Search

Keeble D. E. and Hauser D. P. (1972) Spatial analysis of manufacturing growth in outer South-East England 1960–1967, Reg. Studies 6, 11–36. The paper reports the main findings of multiple regression analyses of the interurban spatial pattern of recent manufacturing growth in South-East England outside Greater London. The variables and hypotheses tested were described in an earlier paper. Manufacturing employment growth varied directly with unemployment rates, population growth, economic potential and expected growth based on industrial structure, but inversely with specialization and industrial size measures. Industrial floorspace growth was related positively to skilled labour availability, population growth, urban size and unemployment rates, but negatively to specialization indices. The paper discusses technical problems such as multicollinearity, spatial autocorrelation and choice of step-wise procedures, and presents special analyses for the new and expanded towns. [ABSTRACT FROM PUBLISHER]

Keeble D. E. and Hauser D. P. (1971) Spatial analysis of manufacturing growth in outer South-East England, 1960–1967, Reg. Studies 5, 229–262. The paper discusses the hypotheses tested and data used in multiple regression analyses of the interurban spatial pattern of manufacturing change in South-East England outside Greater London. This is the most rapidly-growing major industrial region of Britain. Measurements of 39 different independent variables were used to test hypotheses relating spatial variation in seven manufacturing employment and floorspace growth indices to local industrial and firm size structures, level of industrial specialization, the impact of local and metropolitan manufacturing decentralization, market accessibility, labour availability, agglomeration economies and diseconomies, communications proximity, residential space preferences, and planning constraints and policies. A later paper will discuss the results of these analyses. [ABSTRACT FROM PUBLISHER]

This paper presents a forecasting technique which attempts to combine the advantages of both time series analysis and multiple regression. In this two-stage technique, an exponentially smoothed moving average model is used to forecast values of the dependent variable and/or selected independent variables as desired. These forecasts, along with data for other (lagged) independent variables, are then used as inputs to a multiple regression program. The observations are selected sequentially by the regression model so that each equation is based only upon data which would have been available at the time of the forecast, and the coefficients of the equation are updated as new information becomes available. The final section of the paper describes a successful application of the two-stage model to a demand deposit forecasting problem. [ABSTRACT FROM AUTHOR]

The linear discriminant function appears to be a useful statistical technique in sociology and economics, but application of it in these fields has been slow to develop. This paper illustrates its applicability in the analysis of social participation data from three central Pennsylvania communities. The problem of what characteristics of farmers best discriminate between active and inactive participants is used to demonstrate the technique. The results are compared with the results obtained by merely comparing the individual measures for the two groups, and with the results from multiple regression techniques. Some limitations of the linear discriminant function method are pointed out, but it is concluded that this technique is definitely useful in analyzing certain kinds of sociological and economic data. [ABSTRACT FROM AUTHOR]

The article presents commentary and extensions to the dangers of multicollinearity in multiple regression equations by t tests initially asserted within the research of Bruce Cohen and Damodar Gujarati in their report "The Student's t Test in Multiple Regression Under Simple Collinearity," from the September 1970 issue. The particular faults within the mechanical t test specifications are emphasized and criticisms are given regarding a lack of diagnostic methods to detect the error.

This paper reports data on depressive symptomatology in a random sample of 1,645 adults in a Southeastern county. The respondents' scores on a depression scale are analyzed according to age, race, sex, annual family income, education, and a general socioeconomic status score. In order to review simultaneously the relationships between the respondents' scores on the depression scale and the variables under consideration, multiple regression analysis was used. Overall, the regression was significant although the variables of age and race were not. Socioeconomic status was the most significant variable, p < .001; femaleness was also statistically significant at the p < .001 level. The sociodemographic variables used in the multiple regression analysis accounted for 12.6 per cent of the total variance. This suggests that other factors such as psychopathology and/or organic illnesses, and their interactions with social and physiologic factors must be investigated before more definitive conclusions can be reached. [ABSTRACT FROM AUTHOR]

