Search

The article presents a research response to the methodological criticism issued by Tom Van der Meer, Manfred Te Grotenhuis, and Ben Pelzer included within the issue, challenging the case examples provided in the author's sociological analysis of religiosity and volunteering behavior in the United States. The authors disagree with the conclusions made against their original study but also compliment the methodology suggestions raised by the article and provide alternative paradigms for future studies to correct for the problems.

The article presents comment on the work of sociologists H.F. Franke and J.D. Kaul related to an alternative statistical interpretation. Franke and Kaul show that well over 90% of the variation in the average productivity of the workers involved can be accounted for by just three explanatory variables: two dummies of their own construction, which they interpret as managerial discipline and economic depression; and one experimental variable, scheduled rest time. They conclude that the lack of substantial unexplained variance indicates that the unmeasured supervisory and social interaction variables were not very important economically. Despite uncertainty about the exact form of the relation between productivity and time, and despite very great uncertainty about the magnitudes of the effects of some of the other measured variables, there are two qualitative conclusions concerning the first Relay experiment. The data agree very much better with models, which include a smooth function of time as an explanatory variable than they do with models, which do not.

Path analysis requires that exogenous and residual variables not be correlated. An alternative to assuming the two to be independent is to treat the exogenous variables as fixed or non-random. Examination of pat/i models, however, reveals that the usual standardized path coefficients arc biased and inconsistent when exogenous variables are considered fixed. At the same time, the path coefficient divided by the residual term is a large-sample unbiased estimate of the true coefficient. Variants of this basic transformation are discussed and an empirical example is included. [ABSTRACT FROM AUTHOR]

Controversies over the value of statistical significance tests in assessing substantive significance in basic sociological research have persisted because the methodological issues involved have not been clarified. The issues are (1) the meanings or definitions assigned to the terms statistical ant substantive significance ; (2) the nature of relations between the definitions, and (3) the grounds for choosing definitions and their relations. An argument is made for : (a) defining subs'antive and statistical significance independently of each other and (b) forming methodological propositions relating these definitions in which statistical significance tests play no role in assessing substantive significance. The concept of methodological significance is introduced as a basis for choosing between alternative definitions and methodological propositions. [ABSTRACT FROM AUTHOR]

It has frequently been assumed that American women tend to marry men whose socio-economic origins are higher than their own. One rationale for this assumption is the notion that the woman's romantic orientation is more adaptive and directive than that of the man. A review of the evidence reveals that the assumption has not been adequately tested, however. The present study approached the question by applying several statistical techniques to large-sample data on husbands' and wives' fathers' occupations. No overall tendency for American women to marry up was found. If only the two highest occupational categories are considered, however, women apparently do tend to marry up. In addition, farmers' daughters are more likely than farmers' sons to marry the children of white-collar workers. [ABSTRACT FROM AUTHOR]

The matching of experimental and control subjects on critical variables, particularly when used in conjunction with random assignment, is widely recognized as a valuable element in research design. Both in discussions of its usefulness and in applications, matching tends to be treated as an attribute. In fact, however, it is a variable, with a range determined by the number and precision of measurement of the items taken into account in the matching process, and by the size and homogeneity of the pool from which subjects are drawn. Failure to distinguish between well-matched and poorly-matched pairs can lead to significant errors of interpretation. But treating the quality of the match itself as a variable, and introducing it into the interpretation can, it is suggested, contribute to the measurement of the effects of the experimental stimulus. An "index of congruence," designed to measure the quality of matching in an experiment in progress, is described. We propose to use this new method as a complement to established statistical techniques, whose comparative advantages and disadvantages are here briefly discussed. [ABSTRACT FROM AUTHOR]

Many theoretical questions investigated by sociologists hypothesize that one or more independent variables produce certain effects on a dependent variable. For substantive theories with this kind of structure, an explicit translation into mathematical model form is provided by Coleman's work with continuous-time, discrete-space stochastic processes. The statistical technique most appropriate for estimating the parameters of such a model is exactly the same as multiple regression analysis of dummy variables. Aside from the power which multiple regression brings to multivariate analysis, the major advantage of this convergence is the explicit empirical evaluation it provides for the structure of a substantive theory. [ABSTRACT FROM AUTHOR]

