China
Africa
Explore chapters and articles related to this topic
A survey of four Scottish proposed wind farms
Published in Giuseppe Pellegrini-Masini, Wind Power and Public Engagement, 2020
“Ordinal regression”, as presented by Norušis (2011), analyses data, i.e. predictors (independent variables) and the outcome variable (dependent variable), which are of ordinal nature. In fact, the researcher can perform a multivariate regression analysis retaining the ordinal information of the variable, i.e. the scores ordered, for example, on a Likert scale (as in this study), along a continuum that represents the concept measured. This removes the need to consider the variables improperly as interval variables in order to perform a linear regression (Bryman and Cramer, 2009; Field, 2009).
Project management as an anti-planning measure for the effective preparation and implementation of the urban development plans in the Global South
Published in Cogent Engineering, 2023
The role of project management aspects in urban planning and implementation was evaluated by ordinal regression model estimation. The ordinal regression model is a typical regression model that is usually used to predict the behaviour of the ordinal dependent variable (in this case the impact of project management on planning and implementation) premised upon a set of independent variables (project management aspects). However, this model is only applicable if the values are available on an ordinal and arbitrary scale (Williams & Quiroz, 2020). Since data was collected on an ordinal scale for this study, this method is found to be suitable. The model was estimated to evaluate the relative importance and influence of different project management aspects as anti-planning measures to improve the efficiency of urban plan preparation and implementation. In this case, the various project management aspects were considered the predictor variables and the influence of these aspects was considered the dependent variable. The log-linked ordinal regression model was used for the model estimation. Equation 2 presents the log-linked ordinal regression model.
Comprehensive solid waste dumpsite selection in arid cities of Northeastern Ethiopia: A spatial-MCDA approach
Published in Journal of the Air & Waste Management Association, 2022
Teshome Betru Tadesse, Setiye Abebaw Tefera, Mengist Tesfaye Kidane
Depending on the nature of the endogenous variable which is the Likert or ordered variable, the ordinal regression, multinomial logistic regression, and binary logit (if the response variable is recoded into different variables i.e. dummy response) models can be used. If the dependent variable has a category which is a natural ordering, the ordinal regression model fits the data (Gujarati 2004; Rao 2008). To choose a link function, it is helpful to examine the distribution of values for the outcome variable (Elamir and Sadeq 2010).
Exploring the impact of feedback on student performance in undergraduate civil engineering
Published in European Journal of Engineering Education, 2023
The data collected from the questionnaire survey were used to develop an estimation model to examine the relative influence of different types of feedback on students’ performance compared with conventional, unstructured, and informal feedback. For this purpose, log-linked ordinal regression model estimation was used. Since the data collected were on an ordinal scale, and comparative analyses of different feedback systems were required, ordinal regression model estimation was found to be suitable (Ananth and Kleinbaum 1997; Williams and Quiroz 2020). Ordinal regression model estimation offers the advantage of evaluating the impact of different independent variables on the dependent variable, compared with each other or a reference, and forecasts the behaviour of the dependent variable (Lu, Wang, and Tolliver 2019; Williams and Quiroz 2020). In this context, the various forms of feedback were considered to be independent variables and the final result (or impact on learning) was considered to be the dependent variable. For this estimation model, the data obtained from Question 15 of the survey questionnaire (Appendix 1) were used. Using this estimation model, the influence of the independent variables was measured and compared with a reference variable. Therefore, the conventional unstructured feedback (the oral feedback generally given to classes in non-structured form based on the queries of the students or proactively by the lecturers) was used as the reference variable. However, before the model parameter estimation (B) was done, the validity and robustness of the model were checked by using Model fitting information, Goodness of Fit, Nagelkerke (Pseudo R square) and the test of parallel lines (Javali and Pandit 2010; Scott, Goldberg, and Mayo 1997). The higher the B values of the statistically significant variables, the higher the likelihood of their influence. The model estimation was conducted by using IBM-SPSS V.27 Software.