It has long been noted that social dislocation gives rise to mass movements. This paper attempts to relate individual histories of social dislocation to movement adherence, using respondents from Palmira in the Cauca Valley of Colombia. Pentecostalism is viewed as a mass movement, involving working class individuals and total changes in lifestyle. Multiple regression analysis demonstrates that individuals of low SES who have experienced personal social dislocation in terms of migration and employment are more likely to become Pentecostals than their fellows who have not had similar experiences. Primary ties are also important mechanisms for introducing lower class individuals to Pentecostalism. However, they were no more important for Pentecostals who had not suffered personal dislocation than for those who had. [ABSTRACT FROM AUTHOR]

This paper examines the relationship of fertility to three measures of metropolitan dominance: size of place. distance from central city, and the Stoeckel-Beegle size-distance index. Using data from the 1965 National Fertility Study, the number of children ever born to married women under 45 is examined. Fertility differentials exist within all three dominance measures, with rural farm areas having high fertility and central city residents low fertility. However, when duration of marriage, religion, work experience, and education are controlled in a multiple regression model, very little additional variance was explained by metropolitan. dominance measures. It is suggested that fertility differentials are due to the characteristics of the women living in these areas rather than to the influence of the city on the hinterland. [ABSTRACT FROM AUTHOR]

This paper examines some major shortcomings and errors of previous research on the prediction of delinquency and the related conditions of personal and social adjustment, health problems, welfare contacts, and achievement in school. In the Eau Claire County Youth Study a large number of behavioral observations of 1,550 children were conducted in 1961 and 1962. The first problem which shall be noted in this field of research is the paucity of true longitudinal studies. A second major problem generally noted in delinquency research involves the analysis of predictor variables one at a time, with no provision for assessment of interrelationships among predictors. Another problem involves criteria. If tests B, C, D, E, F, and G are all individually correlated with the delinquency criterion Z, then multiple regression analysis makes it possible to combine their predictive power.

By utilizing the multiple regression technique, this paper examines whether various measures of public policies have significant effect in reducing serious crimes and in suppressing the upward trends of crime rates in the 50 largest U,S, cities in recent years. The public policies examined are two conceptually distinct types--control and service policies. The control policies refer to those policies directly concerned with law enforcement and administration of justice while the service policies are those policies which are designed to promote the general well-being and amenities in the community. The crime indicators used are the so-called "index crime" as reported by the FBI. [ABSTRACT FROM AUTHOR]

This paper reports the results of a project to use simple and multiple regression to develop standards for evaluating budget proposals of a university library. Based on a seven-year sample of actual expenditure data, ending with the academic year 1964-65, regression equations were developed for each of the five major categories of library expenditures. These equations were tested against the expenditures approved for the academic year 1965-66. Based on this test, useful standards were developed by employing the regression technique. The technique is proposed as a device for coping with the control of complex activities for which direct standard setting is difficult or prohibitively expensive, in institutional, governmental or business organizations. [ABSTRACT FROM AUTHOR]

ADVERTISING frequently generates controversy. Few industries, however, have so much strong feeling aroused by their promotional activities as pharmaceuticals. This paper discusses some of the specific reasons why pharmaceutical promotion may provide cause for concern and tries by statistical investigation to establish whether or not grounds exist for some of the criticisms to which the industry has been subjected. The paper is in two main sections. Firstly, promotion in the industry and some of the specific arguments which are directed against it are examined. Secondly, the variables used in a cross- sectional, multiple regression analysis carried out on pharmaceutical promotion are described and the results of the analysis presented. The exercise was carried out using promotional data for the calendar year 1966. [ABSTRACT FROM AUTHOR]

This section presents comments and discussion on an article by Lester Taylor on the differences in the response of consumers to changes in different types of income as a possible explanation for the major swings in personal saving that have puzzled economists as of June 1971. Several participants argued that the time series data could not provide reliable answers to the important questions posed by Taylor, no matter how much effort and ingenuity he applied. Both Thomas Juster and R. J. Gordon felt that a time span of a single quarter was too short, given the statistical noise in the data and the variability of actual consumer behavior. Gordon noted that, because personal savings is measured residually as the difference between income and consumption, the errors of measurement in both income and consumption get built into the saving series; this may introduce spurious correlations. He suggested experimenting with variables measured as average changes over two or three quarters in an attempt to reduce the influence of quarter-to-quarter noise. Juster warned that even the modest amount of collinearity among variables reported by Taylor could create problems in the multiple regression approach. Moreover, the exploration for lagged effects surely ran into serious collinearity problems in an attempt to pick out the relative strength of relationships of particular variables to a current quarter's saving and previous quarter's saving.

