Hi. In this module I'm going to introduce the China Multigenerational Panel Databases. These are databases that are close to my own heart since my collaborators and I began constructing them three decades ago, and have been analyzing them and have most recently made them publicly available at ICPSR, as you'll see later. The China Multigenerational Panel Databases describe rural populations in Northeast China from the 18th century to the beginning of the 20th century. The two databases are the CMGPD-Liaoning and the CMGPD-Shuangcheng. Together they offer 2.8 million observations of approximately 350,000 people. These are longitudinal data that follow individuals, households, and communities over time. The families in these data sets can be linked across multiple generations, as we'll see later, that's especially the case for the CMGPD-Liaoning. The basic contents are similar to what you might find in censuses: so we have age, name, household relationship -- the people were organized by household and for everyone living in a household, their relationship to the head of the household was listed; from that, we know everybody's relationship to each other -- gender, obviously, birth, death, marriage, and migration -- that is the key vital events that we need if we're going to study demographic patterns and demographic trends. We know whether or not males in the households held official positions that has worked for the state bureaucracy in some capacity. We also know which administrative population the people in our data set were affiliated with. These data sets were constructed to track the members of specific hereditary administrative populations in Northeast China, and everyone in the dataset was affiliated with one or another population. Finally, we know the village of residence for the people in these data sets. So, these data, because of their volume, their detail, and their longitudinal depth have a variety of applications. The most basic, which we've actually published a number of papers on are family and community contextual influences on marriage, reproduction, migration, and mortality. So, in the last few decades through studies carried out around the world, we've generally learned a lot about the relationships between people's context and their demographic behavior -- when they married, how many children they had, and when they died. These data sets allow us to do the same thing for at least Northeast China, in particular these rural populations. These data sets also allow us to study household dynamics, including household division. In the past, Chinese households were quite large. In some cases you might have one or two hundred people living together under the same household. Cousins, or even second cousins, all living together with their families. We can study the dynamics of these very complex households and look at the factors that precipitated when they divided into new households. These data also allow us to study social mobility in a non-elite population. Most studies of social mobility in historical China have focused on elite populations, the people who were eligible to take the Imperial Examinations. Here we have an opportunity to examine social mobility in a much more mundane way, that is the ability of farmers to secure mundane or low level positions in the state bureaucracy. We're also able to study the social dynamics and the demographic dynamics of lineages by following families over multiple generations. And finally, these data are amenable to comparative studies where we can compare these data to similar data that have been collected for other countries, other regions, such as Japan, Korea, Taiwan, many places in Europe. Now the strengths of these data are that they are prospective. They follow people forward in time. This makes them very different from most of the traditional sources for studying Chinese population history. Most other studies in the past that looked at Chinese population history and family history relied on family genealogies. These were retrospective in the sense that they were compiled by surviving descendants, families. They had a lot of biases, most particularly, they were only available for demographically successful families because these also registers the CMGPDR prospectives. They were compiled by the state bureaucracy. They followed families forward in time. We actually can find the families that eventually became extinct as well as the ones that flourished -- making this a very distinctive data set for the study of population in China's past. The data also offer a great deal of detail on household context. Genealogies, again a common source for the study of China's population and family in the past, typically did not describe the residential context of the people in the genealogies, just their relationships to each other and in particular their patrilineage. We also get a lot of detail on demographic events, much more detail than is typically available in traditional Chinese family genealogies. Finally, and this is very important, we get very good detail on spousal characteristics. So a lot of traditional family genealogies from China did not record any women, daughters, or wives. In these household registers, the CMGPD, we actually get to track wives from the time that they married into their husband's household until the time that they died, or in a few cases they married out or remarried out after their husband passed away. Finally, these datasets have a great deal of geographic specificity. On the one hand, we know exactly where people lived but they also have geographic diversity. As we'll see in a moment the Liaoning data cover a large region and diverse region within Liaoning. Now these data do have some limitations. Many children who died in infancy or childhood were not recorded at all. So we don't have direct records of births. We only know children once they are old enough, perhaps three or four or five years old, to be recorded by their parents in the registers. Many unmarried daughters are never recorded at all. So even though we have very