Wednesday, February 16, 2011
Small Worlds & Six Degrees – Duncan J. Watts (1999; 2003)
Watts, Duncan J. 1999. Small worlds : the dynamics of networks between order and randomness. Princeton, N.J.: Princeton University Press.
Duncan Watts, I think, is a kind of rising star in recent interdisciplinary science in a connected age. Especially, in a field of management science, he and his colleagues argued viral marketing which exploits an existing social networks or communication networks (by the way, my personal view on ‘viral marketing’ sounds not positive and creates some unexpected reactions for readers who have no knowledge over epidemiology or network sciences).
Anyway, while it is a hyperbole or not, a so-called network science has lots of implications and gives scholars very valuable theoretical insights.
First, Watts’ book published in 1999 is a little bit hard for ordinary social scientists to follow the arguments and presentation of results. As he told in a book, his original idea about the topic comes from his studies on natural phenomena via theoretical lens of physics. Results are filtered via mathematical formula (in other words, lots of Greeks!!) and lots of scatterplots. However, if readers feel comfortable for the writing styles, they will find that the 1999 book is really beautifully written with great conciseness and details.
Second, the 1999 book is well structured: introduction, explanation, representation of results, interpretations, and summaries. Readers who are very familiar with the writing style of physics can skip some sentences and save their time.
However, I think ordinary social scientists (like me) will be more benefitted from the 2003 book. Basically, the 2003 book is a sort of lecture note or research diary for ordinary readers. Readers get more knowledge over the context and the academic collaboration of the author, which is also very interesting and valuable for baby-students who want to know how scientific collaboration emerges and happens and forms. (Personally, I am really enjoying over the critics of Stanley Milgram’s research about how much overblown the original paper, despite its theoretical importance)
Thus it is better for social scientific readers to start with the 2003 book and to go forward the 1999 book for understanding Watts’s arguments over small-world phenomena.
Tuesday, February 15, 2011
The Drunkard's walk - Leonard Mlodinow (2008)
I believe the author of the book is a great translator, meaning he is really good at ‘interpreting mathematical or statistical theorems at highly abstract level into ordinary English which is natural language most people can understand.’ With the deep and interesting historical anecdotes and records, the author shows that patterns that people found ‘intuitively’ are frequently wrong and misguide people’s decision. The key argument is simple: Randomness matters and it is the key principle that penetrates human activities.
However, actually, it is really hard concept for ordinary people to understand the meaning of randomness because most human beings cannot generate random numbers as machine (any statistical or computer language does with short time) does. It might be really wonder that ordinary humans understand the meaning of randomness via interesting examples in the book while they are not able to make random events.
Suggested readers are undergraduate students and some undergraduate students who have no history of studying formal statistics or mathematical trainings. Of course, social scientific readers who have studied statistics with three or four courses also get huge help and insights from this book. Personally, some of social scientific findings may not be safe from the randomness argument in the book.
Though it is my personal feeling, the book has too many distractions that are interesting by themselves but prevent readers from focusing the main topic, i.e., randomness. Historical records or anecdotes are good and release readers’ burden for the hardness of the topic, but I feel that they are too many and suddenly intervene in the middle of descriptions.
Saturday, January 15, 2011
Comparing media systems - Hallin & Mancini (2004)
Hallin, Daniel C., and Paolo Mancini. 2004. Comparing media systems : three models of media and politics. Cambridge; New York: Cambridge University Press.
As clearly noted in the book, the authors want to make follow-up study of Siebert et al's “Four theories of the press,” one of the classics investigating the relationship between the form of the press and the characteristics of social/political institutions. While Siebert et al's classic book provided very simplified stereotypes of media system due to the influence of cold war, their book was very insightful because media system is not separated from its larger social system. The basic research question of this book is not much different from its theoretical antecedent, while the dimensions and presented three models of media system are more empirically representative.
