4.3. Statistics

Wilcoxon signed-rank test was used as follows: to assess the significance of changes in reported symptoms and problems and satisfaction with life before and after surgery in paper I; to assess the difference between various preoperative and postoperative anamnestic and clinical measures included in Helkimo´s Anamnestic and Dysfunction indexes and occlusal parameters in paper II; in paper III, to assess the differences between the numbers of patients in the different categories of motivations to seek treatment and between the ages of patients with and without neurosensory disturbances; in paper IV, to assess the differences between the test and control sides in 2-PD and VST test scores at every control visit .

Spearman’s correlation coefficient was used in paper I to assess the relationship of various factors with overall satisfaction and in paper II to study the association between Ai or Di and the specific type or magnitude of dentofacial deformity, between Ai or Di and the patient’s age and between the occlusal parameters or Ai or Di and the specific type of malocclusion. In paper III, Spearman’s correlation coefficient was used to assess the relationship between the patient’s age and the inpatient period or condylar resorption or TMD; in paper IV, to test the association between the subjective sensation and such variables as age, gender, 2-PD and VST; and in paper V, to test the differences between costs in various phases and the tested clinical and cephalometric values.

Wilcoxon’s two-sample test was used in paper I to assess significance in relation to the categorical variables. In paper V, analysis of covariance (ANCOVA) was used to test the costs of orthodontics, surgery and inpatient care and total costs in relation to the different clinical and cephalometric values and measurements. The outpatient surgery clinic cost was tested first by using ANCOVA and then by applying a regression-type model, because none of the categorical variables were significant. The cost differences between the various diagnostic and operative groups were tested with analysis of variance and paired t-tests, and due to the multiple comparisons, a Bonferroni procedure to correct the p-values was added in paper V.

For paper IV, calibration of the two investigators for 2-PD was done, i.e. the inter-examiner difference was tested with non-operated volunteers by t-test and was found to be negligible.