Two distinct criteria are discussed for evaluating the relative importance of several independent variables in determining the variation in a dependent variable. The quantitative criterion is used primarily with numerical data, whereas the causal criterion often appears in theoretical arguments. Simon's method for making causal inferences from correlational data may offer potentialities for combining these criteria. The indiscriminate use of partial correlations and a single multiple regression equation can yield misleading conclusions in evaluating importance. instead, an entire set of simultaneous equations is needed. [ABSTRACT FROM AUTHOR]

The article discusses significance of nationwide marriage and divorce statistics for the United States. One of the recommendations made by the Committee on Marriage and Divorce Statistics, and subsequently approved by the Publications Committee, for the year 1955-56, has been that an article be prepared for setting forth the issues involved in the establishment of a Marriage Registration and a Divorce Registralion Area, and the sociological benefits to be derived from nationwide marriage and divorce statistics. This presentation will not only contribute to a better understanding of the status of vital statistics, but also enable sociologists to assess their own role in the development of marriage and divorce registration in the U.S. In academics also these statistics will be beneficial as well. Sociological needs relative to the collection, analysis and publication of marriage and divorce statistics has been discussed in the article. But a question remains to be answered that what should sociologists, both individually and collectively, do about furthering the program in this field.

This article analyzes the tests of significance in sociological survey researches. Differences or correlations are now routinely tested to ensure the statistical soundness of sociological inferences. The popularity of these tests has even reached the point where their absence occasions criticism. Empirical researches generally fall into two groups, those which test each conclusion for significance but seldom cross-tabulate extensively to discern causal or explanatory factors and those in which substantive analyses are based on extensive cross-tabulations, with no tests of significance. The conditions under which tests of significance may validly be used are almost impossible of fulfillment in sociological research, save perhaps in one or two limited areas, and even when these conditions are met, the nature of the research situations faced by sociologists is such that correct inferences from the tests are equally difficult to reach. It seems useful, therefore, to consider tests of significance in sociological research more generally, to inquire into the conditions under which they may properly be used and under which they may properly be used and the kinds of inferences that may justifiable be drawn.

This article focuses on family life cycle as an independent demographic variable. Demographic variables, all too frequently, are used as independent variables in a mechanical way, without theory to guide in the selection of the variables appropriate to the study of the dependent variable under consideration, and without theory to guide the definitions of the variables themselves. This is primarily due to the fact that data used in demographic research were collected primarily for other purposes. To pay close attention to the meaning of the variables, it is important in using demographic variables as predictors of behavior. Moreover, age of person or of the head of a family is a traditional variable. It has several advantages such as uniformity of measurement, relatively easy and uniform acceptance and the fact that it is a continuous metric scale. The advantage of age as a metric scale may be considerable when one works out formal correlations. But this advantage of age is not obtained when the results are analyzed and presented only in tables showing age in class intervals. This article proposes the family life cycle as an alternative variable and shows some of its advantages, particularly in its greater explanatory power illustrated with several dependent variables.

This article focuses on testing the probability models for laws of social interaction and the predictability of diffusion. Message diffusion is but a one-way person-to-person form of interaction. Message diffusion is, furthermore, an observable and experimentable sub-form of social or cultural diffusion in general. Interactance formula, or general model, for any single kind of group behavior-in-context analyzes human interacting into six dimensions or classes of factors. The simple group behavior of spreading an attribute has been studied as an example of a unified field of social forces. This field was factored into the six general dimensions of the behavior of people in time and space when internally stimulated or "motivated" and externally stimulated in specified ways. A prime methodological interest of these researches is in the homogeneity of these six classes of factors or "sectors" in the human sciences. These factors have been found to be both influential and instrumental in determining various aspects of social interaction.