There are three parts in this book. The first part introduces the concepts and models. Probably the most important thing in the first part is the theoretical definition of four dimensions: (1) the development of a mass press, (2) political parallelism, (3) professionalism, and (4) the role of the government in media system. Based on the degree of a nation-state's score on each dimension, the authors extract three ideal-types (although the authors clearly want to admit internal heterogeneity within each ideal type): (a) Polarized pluralist model, (b) Liberal model, and (c) Democratic corporatist model.
In the second part, the authors describe the histories and social/economic/institutional reasons of nation-states in each idealized model.
In the third part, the authors discuss the underlying similarity between three models, especially recent rampart of Liberal model (e.g., more professionalism, less dependent on political parties or patrons) in terms of differentiation theory.
Very good book, I believe. However, some limitations I felt were: (1) Each model is highly clustered within the spatial boundary. The authors also knew this spatial clustering and concluded the underlying reason is the shared history and mutual influences between coterminous nation-states. I also agree with the authors' conclusion, but this finding, to me, seems to weaken the persuasive power of four dimensions the authors theorized in the first part. In other words, if the spatial clustering is the main force, then the four dimensions are not the explanatory factors to understand the differences between media system, but the explained phenomena caused by shared histories.
(2) Relating to the first point, the book has to emphasize the history (e.g., the invasion of Napoleon into Mediterranean countries), rather than formal social structure or others. In European contexts, it seems okay, I believe. However, it seems highly problematic when the authors' model is applied in Asian or Arab countries. Probably, the prior imperialistic countries wielded higher influence on the form of media system in prior colonized countries. Probably true, but this unilateral influence has to sacrifice or deny the voluntary development in each country.
However, it is still very good book containing many important insights and theories between media and society.
Thursday, January 13, 2011
Comparing policy networks - David Knoke (1996)
Knoke, David. 1996. Comparing policy networks : labor politics in the U.S., Germany, and Japan. Cambridge [England]; New York, NY, USA: Cambridge University Press.
The lead author of the book is very famous for network analysts in social science, especially those whose interests are political networks or social movements. As expected, the main methodology of this book is originated from network analyses investigating the links, types of connections (asymmetric influences, i.e., influencing vs. being influenced), and types of network (e.g., how much dense a nation’s political web about labor politics). Readers should be accustomed to some relation-oriented statistics, such as multi-dimensional scaling, diverse statistics of network properties (e.g., indegree vs. outdegree; betweenness; prominence; mutual connection; density; structural equivalence; or blockmodeling) and some advanced statistics such as structural equation modeling.
Underlying theory of this book is relatively straightforward. A variety of publics (issue public, event public, and others) and classic sociologist’s (G. Homans and T. Parsons) theory of exchange. Probably all sociologists and most of non-sociological social scientists may know what they do mean, although having no specialized expertise.
The subtitle of the book clearly shows that the book adopts comparative study, which aims to compare the similarity and difference between different social/political systems. Three nations – US, Germany, and Japan – were chosen and compared their political webs based on the author’s conceptualization, termed, organizational state approach.
The first chapter is an introduction of what the organizational state approach means, and why it is adopted (simply put, its strength is the possibility of empirical tests). Basically, it means “[Organizational state approach’s] conceptual components argue that modern state-society relationships have increasingly become blurred merging into a mélange of interorganizational influence and power relations”
The second chapter compares three nation-states from the perspective of legal-historical context.
The third chapter introduces the organizations in each nation-state because the basic unit of organizational state approach is a social organization that participates into the policy networks where the collective decision occurs.
The fourth to eighth chapters provided empirical results based on the author’s approach. In chapter four, he examines what is common and what is different in terms of policy interests among three nation-states. In chapter five, he examined policy webs where social agents interact. In chapter six, influences of both action sets and events are examined. In chapter seven, he provides several models about exchange between social agents (information exchange vs. resource exchange that is physical). Mainly this chapter might be considered as the expansion of the prior study conducted by Laumann, Knoke, and Kim (1992). The chapter eight concludes power structure based on results of policy networks. Conclusion – (1) public actors are important in all nation-states, (2) Japan shows centralized, but US shows bipartite with Germany intermediate between two nation-states – is not much surprising but clearly shows general perceptions of each nation-state’s political-institutional structure and/or decision-making culture.
