[MUSIC] Hi, in this module, I'm going to introduce some sources of aggregated data. Places where you can obtain national level indicators relatively straightforwardly, if you want to start experimenting or trying out with your own analysis. Or fortunate in that now a days it's very easy to access aggregated data, national level indicators or even subnational indicators for provinces and regions via the Internet. A few decades ago if anyone wanting to make use of this data, typically have to go to way library and a major research university. Or in a large city in go to the reference section to obtain statistical year books or United Nations reports in publication and so forth. And then transfer that data in to a computer to conduct analysis. Now a days however, it's accessible to anyone with Internet access and indeed you can download it on your phone. Now one of the really important sources of aggregated of data, perhaps your first visit should be to the World Bank website. So the World Bank website has a rich trough of indicators, development indicators on different aspects of economic health, education and other performance that we consider relevant to economic development. A related website is the organization for economic cooperation and development. Which collects statistics of it's own again related to economic development, but also including other aspects like labor, economics, education, and so forth. Again everything is accessible at a website with a user friendly interface. The United Nations has a rich array of demographic indices. Birth rates, fertility rates, migration rates, marriage rates, and so forth. As well as a lot of other development indicators, economic indicators, and so forth related to the ones available at the World Bank or the OECD websites. If you're interested specifically in health, the World Health Organization has extensive data reported to it by member countries on the prevalence of specific diseases. And then other things that we think are important that are related to health. For example, certain kind of health care spending, healthcare infrastructure and so forth. And then for the environment, there is the UNEP, the United Nations Environmental Program. Which is again a rich array of environmental indicators from United Nations member countries. Now of course, if you conduct an analysis, you don't have to restrict yourself to using only one of these sources. Now a days, it's very common for people to download data from multiple sources. Because certain variables maybe only available at one location versus another. So people download data from multiple sources, combine it and then conduct analysis of the resulting aggregated data. If you have a specific interest in mainland China, Hong Kong or Taiwan, you're also very fortunate that now a days much of the key data are accessible again via the web. Not like 20 or 30 years ago where you would have to go to a research library, a reference section to look at actual statistical year books and so forth. For Mainland China,the National Bureau of Statistics should be your starting point. There's a ritury of economic and other indicators for Mainland China available at this website. As well as links to a number of other government agencies in the Mainland with more detailed data on specific topics. If you're interested in Hong Kong the Census and Statistics website says rich detail from the Hong Kong census as well as other data collection efforts by the Hong Kong government. You can even carry out simple tabulations online in your browser of Hong Kong census data at this website. There's also extensive detail on health statistics for Hong Kong at the Hong Kong Health website. And then for Taiwan, the government there has extensive amounts of statistical information available at its website. Again, covering topics like health, economics, education, and again a number of other topics. Now to make use of all of these data, you'll have to learn more advanced methods in a statistics course. I'll mention some of these methods now, but you'll probably need additional training to really learn them properly. The most basic approach is to analyze aggregated data using regression analysis. This has been the main stay of the analysis of aggregated data since the 1950s or even earlier. A lot of pioneering studies that made use of newly available international aggregated data in the 1950s. For example, on the relationships between life expectancy and per capita income make use of basic regression analysis. Over the decades, people working on economic data especially make use of advanced techniques in time series analysis. So when we're looking at economic indicators over the years whether it's unemployment rates, the gross domestic product, other economic indicators. These indicators have unique properties that make it difficult to analyze them solely with basic regression analysis techniques. There are specific techniques we refer to as time series analysis techniques. They are specifically used for indicators that we have on a annual basis or a quarterly basis and which account for the special properties of such data. More recently as people has become interested in establishing cause and effect in looking at relationships among variables in aggregated data. There been a lot of studies that make use of instrumental variables, we mention this in the earlier lectures. Again the idea is to find some other variable that is a source of exogenous variation in what we think is an independent variable. And then use an instrumental variable analysis to isolate the causal effect of variations in a proposed independent variable in shaping variation in some outcome variable which we think is the dependent variable. Finally, many papers now for natural or quasi experiments, policy changes, disasters, other sorts of shocks that allow for, again, some claim about cause and effect. Now of course, if you're going to make use of aggregated data, using all of these techniques which ever technique that you use. You need to be mindful of the potential problems with aggregated data that I talked about at the end of the last module.
