Computational protocol: Relationships of work characteristics to job satisfaction, turnover intention, and burnout among doctors in the district public-private mixed health system of Bangladesh

Similar protocols

Protocol publication

[…] SPSS (version 21) was used for processing and analyzing the data. Internal consistency of the scales was assessed by Cronbach’s α statistic as recommended by Nunnally [] and was acceptable (Table ). Linearity and normality were confirmed by partial plot and p-p plots respectively. Presence or absence of outliers was determined using Cook’s distance (cut-off point <1) and that of multicollinearity by variance inflation factors (cut-off point <10) as well as by tolerance factors (cut-off point >0.1) as described by Tabachnick & Fidell []. No serious violations of assumptions were noted.Means and standard deviations were inspected. One-way multivariate analysis of variance was performed to test any significant differences in work characteristics and outcomes between the doctors of the public primary and public secondary level facilities, and between all public and private doctors. Bivariate correlation analyses were conducted to detect the direction and strength of relations between work characteristics and outcomes. Multiple regression (forward method) was used to identify the key predictors of the outcomes of interest. Ninty 5 % bias corrected and accelerated bootstrap confidence intervals (BCa CIs) were used to assess significance of correlation. An online programme (http://vassarstats.net/rdiff.html) was used to test any significant differences in the corresponding correlation coefficients between the public and private sector using Fisher’s r-to-z-transformation. A p-value <.05 was regarded as significant. Missing data were treated by using the ‘listwise’ option of SPSS. […]

Pipeline specifications

Software tools SPSS, VassarStats, rDiff
Application Miscellaneous