The last chapter summarizes the findings of the author’s series of studies and confirmed his adopted approach (i.e., organizational state approach) is quite plausible. The figure 9.1 (on page 219) may summarize the author’s finding:
(Formal institutions, Informal institutions)
Contentious (USA model) – (Presidential system, Multiple centers of power)
Collaborative (Germany model) – (Parliamentary system, Multiple centers of power)
Coordinated (Japan model) – (Parliamentary system, Single center of power)
Autocratic (Not mentioned) – (Presidential system, Single center of power)
In general, findings are clear, while some figures are not self-evidently clear (probably, due to my unfamiliarity of styles of the author). Findings are not theoretically interesting by themselves, but sound plausible and persuasive. Especially, the author’s descriptions of historical backgrounds greatly helped readers like me.
Tuesday, December 28, 2010
Human Information Processing - Schroder et al. (1967)
Very old book, but contains many provocative ideas and well-organized thoughts. I am informed that this book is famous among information scientists and cognitive psychologists.
The book is composed of three parts. The first part introduces the authors’ theory (or theories). The main point is the relationship between environmental complexity and complexity of human information processing. Of course, there are other hypotheses that are examined, for example, variations between individuals. The main point is that there are inverted U-shaped relationship between environmental stimuli and complexity of human information processing. In other words, there are optimal point that is achieved around the middle range of environmental stimuli. If the stimuli were less than an optimal point, organisms (including people) are not likely to develop complex thoughts, simple rule based decision-making. However, if the stimuli went too further beyond an optimal point, organisms also less inclined to develop complex thoughts. Thus, educational devices (training, according to the authors’ terms) become optimally effective only when the complexity of such devices (i.e., stimuli) is complex enough.
The second part provides a series of empirical findings that are supportive of the theories in the first part. Most of materials are very old, indicating that methods (based on manual content analyses) and testing tools (based on basic statistical analyses) are simple but robust.
The third part carries the code-books and procedures the authors used, in order to construct their measures. Probably researchers whose research emphases are in applied fields would be interested in the third part.
Substantial and technically informative book, if readers are interested in human psychology of decision-making or reasoning. Some of the chapters, especially those dealing with ambivalence of human reasoning in the first three chapters, sound interesting for public opinion researchers
Deep South - Allison, Burleigh, and Gardner (1941/1969)
When I studied in South Korea, I am always hungry for recent English books because they are not relatively easy to obtain there. Old books could be accessed in the university library but there I was seeking for new books since I thought their content might be more advanced than the old ones.
However, my views totally changed after I arrived in the U.S. The most recent book (sometime books published in 2011, while I am still living in 2010) can be read, but most of the trials was not always enjoyable. New books are necessary to update relevant literature, and poke at the recent topics and arguments that are under debate. However, the topics new books dealt with are sometimes tedious or sometimes too technical or sometimes sound like tautology (Although not always).
Instead, I recently feel the beauty of old books and the cogency of their accents. Deep South is a sort of beautiful books (Similar approach was preceded by William DuBois’s The Philadelphia Negro in 1899).
The first few chapters start with interviews with white and black informants who live in deep south (termed Old City, and rural places). Most of stories are very tangible and clearly show the relationships between whites and blacks. While some of the terminologies are a little bit awkward from the perspectives of recent days. For example, ‘caste’, one of the main terms in the book, sounds very unfamiliar to recent American residents because the term might not seem to fit with pluralistic and democratic regimes any more. Probably recent sociologists or political scientists (while I am not sure about anthropologists) adopt racial relationship or symbolic racism or racial dominance, rather than caste. However, the strongly accented term might be better to directly point out the social structure that are intermingled with racial caste and economical class with more clarity. Recent terms portraying racial relationships might sound politically correct, but they seem to hold less power to clarify the seriousness of racial problems in the country.