The argument of this article is that the use of mathematical models for predicting social phenomena requires matching the social and mathematical conditions and that all-or-none elements often help this matching as illustrated in the particular cases of the normal probability distribution and the logistic growth curve. It is assumed that sociologists, being scientists, are trying to predict social phenomena better. It is noted that they are using mathematical models increasingly for such predicting. For a mathematical model to be used as a rational social hypothesis requires both that the model fit the data closely and reliably and that the mathematical conditions match the social conditions. Thus, in the case of the normal curve and the logistic S-shaped growth curve, the mathematical derivation from all-or-none elements can be matched by suitable combining of present-or-absent social variables. Knowing this structure of binary elements enables the social engineer, for example, to convert normally distributed phenomena into logistically distributed phenomena. This suggests that binary elements may become useful tools for changing some social distributions.

The ascription of low status to Social Theory is a passing phase in the history of sociology. When the change in the status of Social Theory occurs, it will not mean that Social Theory will return to the position it occupied in the nineteenth century when it held a virtual monopoly on sociological interest. Social Theory will be cultivated as one of the major interests of sociology closely integrated with other fields. It will not mean that every aspiring sociologist will have to try his hand at it. Each generation cannot be expected to produce more than a limited number of creative minds who possess the special gifts required for the construction of general theory. It will mean, however, that each generation will provide opportunities for creative work in Social Theory. It will seek to maintain Social Theory as a going concern by attracting and cultivating potential candidates for this task. This is not the case today when the prevailing emphasis on rapid specialization and the pressure of the demand for technically trained personnel discourages brilliant and creative minds from joining the profession on the one hand and on the other, diverts talented students by giving them neither time nor encouragement to test their abilities for the construction of general theories.

This article focuses on contributions of sociometry to research methodology in sociology. According to sociometry, society systems are preference or attraction-repulsion systems. This is claimed to be true not only of human, but also of sub-human societies. It also claimed that human preferential systems cannot be examined adequately by the old methods of fact-finding objectivity as statistical methods and observational methods. The method themselves and instruments derived from them have to undergo a process of subjectification in order to return to the researcher endowed with a more profound objectivity, having gained a grasp of the social processes on the depth level. This new sociometric objectivity can well be contrasted with the old positivistic objectivity. Sociology owes to sociometry a few genuine generalizations among which are most prominent the law of social gravitation or mobility, the sociogenetic law, the sociodynamic law, the reality test of social configurations, the law of the social atom and the tele phenomenon.

The article presents comments of the author on the article "The Logical Adequacy of Homans' Social Theory," by Ronald Maris. Maris attempts to formalize some theses of scholar George C. Homans' in order to permit deduction of consequences thought to be of pragmatic interest. However, his methods of formalization bear little relationship to the normal roles played by the symbols used, and so he fails to realize that his deductive procedures, although standard for the symbols, are invalid for the uses he makes of them. The outcome is that he must choose between following the notation or rejecting most of his derivations. Maris offers as a derived but apparently self-contradictory statement. He then suggests that it is not really self-contradictory because "it is a logical consequence of the satiation principle the more rewarding an activity in the past, the less rewarding is future activity. The difficulty that lies in Maris' strategy of overcoming logical gaps with empirical derived rules of inference such as Conversion. Once one finds that the sense of one's symbols cannot yield certain conclusions, the next move is to make explicit what further principles are needed to obtain such results and to count them, if they are not themselves logically true, as empirical premises.

