Figure 1

Regression Tree for predicting R&D Culture Index Score with mean score (and number of respondents) given at the terminal nodes. The statistical modeling used in the above analysis adjusted for the following covariates: age, gender, work status, professional group, professional registration, postgraduate qualification, higher training qualification, number of years registered, research training, time in post, research involvement, named applicant on research funding application, named applicant on research ethics application, author of peer-reviewed publication, author of non-peer reviewed publication, presenter of research conference paper. (professional group, research training and research involvement were the covariates kept in the final multiple regression model and the final tree regression model).