Seven years ago, my wife and I thought we were going to be clever and original if we named our daughter "Hannah". We didn't know any Hannahs in our generation. But now it seems every other family had a "Hannah" in the 1990s. After reading Lieberson's book, I understand that we got swept up in one of the many waves of name-fashions that he so ably chronicles.

This is an entertaining book, remarkably so considering the author's intention which was to write a serious work of academic deep-think. So there's a great deal of sociologist talk, which is decidedly not entertaining. But, just as the cartoons redeem even the worst issues of the _New Yorker_ , this book is worth getting just for the many statistical charts. You can follow the spectacular career of "Jennifer", the ups and downs of Biblical names, learn about names and social class, and so on.

Finally, I recommend this book to economists who are interested in fads and herd behavior.

Lieberson has written a scholarly, witty, and extremely informative book about the factors influencing parents' choices of first names for their children. Using data from 7 countries plus the states of Illinois and California, he shows that "fashion" affects children's names just as much as it does choices in clothing or music. Names became objects of fashion several centuries ago in the West, when, among other influences, state regulations and religious customs loosened their hold over what names parents could choose. With the changes concomitant upon nations entering the modern era, name choices subsequently became more matters of individual preference rather than custom and tradition. However, parents made their choices within the context of changing tastes driven by forces "internal" to the naming process itself, rather than being "determined" by external technological or mass media forces.

The sounds of names themselves display explicable trends, such as the preference for names ending in "a" or "n." Groups of names with similar endings rise and fall together, in fairly orderly, long-term movements.

Lieberson does a brilliant job in presenting evidence, using simple graphs and tables, rather than elaborate quantitative statistical analysis. His chapter on trends in name choices among ethnic and racial groups is particularly compelling, as he shows the joint affects of internal mechanisms (e.g. how names "sound") and external influences (e.g. a group's desire to assimilate quickly).

Want to know why your parents named you "Judy" rather than "Judith"? This book has the answer!

A MATTER OF TASTE is a powerful contribution to our understanding of the factors underlying the popularity of first names. Lieberson has brought together a wealth of ideas, concepts, and principles to the analysis of social change. He has used empirical data from the research on names to do this. The data come from several locales including various parts of the United States, England and Wales, Scotland, Denmark and France. Extensive attention paid to the media influence (or lack of) on the popularity of names. For anyone interested in first names this is a valuable background source to understanding their importance.

Lieberson has some great things to say, but he will lose many readers because of his turgid phrasing.

Lieberson's main thesis is that quantitative social research (e.g. sociological studies of racial differences) makes unwarrented conclusions from multiple regression of observed social data. Lieberson criticizes the current researchers because they ignore factors like selectivity, assymetrical causal relationships (if you remove a bullet from someone's chest, it doesn't bring them back to life!) and time lags in social responses.* Lieberson does a good job of showing the validity of his points and of showing how they are often ignored by social scientists who are overly wedded to their statistical computer programs. His comments would be a little more strongly stated if he went on and showed some example of policy missteps based on flaws that he documents. (But I can understand his not wanting to step into a political firefight.)

The book is written in a very academic manner; he refers to himself in third person, overuses nominalizations, etc. I struggled to keep my brain engaged as I read through some of his examples and followed the complicated logic. He could have improved here by using simpler words, using illustrative diagrams, by clearly identifying sidebar discussions and by using summary syntheses. (The Bell Curve is a model of how to do this well--granted Lieberson is writing for a more academic audience, but still some sympathy for the reader might induce Stan to reduce the overly dry "academic tone".)

A couple of methodological questions/issues:

1. When Lieberson describes the dangers of controlling for variables when selectivity is present, he ought to note that this "control danger" exists any time there is a significant unknown variable skewing the data (selectivity or just an "X3" factor. Agreeably selectivity is more likely to be prevelant in social research (and ignored), but in his long, torturous math explanation (show it on a number line, Stan!), he should cite the similar problem of a hidden X3 or a hidden selectivity.

2. Lieberson never clearly addresses what natural scientists or engineers (those who use regression and DOE and read books like Box, Hunter, Hunter) would call an interaction effect:

For instance, Y is a possible function of A, B, C... After testing for each factor individually, you will want to test for the effect of A*B. (above and beyond individual A, B effects. Many of the difficult results that Lieberson describes in social research do not need an "X28" to explain them but can be explained by interaction effects.

3. I wonder if sociologists like Lieberson could learn something from looking at more applied forms of social research (election predictions, actuarial research). Actuarial work for insurance work is a great example to prove Leiberson's point about causal factors acting with a set of random results. And the economic incentive to be "right" may serve to overcome some of the pitfalls that Lieberson documents. I also wonder what business researchers and other practical creatures should "take away" from Lieberson.

* A nice physical analogy for assymetricity is magnetic hysterisis. Technical folks will often talk of a "hysterisis effect" when describing social issues that have assymetic response.