From January 1998 to December 2000, laparotomy was performed on all primary rectal cancer undergoing radical excision. From January 2002 to September 2004, all cases about to undergo radical excision were considered for laparoscopy. Patients with fixed tumor or T4, indications for synchronous hepatectomy, emergencies, and medical contraindications were not included. The study was based on the intention-to-treat principle. RESULTS: Short-term outcome was compared between the laparoscopy group (n=104) and the laparotomy group (n=68). Demographic, general and tumor data, and rates of preoperative irradiation were comparable, as were surgical procedures and perioperative management. Hospital mortality (1 and 2.9 percent, P=0.33) and three-month overall morbidity (43.3 and 48.5 percent, P=0.49) were comparable between laparoscopy and laparotomy groups. Surgical complication rates were comparable (39.3 and 35.5 percent, P=0.58), but a significantly lower medical complication rate was observed in laparoscopy patients (8.7 and 20.6 percent, P=0.025), mainly because this group had fewer respiratory complications. Hospital stay was shorter in laparoscopy patients (10 and 14 days, P<0.001). Oncologic quality criteria were comparable, in terms of number of lymph nodes, lateral and distal margins, and delivery of postoperative chemotherapy. CONCLUSIONS: The laparoscopic approach to restorative mesorectal excision for cancer does not increase postoperative morbidity or reduce oncologic quality. Our results suggest that the short-term outcome is probably improved with this procedure.

Chi-Square Distribution↗