complete records of wives, that is when men marry their wives are added to their household register and we get to track them forward, quite often unmarried daughters simply are not recorded in the registers. Mortality of the very elderly, that is people over age 75, is not recorded very well. Only official positions are recorded. We don't have other economic activities. If people were engaging in various jobs or investing or doing other things that were not part of the state bureaucracy, we don't have those occupations. And finally, rather technical detail, when we think about events like birth and so forth, we only know their timing to within a one to three year window. We don't actually know the exact dates on which most of these events occur. So each of these two databases have some distinctive features which I'll summarize now starting with the CMGPD-Liaoning. The CMGPD-Liaoning is triennial, that is every three years, from 1749 to 1909. It comprises 1.5 million observations of roughly a quarter million people. It organizes people by administrative population, village, and then household. Linkage of these data across this time has divided those quarter million people into approximately one thousand distinctive patrilineages. These patrilineages can be analyzed as units in their own right to study the persistence of lineage status over time, for example, where the implications of lineage characteristics for the outcomes of members. The other distinguishing feature of the CMGPD-Liaoning is its geographic diversity. The area covered by the villages in this map, the villages indicated by black dots, is roughly similar to that of The Netherlands or New Jersey. Now within that region there is some diversity. So, on the one hand, we have many villages in a agricultural plane or on the provincial capital which is now the city of Shenyang. We also have villages up in the northeast of the province, essentially a hilly, remote, and somewhat poor area. And even until a decade or two ago it was a relatively inaccessible region where people engaged in forestry, hunting, and other activities, with some agriculture. So, very different from the settled farmers living around Shenyang. And then finally, there are many villages down on the coast of Liaoning, in near the city of what is now Yingkou, the area of Gaizhou, many of these were fishing villages or in some cases engaged in commerce or they had orchards, raising fruit and so forth. So the economic and geographic contexts within the CMGDP-Liaoning are quite diverse. One of the unique features of the CMGPD-Liaoning that I'd like to emphasize is the opportunity for multigenerational linkage -- that is the ability to trace ancestry over multiple generations. So almost everybody in the CMGPD-Liaoning, we can find their fathers, so it's close to 100 percent in this line that's indicated by the arrow. Now if we want to go for something a little more exotic, great-great grandfathers, by the end of the 19th century over 80 percent of men born toward the end of the 18th century can be linked back to their great-great grandfather. And then, if we go even further, we want to link men to their great-great-great-great grandfathers, around 50 percent of the men born at the beginning of the 20th century can be linked through the registers to their great-great-great-great grandfathers. So, through this linkage we can then identify people's distant relatives, even relatives that are living in other villages and we can study long term persistence in family status and the implications of being embedded in large kinship networks. Now, we'll talk about some of the distinctive features of the CMGPD-Shuangcheng. The data were compiled originally, or at least they survived for, the years 1865 to 1913 on an annual basis. So we have 1.3 million observations of roughly 108,000 people. The data again are annual so that distinguishes it from the Liaoning population which was triennial data. We also have linked landholding data. Actually, information on landholding is quite rare for historical populations. So this is one of the really important features of Shuangcheng, not just for Chinese data but actually as any historical data set. The CMGPD-Shuangcheng also has registered ethnicity, the population, although they all belong to a Qing Dynasty organization known as the Eight Banners, they all had a registered ethnicity, either Manchu, Han, or one of the other Mongol, etc., so we can do comparisons across ethnic groups within this population. They were recent settlers so they were resettled from elsewhere in China, actually at the beginning of the 19th century, in a deliberate effort by the state to open up this region. These settlers were divided into two types. There were the Metropolitan Bannermen who were recruited from Beijing, the capital, and moved here. They ended up living in the central villages indicated in the circle on this map, this nicely laid out grid of forty villages, eight banners with five villages each. The remainder of the migrants were Rural Bannermen that were recruited from elsewhere in Northeast China, who actually arrived in the area before the Metropolitan Bannermen with the idea that they would prepare the land and facilitate the subsequent in-migration of the Metropolitan Bannermen. The state took a lot of measures to try to preserve the distinctions between these two groups. And, in fact, one of our collaborators -- Shuang Chen at the University of Iowa -- has a book from Stanford University Press, discussing the state's efforts to maintain these privileges. These data are now available at ICPSR to download along with documentation, CMGPD- Liaoning and the CMGPD-Shuangcheng. I encourage you to go and download these data and analyze them. Even if you're not specifically interested in historical China. These are complex data but with good documentation. So it's a good opportunity to learn some techniques for data management which I've been advocating; it's something that's important for you to do regardless of what type of data you may want to work with in the future.
