S plus 8.0 for linux

Statistical analyses were performed with S‐plus 8.0 for Linux (Tibco software, Palo Alto, CA). The data were first analyzed descriptively for mean values and percentage distribution. To detect significant differences, p‐values were calculated by means of a generalized linear mixed model for binary outcomes using a logit‐link. A univariate separate model was built for the different sinus relations (distant, <2 mm, touching, protruding, and touching or protruding) separately. Explanatory variables were tooth category, root category and age category.
Comparison between the root categories were assessed for each tooth type separately. The difference between age categories were assessed once for each tooth category and once for each combination of root and tooth category.
In addition, the occurrence of roots with a close sinus relation was compared between the genders.

Statistical analyses were performed with S‐plus 8.0 for Linux (Tibco software, Palo Alto, CA, USA). The mean and percentage of contact area and volume were calculated for assessing the individual and combined relationship of mesiobuccal (MB), distobuccal (DB), and palatal root (P) with the MS. A one‐way analysis of variance (ANOVA) test was applied to detect a statistically significant difference (p < .05) for both contact area and volume between the roots and MS. p Values were adjusted for multiple comparisons according to Tukey. The relation between age and contact area was assessed by a weighted linear mixed model with age as an explanatory variable and contact area as a continuous variable. Tooth, nested in patient, was used as a random variable. A weighing was applied that was proportional to the age. Inter‐ and intra‐observer agreement was calculated using the intra‐class correlation coefficient (ICC).