Page 82 - Demo

Chapter 580the total self-efficacy score 1 year following graduation. The change in self-efficacy was calculated by subtracting the baseline total self-efficacy score from the followup total self-efficacy score. Average self-efficacy was calculated by dividing the sum of baseline and follow-up total self-efficacy scores by two. The relationship between the baseline self-efficacy and the follow-up self-efficacy was initially studied by a Bland-Altman plot (Kirkwood & Sterne 2003) and then described by an errors-in-variables model fitted by orthogonal regression (Deming regression) (Cornbleet & Gochman 1979). For comparisons of programmes, chi-square tests, Fisher’s exact tests, MannWhitney tests, and t-tests were used. Predictors of the change in self-efficacy were studied by linear regression and multiple regression analyses. To this end, categorical independent variables were dichotomized due to the modest sample size. Referral behaviour was also studied by multiple regression analyses. The results of the multiple regression analyses were presented as regression coefficients with standard errors or as partial correlation coefficients. Stata Release 15 (Stata Corp. 2017, College Station, TX, USA) was used for all statistical analyses.ResultsAll participants completed the Endodontic General Self-Efficacy Scale questionnaire. Since the change in self-efficacy for the graduates of each school was similar, the data were combined (Figure 1). Self-efficacy of most of the participants increased within the first year following graduation (Figure 1a). The correlation between the change in self-efficacy and average self-efficacy was statistically significant (rho = −0.32, p = .01). The increase in self-efficacy was higher for those whose baseline self-efficacy was lower, and lower for those whose baseline self-efficacy was higher. The ‘Deming-regressionline’ (Figure 1b) shows the relationship between follow-up and baseline selfefficacy. The line has intercept 11.7 (se = 2.6) and slope 0.69 (se = 0.09) and was obtained from an errors-in-variables model fitted by orthogonal regression. The expected change in self-efficacy was 8.6 for those with baseline self-efficacy of 10 (i.e. the lowest possible total self-efficacy score), and it was −0.8 for those with baseline self-efficacy of 40 (i.e. the highest possible total self-efficacy score).Annemarie Baaij.indd 80 28-06-2023 12:26