This document discusses the differences between scientific and postmodern views of language, as well as the differences between correlation and causation. It notes that qualitative research can take a scientific view of language, and qualitative descriptions can sometimes be viewed quantitatively. It then provides examples of correlation not implying causation, such as the relationships between height and weight, or temperature and ice cream sales. The document outlines three requirements for determining causality between variables and a two-step process for making causal inferences.
This document discusses the differences between scientific and postmodern views of language, as well as the differences between correlation and causation. It notes that qualitative research can take a scientific view of language, and qualitative descriptions can sometimes be viewed quantitatively. It then provides examples of correlation not implying causation, such as the relationships between height and weight, or temperature and ice cream sales. The document outlines three requirements for determining causality between variables and a two-step process for making causal inferences.
This document discusses the differences between scientific and postmodern views of language, as well as the differences between correlation and causation. It notes that qualitative research can take a scientific view of language, and qualitative descriptions can sometimes be viewed quantitatively. It then provides examples of correlation not implying causation, such as the relationships between height and weight, or temperature and ice cream sales. The document outlines three requirements for determining causality between variables and a two-step process for making causal inferences.
The University of Electro-Communications Graduate School Second semester 2022
International Politics and Economy 3rd time Causality (1) Two stages 2 Scientific = modern view of language Postmodern view of language Two histories: empirical history and postmodern history Anthropology Overall understanding of human activities through observation by researchers Last class 3 Q. Do scientific and postmodern views of language correspond to the so-called quantitative-qualitative distinction? A. Not for two reasons (1) Even qualitative research can take a scientific view of language ex. Empirical history (2) Some argue that qualitative descriptions are actually quantitative ex. "Crime is very serious in this country." If the referents of ``this country'', ``crime'', ``very'', and ``serious'' are clear, they are the same as binary (0/1) expressions. 4 Previous test answer example 1 About your research ・Safe remote control of robot arms that are resistant to cyber attacks ・Detection of lifestyle changes using lifelog data ・Fabrication of cold atomic clusters ・Optimal parameter settings for nuclear reactor safety systems ・Performance evaluation of biometric identification system ・Performance improvement of limb robot 2 Is your research scientific? ・ It is scientific because it can be judged with clear criteria according to the observed results. ・Criteria for “safety” and “improvement” are subjective and unscientific 5 Correlation and causation variable A symbol that takes a value (x, y, height, weight...) Correlation When the value of one variable changes, that the value of another variable also changes Positive correlation: one increases, the other increases Negative correlation: one increases, the other decreases Causality A cause and effect relationship between one variable and another 6 Correlation ≠ Causation Height and weight Taller people will weigh more A heavier person would be taller If I gain weight, will I grow taller? Temperature and ice cream sales Tobacco and healthcare spending A typical example of pseudo-correlation Both expenditures are correlated with income [income – Spending, tobacco == medical care, Spending] 7 Three requirements for causality covariance that the two variables are correlated Time lag Changes in causative variables (explanatory/independent variables) precede changes in consequential variables (explained/dependent variables) Control Confirmation of changes in explanatory and dependent variables after removing the influence of other variables cf. ceteris paribus 8 Two steps to identify causality descriptive inference "What's going on?" To describe an event in a probable (agreeable form) causal inference "What is the cause and what is the effect?" Demonstrate plausible causal relationships based on descriptive reasoning 9 descriptive reasoning? Descriptive reasoning ≠ factual identification Whether it is a natural phenomenon or a social phenomenon, there are times when we ultimately do not know what the facts are. ex. traffic accidents, atmospheric pressure, number of suicides Shareable Facts If you cannot identify what the facts are, consider what many people accept to be facts as facts and proceed to identify causal relationships. 10 Procedure for causal inference Question setting ex. What kind of policies are necessary to extend healthy life expectancy? → dependent variable (y) Healthy life expectancy Explanatory variable (x) Factors affecting healthy life expectancy Descriptive reasoning ex. healthy life expectancy = Period during which you can live without any restrictions on your daily life due to health problems Policy (Public burden of medical expenses, dietary guidance) Personal attributes (age, occupation, annual income, gender...) Causal inference Consider whether the target policy changes and other changes in healthy life expectancy at certain times