The Casual Theory and The Result of a Multivariate Linear Regression Analysis - Coursework Example

CASE STUDIES Case studies Case Study A2 In order to fulfill the assignment above, I have chosen the paper Case Study A2 and I intendto resolve my paper so that it can answer the questions as per the assignment. I have described the casual theory and the result of a multivariate linear regression analysis. Moreover, in my paper I have discussed dependent and independent variables as to how they correlate and functions. The paper has touched on the causal theory. It explains what is causal theory and the causal theory used in the case study selected. Additionally, it has explained a brief exposure on the hypothesis findings in the case study. The hypothesis study includes the alternative and null hypothesis. The latter is explained on how it contradicts with the causal theory. In addition, the paper introduce to the reader on the empirical studies and the linear regression used in the case study. Furthermore, there is a good explanation of reverse causality in the case study and how the confounding variables are given interest in the case study. Casual Theory A Casual Theory is a tentative explanation and conjecture that explain the causes of some phenomenon of interest to a researcher.1 In Case Study A2, the casual theory is that the growth rate of a country is influenced by the effect of GDP per capita, mobile cellular subscriptions, foreign direct investment net flow in the percentage of GDP and the rate of primary education enrollment. In this case study, the dependent variable is policy score while the independent variables are GDP per capita, mobile subscribers, foreign direct investment and primary education enrollment. In the case study A2, a good example to explain the problem is the theory of the economic voting. This theory states that “the voters hold the government accountable for the consequence of its economic policy.” Therefore, in this our case the independent variable is economic performance while the dependent variable is election policies. Economic performance of the government party influences the national election outcome and this can be derived that the outcome of the election will depend on the economic performance of the of any party in the government. The voters usually punish the governments’ party during the national election based on the economic performance of the government. Hypothesis Hypothesis theory is based on the relationship of variables that is expected to be seen.2 For example, in our case study, the hypothesis is that the growth rate of the real gross domestic product influences the vote share of a government party during the national election. For example, with an increased growth rate of the real GDP, there is a high number of the government party’s share of vote in the national election. 3 Below is a graph that shows an explanation of the above statement. The graph shows the incumbent party’s vote share in percentage in the Y-axis against the gross domestic product per capita growth rate as percentage On the X-axis. Figure 1: Graph of National Executive Election The above graph shows that the incumbent party’s share is affected by the percentage of the real growth rate per capita. It is good to note the relationship between the economic growth rate in terms of GDP per capita and the incumbent part’s vote share. Therefore, the graph is a clear indication that the party’s vote share is influenced by the rate of economic growth rate. Figure 1 is a copy from the case study selected. Null Findings There may also a null hypothesis that the scientists may want to test. This then can be described as follows. Figure 2: Graph Showing Null Finding. Null hypothesis means that there is no relationship between the two variables.4This is well represented in the above graph. The incumbent party’s vote share percentage is not affected by the real gross domestic per capita growth rate. The Null Hypothesis is a contradictory hypothesis to the Causal Theory because from the case study, the null hypothesis is drawn to show that it is not a must that there is a causal relationship between the variables. The finding of null hypothesis is important because it enables the researcher to take a test of her hypothesis by conducting their study. Therefore, the results in the above graph is an evidence that the researcher in my case study never assumed things and that is why they found another control variable. For the future studies, null findings help the researchers not to be rigid to their theory or alternative hypothesis. For instance, the null hypothesis on the graph above shows that there may be no relationship between the result of election and economic growth rate. This calls for a more elaborate study to know what influences the results of election. Therefore, for this case, null finding helps to build more caution in our studies. Empirical Studies An empirical study is a process that requires scientist to evaluate collected evidence to make a correct judgment of whether the evidence favors the hypothesis or do not.5 For example, in our case study, first, the theory can be explained by collecting data for multiple national elections held. Empirical studies or analysis will help the scientist to test his argument about the effect of the economic growth rate to the outcome of the national election. Therefore, to support this argument, a copy of a graph from the case study selected previously is taken. The graph shows different levels of election outcome in a national election. The graph below is an example of United States Presidential elections held in different years. Figure 3: Graph showing the United States Presidential elections over years. The empirical study in this case shows the different effect of the real GDP growth rate to the share of votes for the incumbent President’s party. Apart from one single dependent variable, there may be other control variable in place. The control variables also have an influence on the independent variable. It is not real that you find just only one dependent variable in the real case. There are other factors (control variable) that are also responsible for the outcome. For example, in our cases study, there is another variable that causes the outcome of the national election in Canada and United States. Below is a graph presentation of a general outcome of national election as a result of effect of percentage change in unemployment rate. It is a representation from the selected case study. Figure 4: Graph showing the Effect of Rate of Unemployment on Election Results The control variable (percentage change in unemployment rate) has an effect on the election results. As the rate of unemployment reduces, the incumbent party’s vote share increases. It is the work of scientist, when testing an empirical test, to consider some other factors that may control the result of the study. Linear Regression Linear regression is a presentation that shows the relationship between two variables linearly. One variable must independent and the other dependent. The graph below shows in a straight line the relationship between the two variables. Basing this argument on the case study presented in the paper, the linear regression graph between the real economic growth rate of a country and result of the election is as follows. Figure 5: Graph of Linear Regression The above graph sows how the vote percentage share is influenced by the economic growth rate of United States. Reverse Causality Reversal causality is the condition that two variables in a study may have a mutual causal effect.6The question of reverse causality is answered on the case study selected. There is no real evidence that there could be a double cause on the variables. When a scientist who conducted a certain study on the case study selected did not investigate whether the outcome of the economic growth could be affected by the incumbent party’s vote share, this calls for another study to answer this question. The study may be conducted a few years after the election. Confounding Variables Confounding variables are variables that may be hiding or has an effect to either independent or dependent variable.7 It may also apply at the same time. Therefore, as it is outlined in the case study selected, it is appropriate before making final decision the scientist to find whether there is another variable. There may be another driving force causing the results of election other than economic growth rate only. Additionally, confounding variable may have a hand in either the result of election or the economic growth rate or the both at the same time. It is important to note that confounding variables are also control variables. In the cases study selected by the paper, the researchers omitted some of good causal variables like unemployment rate. The paper has shown it vividly that the unemployment rate decreases while the party’s vote share increases. Therefore, these confounding variables should have been used in the study. However, without some of these control variables, the result would not be the same as compared to when they are put in place. Bibliography Baron and David Kenny. "The Moderator–Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations." Journal of personality and social psychology 51, no. 6 (1986): 1173. Bates, Douglas and Donald Watts. Nonlinear regression: iterative estimation and linear approximations. John Wiley & Sons, Inc, (1988) Evans, Gareth, and Altham. "The Causal Theory of Names." Proceedings of the Aristotelian Society, Supplementary Volumes (1973): 187-225. Frank, Kenneth . "Impact of a Confounding Variable on a Regression Coefficient." Sociological Methods & Research 29, no. 2 (2000): 147-194. Lacey, Hugh. "The Causal Theory of Time: A Critique of Grunbaums Version." Philosophy of Science (1968): 332-354. Kwiatkowski, Denis, Peter Schmidt. "Testing the Null Hypothesis of Stationarity against the Alternative of a Unit Root: How Sure are We that Economic Time Series have a Unit Root?" Journal of econometrics 54, no. 1 (1992): 159-178. Watts and Jerold Zimmerman. "Positive Accounting Theory." (1986). Read More