Annexes values for the dataset are decided such that anything below the minimum is assigned 0, anything above the maximum is assigned 1, and everything else is scaled evenly between 0 and 1. Depending on the nature of the data, the banding process can take different forms. For example, in the ‘Secondary school enrolment’ indicator, a higher score reflects a more desirable situation. Therefore, in this case the banding process requires the largest data point to be assigned a value of 1. Conversely the lowest data point in the indicator is assigned 0, while all other data are scaled relative to these two points. This process is referred to as forward banding. On the other hand, in the indicator ‘Youth mortality’, a lower score reflects a more desirable situation. In this case, the data are reverse banded, so the lowest value is assigned 1 in the banding process, while the highest is assigned 0. Therefore, in year y, a forward banded score is calculated for indicator i by Equation 1. A reverse banded score is calculated by Equation 2. Equation 1. Banding equation Bandedi = Country indicator value in year y i − Mininum cutoffi Maximum cutoffi − Mininum cutoffi Equation 2: Reverse banding equation Reverse Bandedi = Country indicator value in year y i − Mininum cutoffi 1− Maximum cutoffi − Mininum cutoffi An integral part of this process is to set appropriate minimum and maximum cutoff values for the banded scores; empirical and normative methods are available to do this. While some data may be distributed normally and therefore lend themselves well to standard and well-defined mathematical techniques, such as defining outliers as those greater than three standard deviations from the mean, other datasets do not follow well-behaved trends. The final choice of which technique to use must depend on a number of considerations: the nature of the data, the underlying distribution, the purpose of the index, what information is being conveyed, and so on. When investigating the global datasets used in the YDI, very few of the distributions can be classified as normal. The presence of outliers affects not only the average, but also the variance, skewing both the minimum and maximum values. To account for this, the Institute for Economics and Peace in some instances set artificial minimums and maximums to ensure results are not too heavily influenced by outliers. In the cases where outliers are present, the lower bound set for the banding process are set as the lowest data point that is within 1.5 times the inter-quartile range below the first quartile (where the inter-quartile range is defined as the distance between the first and third quartiles). Similarly, the upper bound set for the banding process are set as the largest data point that is within 1.5 times the inter-quartile range above the third quartile. The following section details where this process has been used and illustrates the application of the process to the distributions of the indicators. The bands obtained by this process are shown in Table A2.7. Outlier results have been trimmed. A2.5  Weighting indicators and domains Table A2.8 shows the indicators and weights used in the YDI. The YDI assigns equal weight to the domains of Education, Health, Employment, and Participation (civic plus political participation). Within domains, each indicator is weighted by its relative importance to the other indicators in the domain. There are a number of methods available5 to decision-makers, including data envelopment analysis, benefit of the doubt approach and unobserved components. The YDI uses a set of three primary indicators to align overall scores with broader human development priorities: youth literacy rate, youth mortality rate and youth NEET rate. Primary indicators are weighted more heavily than others in the index. It is important to note that these primary indicators have a major impact on domain scores, in some cases giving countries a more pronounced domain score regardless of their overall rank in the YDI. A realignment of weightings to better represent the current development issues has resulted in rank changes for many countries in this iteration of the YDI. The only time series comparison recommended by IEP is to use this iteration across the years provided. \ 143