European Sociological Review, Vol. 5 No. 1, May 1989 © Oxford University Press 1989 81 Searching for ideal types: the potentialities of latent class analysis JACQUES A. HAGENAARS AND LOEK C. HALMAN ABSTRACT Typologies play an important role in sociological theory and research. Basic to the use of (ideal) types is the notion that a subject's overt behavior can be conceived of as governed by his/her belonging or closeness to a particular underlying pure type. Many statistical techniques are in use to detect or construct these fundamental types, especially factor analysis. Much less attention has been paid to the possibilities that latent class analysis has to offer. Through an elaborate example, it is shown that the basic ideas of latent class analysis correspond eminently well with the use social scientists make of (ideal) types. Several important extensions of the basic latent class model along with significant new developments are discussed. INTRODUCTION Typologies, whether or not in the form of ideal types, play an important role in sociological theory and research. An outstanding example is Max Weber's 'ideal' ethos of Protestantism and the Spirit of Capitalism, or his Reiner Typen legitimer Herrschaft. Besides Max Weber, many other social scientists from the past and the present have created typologies which have had a great influence on the directions sociological theory and research have taken. To mention just a few: Tbnnies' dichotomy Gemeinschaft- Gesellschaft; Riesman's typification of societies in terms of tradition-, inner-, and other-direc- tedness; Redfield's folk and urban society; Almond and Verba's typology of political cul- tures; Bell's characterization of societies as pre- industrial, industrial and post-industrial; and, finally, Inglehart's division of people in materi- alists and post-materialists. (For a survey and a still longer list, see McKinney, 1966, chapter 5; see also, Bell, 1974; Inglehart, 1977.) The categories of a typology, the types, can be conceived of as the cells of a multi-dimensional table whose dimensions are formed by the rel- evant attributes which make up the typology. Or, in other words: types are just locations in a so-called property space (Barton, 1955, p. 50; Sjoberg and Nett, 1968, p. 254; Riley, 1963, p. 343). Usually, a typology occupies only part of the property space and consists of only a few cells of the multidimensional table. Very often these are the extreme cells or locations. But extremity as such is not important. What does matter is that the particular combination of characteristics a certain cell or location repre- sents, makes sense and can be interpreted mean- ingfully. By focusing solely upon certain 'pure' types and thus exaggerating certain specific features of the population under study, one tries to get a better and deeper understanding of the complexities of social reality: the 'overt' behavior of people is explained by, and is made intelligible from, their closeness to these pure, fundamental types. In this sense, people are assigned to particular types. There are several statistical techniques that can be very helpful for discovering a suitable typology and for classifying cases in terms of this typology. Well-known and widely used tech- niques are factor analysis, discriminant analysis and cluster analysis. Much less attention has been paid to latent class analysis. Although the basic ideas underly- ing latent class analysis were developed as early as 1950 (Lazarsfeld, 1950a,b), it is only rather recently that serious (statistical) shortcomings of the older procedures for carrying out latent class analyses have been overcome (Clogg, 1981; Downloaded from https://academic.oup.com/esr/article/5/1/81/517546 by guest on 10 October 2021