【BUSS5221 – WEEK12-Amy】 Page 1 of 6 BUSS5221 WEEK12 Data-based argument 主讲人：Amy 【BUSS5221 – WEEK12-Amy】 Page 2 of 6 1.Weaving Numbers into narratives Data alone do not answer questions or clarify problems on their own. At their best they help us understand what is happening They have limited functions in proving how and why things happen. 2.Narrative-Numbers nexus: Principles concerning communication with and through data Listening to data carefully and working sensitively with it to help you devise robust arguments that are compelling as well as robust Principles concerning critical assessment of data based arguments Positivism vs constructivism 3.Communication goes both way As in any aspect of life, communicating with data is complex. It is important that you develop the skills of being able to overcome biases and prejudices so that you can absorb what the data are trying to convey. Presenting and reconfiguring data in different ways is one way to help with this process. 3.1 Listening to the data 【BUSS5221 – WEEK12-Amy】 Page 3 of 6 3.2 Data needs to be communicating 【BUSS5221 – WEEK12-Amy】
Page 4 of 6 3.3 Visualisation as both ‘listening’ and ‘communicating’ Snow’s 1854 Cholera Map of London 【BUSS5221 – WEEK12-Amy】 Page 5 of 6 Minard and his account of Napoleon’s army’s fate 【BUSS5221 – WEEK12-Amy】 Page 6 of 6 4. Assessing the quality of analysis Positivist/naturalist tradition Constructivist tradition Facts are assumed to be phenomena that are discovered Draws on the model of natural sciences Quality of an argument is assessed by how well facts as discovered support or disprove theories Facts are not assumed to be already existing and waiting for discovery but that as our understanding of things has emerged and evolved, our ability to make sense of the world has improved Need to understand under what conditions facts are constructed Contributions from the positivist tradition: minimum technical requirements Reliability Validity Accuracy Reproducibility Contributions from the constructivist tradition: the importance of context: Question the source of the information Beware of unfair comparisons If it seems too good or too bad to be true Think in orders of magnitude Avoid confirmation bias Consider multiple hypothesis