Experimental Designs in Sociological Research

Email

Whatsapp

Facebook

Pinterest

Twitter

Copy link

Experimental design in sociological research involves systematically manipulating variables to examine cause-and-effect relationships within social contexts. It typically includes controlled settings, where participants are divided into experimental and control groups to isolate the effects of a specific variable. This method allows researchers to test hypotheses with a high degree of reliability and internal validity.

In sociology, experiments can be laboratory-based, such as studies on group behavior or conformity, or field experiments, which are conducted in real-life settings like schools or workplaces. Key elements include random assignment, control of extraneous variables, and measurement of outcomes.

Despite its strengths in establishing causal links, experimental design in sociology faces limitations, such as ethical concerns, artificiality of settings, and difficulties in generalizing results to broader populations. Nevertheless, when carefully implemented, experimental methods provide valuable insights into social processes, attitudes, and behavior, complementing other qualitative and observational research approaches.

THIS Revised Edition is primarily concerned with tests of experimental designs in the free community situation. To this end four new chapters have been added to the original seven: Chapter VIII, ""Analysis of Variance and the t-Statistic: Underlying Assumptions""; Chapter IX, ""Non-Parametric or Distribution-Free Statistical Methods""; Chapter X, ""The Ex Post Facto Design: Replication and Extension""; and Chapter XI, ""Some Problems in Psycho-Social Measurement.""

The advantages of the non-parametric approach over the usual parametric methods appear to be: (1) their underlying assumptions are less numerous; (2) they are less time-consuming in application; and (3) they yield probability estimates which are usually more conservative tests than the usual parametric procedures. But they have the disability of errors of the second kind, i.e. they may lead to acceptance of the null hypothesis, when in fact some alternative hypothesis is true.

Some suggestions are made toward a resolution of the difficult problem of randomization in the free community situation in Chapter VIII. Probability tests are now available for non-random small samples as well as for random samples by the Haberman method outlined in Chapter IX, with probability table in Appendix D. Re-cent examples and extensions of the ex post facto procedures are presented and examined in Chapter X. Some problems of psycho-social measurement and the possibilities of resolving these appear in Chapter XI.

This Revised Edition is in no sense a treatment of these subjects in the tradition of mathematical statistics; nor does it attempt to offer a thorough account of the newer methods, since extended treatment including proof will be found in accessible texts on statistical method to which the student is referred. My effort has been to bring to the attention of the empirical research worker and the serious student certain recent developments in method that offer promising leads at specific points of research designs to applications and to adaptations of method which may lead to the resolution of the practical and theoretical problems involved.

Appendix C attempts to explain my position relative to certain comments on methodology and terminology which appeared in some reviews of the first edition of this book.

Acknowledgements are due the editors of The American Socio-logical Review, Social Forces, The Journal of the American Statistical Association, and Science, for permission to reproduce materials from my articles in these journals. A special word of appreciation is due Sol Haberman of the University of Minnesota for permission to reproduce herein, before publication elsewhere, his own contribution of a non-parametric method based on random small groups and peculiarly adapted to use in testing the results of empirical ex-perimental designs applied in the free community situation. This method of partially ordered systems of expectations is described at the end of Chapter IX. His application of the method to non-random small groups appears in Biometrika, vol. 42, in 1955, and in his doctoral dissertation, Special Distributions of Kendall's Tau (+), University of Minnesota, 1955.

AMONG the problems of human relations that concern us today and demand attention there are a few which have been studied by experimental methods, not in the laboratory or yet in the formal class-room situation, but right out in the natural community. What can we learn from these experiments about the possibility of an affirmative answer to such pressing questions as: Can juvenile delinquency be treated to diminish it, or to prevent its recurrence? Does healthful housing in public projects lead to improved individual and social adjustment of the members of low-income families? Is the morale of needy persons aided by a work program superior to the morale of those who receive only poor relief? Can successful adjustment to our increasingly complex environment be achieved by formal school education or student counseling programs?

The experimental studies described and analyzed in this book seek the beginnings of answers to these questions. Since each study is a single experiment, as yet not repeated, the results do not reveal any dramatic or spectacular achievements any more than a single experiment in physical science solves a problem. Nor can we generalize to the community as a whole from the results of any one of these experiments. But it does seem evident that the results are sufficiently favorable and objective to hold forth the promise of at least subsequent success, if only sufficient determination and patience can be exercised in repeating these experiments until we discover the way to verification. The needed next steps are to repeat them in the study of the same type of problem elsewhere but under like conditions. It is with the hope that this will be done that the present book has been written. As illustrated in these studies, the newly developed research methods of experimental design, applied in the community situation, seem to point the way by which we can overcome, in the course of time, some of the complexities of social interaction that have hitherto baffled rational and objective description of certain acute problems of human relations. At the same time that this optimistic note is expressed, a word of warning should also be recorded, namely, that progress with this method of experimental design may be slow and will undoubtedly require a considerable span of time and numbers of repetitions of an experiment before the stage of reliable generalization and dependable prediction can be attained.

Programs to prevent the recurrence of social ills always rest upon the assumption that social cause and effect sequences are known, whether or not this assumption is implicit or explicitly stated in the program. More research in this area is needed to implement social planning and social action programs. The use of the method of experimental design to discover and to expose the causal complexes in the social situation of a given period has just begun to claim the attention of research workers in human relations.

My own interest in the possibility of an adaptation of the experimental method of physical science to the study of the problems of human relationship dates from 1916, and my first published papers on this subject appeared in the Scientific Monthly of February and March, 1917. Since then some development has been recorded in our knowledge of how to use experimental designs as a method of sociological research. Thus the present book is an account of nine experimental studies over a period of years by my students, Helen Christiansen, Nathan Mandel, Julius A. Jahn, John N. Burrus, T. E. Kyllonen, and Marvin Taves, done under my direction, and also an account of several comparable studies by Naomi Barer, Stuart C. Dodd, Reuben Hill, and Harry M. Shulman.

"