Probably the chapter 7 would be the most widely known chapter to social network analysts. Basically, the authors adopted two-mode network which is comprised of two measures, one is persons, and the other is social events (and their participation). Based on the presence/absence of a person in an event, researchers transform the two-mode network into one-mode one (i.e., two persons can be assumed to be connected, if they simultaneously participated into one same event). However, the evidence does not demand much knowledge over network-related statistics. Anyway, simple statistics clearly support the notion of social cliques that are the bricks of social structure that are hierarchically stratified.
Other chapters provide more participatory observations relating to the social structure of deep south that is divided into (1) white-black, (2) poor-rich, and (3) rural-urban.
While reading the book, two things might be noted because the authors were also influenced by their time. First, black English was all carried with non-standard English, for example, expression of ‘colored’ is written as ‘cullud.’ Even if poor whites’ English were written with grammatically correct form, but wealthy blacks’ English were written as uncorrected form. It reminds me one of readings that showed Japanese English that was written as Native Americans heard. Second, ‘nigger’ was too frequently used. Probably this word MUST be avoided in recent time. Two things might tell readers how much American society changes and be careful when discussing racial issues.
Anyway, it is a readable book that I enjoyed.
Monday, December 13, 2010
Reasoning and Choice - Sniderman, Bordy, and Tetlock (1991)
Some books are repeatedly cited with many reasons. Some are highlighted because of their initiation of certain research topics. Some are highlighted because of novelty of their arguments. Some are focused because their findings are well summarized and organized, serving like a final report in discussed topic. Whatever the category they belong to, books with many citations deserve to read carefully, I believe.
This book, I think, seems the third category. Findings and critics (meaning the accurate or correct judgment of an external world phenomenon) are well summarized and how they sort them out in order to fit their arguments. Especially, the first chapter is well constructed what topics the authors want to deal with, and what suggestions for future research could transform the discipline into the better.
According to the authors, there are six themes with which they want to deal in their book: “the revolt against minimalism, the concept of consistency, the role of feelings as well as beliefs in political reasoning, the “heterogeneity” assumption, the role of education in democratic citizenship, and an account of .. dynamics of reasoning and choice” (pp.1-2). Of course, the six themes are well-interconnected with each other.
Arguments, I believe, do not wait for my summary. Here (because this writing is for myself, not for other readers) I want to note one dissatisfaction over their findings and conclusion.
While many paragraphs deal with two psychological mechanisms (i.e., differentiation and integration), I think their definitions are not so harmonious with their uses of terms. I believe that these seemingly unfitted use of terminology does not belong to the authors’ faults, but conventions of opinion surveys or quantitative measures. First, there are many places where we can observe ‘idea-elements’ following Phillip E. Converse (1964), but there are actually no ‘ideas’ in measurements. Ideas are imposed by researchers or survey designers with the form of preference of certain issues (e.g., affirmative action) or arguments (e.g., people with AIDS should be quarantined) or principles (liberalism-conservatism). However, ideas, in normal discourse, denote some thoughts or cognitions, rather than preferences. Also ideas usually imply that thoughts or cognitions are created or voluntary, rather than given.
Second, thus belief system seems like judgmental structure of preference, rather than a system of knowledge or others.
Third, thus the mentioned belief system has to be measured as the strength between preference A and preference B, rather than a system of key ideas or thoughts or cognitions.
Fourth, opinion polls only ask respondents to assign themselves on a given set of locations.
Thus differentiation is less likely to be measured quantitatively because the dimensions that are obtained are previously achieved by a researcher, not by the survey respondent. While integration is possible to be measured, it has to be mixed with random guessing when respondents’ a web of belief is not sufficient enough.
However, as I’ve already pointed out, this book is book and sophisticated enough. A worthwhile book for reading and also holding for later consultation (although some chapters are out-of-dated because of the lead author’s later publication, such as “Scar of race” or “Reaching beyond race”)