Empirical Analysis. In this section I will report the results from the proposed nonhomogeneous Hidden Markov Model together with a static model which did not consider the relationship dynamics between players and the mobile application. I will especially focus on the part of results concerning my three research questions on points pressure, Gambler’s fallacy and Hot-hand, and the impact of social status.
Empirical Analysis. THE GAP APPROACH The data collected by the ENVIPOLCON project describe the development in the period 1970 – 2000 for 40 environmental policy issues in 21 European countries plus – for reasons of comparison – the USA, Mexico and Japan. The policy issues cover the whole range of environmental problems (air, water, waste, noise, energy and climate, nature protection, etc.). Moreover, 682 Journal of European Public Policy the list includes policies relating to tradable products and production processes as well as procedural measures such as environmental impact assessment or eco- labelling. Finally, there are policies which have become subject to international regulation at any point in time during the period 1970 – 2000 and policies which have not. The data were collected by way of detailed questionnaires filled in by established experts in the respective countries and carefully checked and double-checked for completeness and comparability by the project team (Xxxxxxx et al. 2008). For this paper, two types of data were used: Downloaded by [Radboud Universiteit Nijmegen] at 04:52 09 July 2012
Empirical Analysis. The nature of the data implies a need for either a fixed effects model or a random effects model. This is due to the fact that the data is in the form of panel data. While a pooled ordinary least squares (OLS) model may also be utilized for panel data, a pooled OLS requires the assumption that there are no unique variable characteristics of the entities in the data set. School districts do have unique characteristics that vary and therefore, I cannot use a pooled OLS for the panel data in this study. Thus, instead, I consider only a fixed effect model and a random effects model. Fixed Effects Model I consider a fixed effects model in this study, which is represented by the equation: 𝐷𝑟𝑜𝑝𝑜𝑢𝑡𝑖𝑡 = 𝛽1 𝑂𝑛𝑒𝑌𝑒𝑎𝑟𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤1,𝑖𝑡 + 𝛽2 𝑇𝑤𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤2,𝑖𝑡 + 𝛽3𝑋3,𝑖𝑡 + 𝛼𝑖𝑡 + 𝑢𝑖𝑡 where 𝐷𝑟𝑜𝑝𝑜𝑢𝑡𝑖𝑡 represents one of the dependent variables in question, as described in Section 3A, 𝑂𝑛𝑒𝑌𝑒��𝑟𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤𝑖𝑡 and 𝑇𝑤𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤𝑖𝑡 represent the independent variables in question, 𝑋3,𝑖𝑡 represents a vector of explanatory variables utilized as control variables, 𝛼𝑖𝑡 represents the unknown intercept for each entity, and 𝑢𝑖𝑡 represents the error term. 𝑂𝑛𝑒𝑌𝑒𝑎𝑟𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤1,𝑖𝑡 and 𝑇𝑤𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤2,𝑖𝑡 are dummy variables. While fixed effects is useful to this study due to the nature of the data, it is important to note that fixed effects may not be a flawless model for this study. This is mostly due to the fact that fixed effects models are not the best choice for variables that change slowly over time. Dropout can be considered a slowly changing variable and therefore, there are some limitations with the use of a fixed effects model in this study. Despite this limitation, we find that a fixed effects model is still a superior choice to a random effects model for this particular study. Random Effects Model I also consider a random effects model for this study, which is represented by the equation: 𝐷𝑟𝑜𝑝𝑜𝑢𝑡𝑖𝑡 = 𝛽1 𝑍𝑒𝑟𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤1,𝑖𝑡 + 𝛽2 𝑂𝑛𝑒𝑌𝑒𝑎𝑟𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤2,𝑖𝑡 + 𝛽3 𝑇𝑤𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤3,𝑖𝑡 + 𝛽4𝑋4,𝑖𝑡 + 𝛼 + 𝑢𝑖𝑡 + 𝜀𝑖𝑡 where 𝐷𝑟𝑜𝑝𝑜𝑢𝑡𝑖𝑡 represents one of the dependent variables in question, as described in Section 3A, 𝑍𝑒𝑟𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤1,𝑖𝑡, 𝑂𝑛𝑒𝑌𝑒𝑎𝑟𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤2,𝑖𝑡, and 𝑇𝑤𝑜𝑌𝑒𝑎𝑟𝑠𝑈𝑛𝑑𝑒𝑟𝐿𝑎𝑤3,𝑖𝑡 represent the independent variables in quest...
Empirical Analysis a. Data As aforementioned in the introduction section, variables I plan to use include the trade volumes in pharmaceuticals between the US and Singapore, the price of pharmaceutical products in the US, and employment in the US pharmaceutical industry. Data on the exports and imports of pharmaceuticals is available from 1991 to 2019 at the UN Comtrade Database, and the reporter is the US. Data on employment is retrieved from the US Bureau of Labor Statistics (BLS), which also covers from 1991 to 2019. The employment indicator “All Employees, Thousands” on the BLS is available in the monthly form, so I calculated the annual average to represent the annual employment. I also included employment from the following two industries to represent the total employment in the entire pharmaceutical industry: the pharmaceutical and medicines industry and the pharmaceutical preparation industry. While the pharmaceutical and medicines industry focuses more on manufacturing, the pharmaceutical preparation industry needs to process drugs for final consumption. Data for employment is seasonally adjusted. Finally, data on the price level, which is available from 2000 to 2017, was retrieved from the Federal Reserve Bank and the price indicator is the personal consumption expenditures price index in the pharmaceutical industry — Pharmaceutical and Other Medical Products Expenditures Price Index. Other variables in my empirical analysis include US exports of pharmaceuticals to the world from Comtrade, US imports of pharmaceuticals from the world from Comtrade, annual GDP for the US and Singapore from the World Bank, the USD to the SGD exchange rate from the Federal Reserve Bank, and the producer price index (PPI) in pharmaceuticals from the Federal Reserve Bank. As I will explain in the analysis subsection later, I will employ two different strategies to examine the effects of the USSFTA on the volumes of trade for this study—Ordinary Least Square (OLS) and Difference-in-Differences (DID). The OLS strategy is related to the gravity model of trade and it is commonly used. However, it may be vulnerable to endogeneity. I therefore employ the DID method, a quasi-experimental design, as a complement to elaborate the casual effects of the USSFTA. For DID, I will use Hong Kong as a counterfactual case to see what would take place if the USSFTA did not exist. Thus, data on the trade flows between the United States and Hong Kong in pharmaceutical products, as well as Hong Kong’s annual ...
