ANALYSIS METHOD Clause Samples

ANALYSIS METHOD. The analytical method used in this research is using 9A version of Computable General Equilibrium (CGE) of Global Trade Analysis Project (GTAP) used to see the impact of the ASEAN-Canada cooperation framework on economic and sectoral potential in ASEAN member countries. This study conducted two simulations simulated by cutting 50% tariff and elimination of 100% tariff (full liberalization) for all commodities between ASEAN and Canada. The selection of CGE models in this analysis is due to the purpose of this study is to calculate how much profit is gained on the enactment of the ASEAN-Canada FTA Agreement on the economic potentials obtained when the entry into force of the ASEAN-Canada FTA or in other words ex-ante analysis. There are several empirical studies that calculate the impact (FTA) and its members in terms of macroeconomic and sectoral. ▇▇▇▇ ▇▇.▇▇ (2010) uses a GTAP analysis tool to see the impact of the ASEAN-China FTA (ACFTA) FTA on trade, exports and imports, and GDP. The results of the implementation of ACFTA will have a significant impact on trade, production and GDP of ACFTA members either bilaterally and with other member countries. Other researches, ▇▇▇▇▇▇▇▇, ▇▇▇▇▇▇▇ and ▇▇▇▇▇, ▇▇▇▇▇▇▇▇ (2014) in his paper on the impact of tariff reductions in NAFTA on the trade and welfare of its members, said that the impact of tariff reductions would increase welfare in Mexico by 1.31%, USA by 0.08% and Canada decreased by 0.06%. Furthermore, for this study, state aggregation was conducted into 11 groups in GTAP version 9A as presented in Table 3.1.

ANALYSIS METHOD. The fault classifications of the individuals are analyzed using the Kappa statistic [Altman91]. This type of statistics is a standard method to evaluate inter-rater reliability. In a software engineering context, it has been used in, for example, process assessment [ElEmam99] and for evaluating ODC [ElEmam98]. It has also been used to evaluate the goodness of a model in comparison with the actual outcome as discussed in for example [Wohlin00]. The agreement in terms of classification can be measured by an agreement index, often referred to as Kappa statistic [Altman91]. Briefly, the Kappa statistic can be explained as follows for the simple case with two raters (or classifiers) and two fault classifications (A or B). Table 2 illustrates this. The cells state the proportions of the faults with a given rating according to faults of Type A and Type B. For example, p11 = 0.20 means that 20% of the faults are considered to be of Type A by both classifiers. The columns and rows are summarized (last column and last row respectively in Table 2), which is indicated with p01, p02 p10 and p20. The agreement index is then defined as κ A E P – P = ------------------- 1 – PE To be able to understand the degree of agreement, the Kappa statistic is usually mapped into a rank order scale describing the strength of agreement. Several such scales exist, although they are by and large minor variations of each other. Three scales are presented in [ElEmam99]. Here the scale suggested by ▇▇▇▇▇▇ [Altman91] is used. It is shown in Table 3. 0.21-0.40 Fair 0.41-0.60 Moderate 0.61-0.80 Good 0.81-1.00 Very good All unique pairs of participants (classifiers) were analyzed. The reason is that it is important to see whether some individuals agreed while others did not. Thus, in total 28 pairs may be generated from the eight participants, i.e. n*(n-1)/2 pairs where n is the number of participants. An eight by eight matrix is created for each pair as shown in Table 4.