The article presents comments of the author on the article "The Logical Adequacy of Homans' Social Theory," by Ronald Maris. Partisans of statistical and mathematical approaches to sociological theory like to regard their writings as major "advances" on the frontiers of scientific sociology. Lest Maris' article be construed as an advance in this or any other realm, some comments should be made on his formal demonstration of the logical adequacy of scholar George C. Homans' theory of interaction. Maris' use of the expression "logical adequacy" in the title of his article is perhaps misleading. He states in a footnote that, strictly speaking, his derivations should be called "logic-like propositions." It is never made clear, however, why the article was not, therefore, entitled "The Logic-Like Adequacy of Homans' Social Theory." The use of "logic-like propositions" in a demonstration of logical adequacy would seem to be analogous to utilizing "mathematic-like equations" in an examination of the mathematical adequacy of some theory of mathematical sociology. The discussion of "mathematic-like equations" or "logic-like propositions" leads us into immense problems of language. For mathematics, it seems clear that outside of a formal system, the notion of "equation" makes little sense.

The article discusses oblique references. The statement by researchers that delinquency is fundamentally related only to anomie and not to poor socio-economic conditions is interpretable only if it is recognized that anomie and socio-economic conditions are fundamentally related. This means that the hypothesis that relationship is statistical, is misleading. It is possible to agree with the conclusion that two (at least) factors are derived from the data and still not accept the apparent rigidity of the research interpretation. The distribution of the items on the two factors is within a restricted range, the seven relevant ones being within approximately 62 degrees. Further, the items are distributed even though there are two apparent clusters. This does not suggest the discreteness of interpretation of the items suggested in the research and supported by various techniques. In fact, The basic issue of clarity of structure involves consideration of the sampling of variables and not merely the analysis of these particular data.

This article focuses on a statistical method of graph associational distances between occupations, devised by sociologist Edward Laumann and Louis Guttman. However, they do not seem to be aware of the very serious but easily correctable misuse of the method, which lies in the relative sizes of occupational categories. The method of smallest space analysis used by Laumann and Guttman graphs, in the fewest number of dimensions with as few errors as possible, is a set of associational distances between occupations, where members of occupations which are "close" to each other are likely to associate with each other and members of occupations which are "distant" are unlikely to associate with each other. Once one has the positions of occupations in the space one can interpret clusters of occupations or the dimensions of the space. This is similar in some ways to factor analysis, which might be looked at as graphing the relationships between variables expressed in a correlation matrix. Moreover, the difficulty with the application of the method by Laumann and Guttman is that they do not take into account the effects of sizes of occupational categories. If an occupation is numerically large in the population, either because it really is large or because many occupations have been grouped into one category, many people will have important alters in that category just as a consequence of that category's size.

This article analyzes the effect of cumulative scale analysis on the prediction of life processes. The cumulative scale is a recently evolved technique for describing and measuring patterned human behavior and is used primarily to ascertain the patterning of attitudes along a single dimension. Cumulative scaling techniques and sociometric data have also been used to distinguish patterns of response toward objects and patterns of response from subjects. However, the study of process, regardless of the technique used, consists of relating a patterned series of events to a time order which, by definition, is scalable. If a scalable series of social or psychological events could be related in statistical language to the time dimension, a model of patterned progressive behavior could be evolved. Such a model would involve the correlation of two cumulative scales measuring discrete phenomena. In the literature of sociology many processes have been hypothesized but have not been stated in a testable manner. If a process is involved, the sequence of events should prove to be scalable and the scale should be highly correlated with a time scale. On the whole, sociologists have been little concerned with time as a variable except for comparisons before and after an event, or at two points in time.

In order to get some notions about the nature of the data and the kinds of statistical analysis that do and do not occur in sociological research, an examination has been made of all articles published in the periodical "American Sociological Review," from 1944 through 1953. Roughly 48 per cent of these articles reported data that conceivably could be subjected to some sort of statistical analysis. This classification uses the broadest possible definition of statistical analysis, assuming that any ordering of statistical data, qualitative or quantitative, represents statistical analysis. Included in the 48 per cent are some reports that present incomplete data for purposes of possible statistical analysis but from which it can be inferred that the investigator has more complete data, which he did not present. The other 52 per cent of articles were discussions or studies whose data could not be subjected to any sort of statistical analysis. Included in the 52 per cent are some articles that refer to statistical data and results of statistical analysis in other reports but indicate no statistical manipulation by the writer.