**3. Data description and methodology**

The data set, which is chosen to be analysed, is the appropriate source to examine the desired aims. Also, the selected methods help to reach the outlined result.

#### **3.1. Data description**

The source for the most part of the data is the Datastream maintained by Thomson Reuters Corporation. Also, calculations such as price-to-earnings ratio are done by the Datastream and therefore provide a universal basis across the different regions under consideration. Equity indices and ratio calculations for the Deutscher Aktien Index 30 and the China Securities Index 300 were partially extracted from the Bloomberg database since they could not be obtained from the Datastream. December 2015 has been chosen as an end mark of for the underlying data.

#### **3.2. Methodology**

The methodology mainly focuses on alternative approaches while still incorporating some impulses from classical economic theory and the efficient market hypothesis. This is in line with many other research papers which are also based on the macroeconomic environment, alternative measures and the expectations of private and institutional investors [43, 44].

Since markets are interconnected as for example confirmed by the study [45], a holistic perspective for comparison is employed. In this regard the industry under consideration must be specified. For this paper, it is the technology sector which is analysed and even though most of the researchers agree that the dot-com bubble occurred in that sector, others have also identified spillover effects into other sectors such as financial, general industrial and non-cyclical services as well [31].

To tackle the research question, first, a combination of macro-economic indicators was used to compare the dot-com bubble with the recent financial crisis. On the one hand, a global perspective is pursued, but on the other hand, data on an aggregate level was not always available. Therefore, the United States, Germany and China are chosen by their economic dominance and highest nominal gross domestic product as representatives in the named regions to base the local economic indicators on. This way a top-down complementary approach from broad regional equity indices to country specific indicators and indices was employed throughout the paper. The term regional in this regard refers to the regions of North America, Europe and Asia, while local refers to the United States, Germany and China.

Additionally, but due to the limited scope of the paper and the main focus being set on the overall economic situation, a rather simple but meaningful test was employed, the Welch-Satterthwaite t-test. This test in the context of this paper is focused on the comparison of means of returns of two samples. With respect to the regular t-test, the Welch-Satterthwaite t-test gives the advantage of yielding accurate results despite unequal variances of the samples. It is, however, important to strictly differentiate the time-period, index and sector under review.

So after an overall evaluation of the economic situation, a distinction is done by the comparison of means of returns between the Nasdaq Composite index and the Datastream technology index in the 1990s of the twentieth century for three phases of the bubble – pre-bubble, inflation and crash. For this purpose, assumptions about the timing of the phases had to be made. The pre-bubble period was chosen to start in December 1990, which marked the lowest point of the Nasdaq Composite index in the nineties years of the twentieth century. The bubble inflation period was set to start in November 1998 since this year is based on the academic research [16], while also marking a relative historic low of the index. The peak obviously occurred in March 2000 and marked the transition to the crash-phase. Here the lowest point before the eventual recovery was chosen – October 2002. For the recent years such a clear distinction was not possible and therefore a time-frame as a whole from April 2009 to December 2015 was chosen, marking a historic low and high, respectively.

Since the movement of the indices alone cannot support nor reject a bubble, a dedicated section takes into account key performance indicators based on which a fundamental value can be estimated.

Due to the limited scope of the paper, the influence of business cycles, exchange rates and non-financial factors such as page-views have been of minor coverage in this paper, and hence provide room for further analysis.