The article responds to an article by R. Mead in the March 1971 issue of the "Journal of Ecology," which proposes a unified mathematical approach to the analysis of ecological observations. It mentions that Mead has mistakenly interpreted his article as an advocacy of the use of multiple regression analysis for fitting high degree polynomial expressions to such observations. It states that the purpose of a first mathematical representation is to provide an intelligible description of observational data.

The article points out a number of misconceptions about the usage of regression models in ecology. The first misconception to be removed is that multiple regression is valid only when the independent variables are uncorrelated. Another misconception is that the estimated regression coefficient for a specific variable in multiple regression represents the dependence of the dependent variable on this independent variable only for the particular values of the other independent variables occurring in the data sample.

This paper attempts to estimate long-run cost functions for the nationalized passenger road transport industry in India. Road transport marginal costs are compared with Indian railway marginal cost data to determine the social usefulness of alternative transportation systems. The study also examines the question of economies of scale in road transportation which, despite a number of studies over the past 15 years, remain an open issue. Statistical cost curve fitting through multiple regression analysis has been used in this study. The data used are for the nationalized portion of the industry in India. This and the empirical analysis are based on data for fiscal year 1965-66. There are 26 autonomous state corporations for which data were available. Out of the 26 firms, 6 are engaged only in intra-city service while 5 are engaged only in operation in mountainous regions. The remaining 14 firms operate mainly on an inter-city basis. Since terrain and urbanization will affect the cost structure of the firms, they should be held constant.

An investigation of the relationship between the pricing behavior of retail gasoline outlets and a number of identifying characteristics reveals the limited usefulness of association tests. The authors employ multiple discriminant analysis to panel-type data to measure the ability of certain variables to predict price aggressiveness. [ABSTRACT FROM AUTHOR]

This article presents the results of a series of multiple regression analyses in which past market share, advertising expenditure share, and test scores for television commercials were used of predict changes in market shares. The results suggest that the qualify of advertising message content-and-presentation, as measured by that test scores, is closely related to short-term changes in market shires for food and drug products. This relationship implies that marketing models based on dollar expenditures alone cannot provide sufficient explanations of market response to advertising. [ABSTRACT FROM AUTHOR]

Between the late 1920s and the 1950s U. S. butter consumption declined by one-third, while margarine consumption quadrupled. One suggested cause is the repeal of various state laws which restricted the distribution of margarine. The author of this article shows how the validity of this argument can be tested. He adjusted national average time-series data to determine the effects of state laws, and shows how the repeal of laws prohibiting the sale of colored margarine has resulted in higher margarine consumption and lower butter consumption. [ABSTRACT FROM AUTHOR]

The article is a reply to economist John Pencavel's comments about the article "Interindustry Variations in Voluntary Labor Mobility," which was in the January 1969 issue of "Industrial and Labor Relations Review." Standard multiple regression techniques were used in the study to determine the quit rate across manufacturing industries. Pencavel suggested modifications to the specification of equations, provided alternative estimates, and objected to the use of weighted regressions. He also offered evidence on the impact of wage supplements on voluntary separations. The authors disagree with Pencavel's assessment.

A Guttman scale was constructed to measure attitudes of Yale University School of Medicine students and graduates toward possible changes in the organization and delivery of medical care in the United States. When analyzed by multiple regression, the best predictors of a respondent's score on this scale were his political ideology, the organization of his medical practice and whether or not he was still in medical school Type of specialty, father's political philosophy and whether or not a respondent was or planned to become an AMA member were strongly correlated with political ideology. Many other factors did not have independent correlations with attitudes when subjected to multiple regression analysis. The striking intraprofessional variations found in medical and political attitudes document the heterogeneity of American doctors. Any spokesmen for the medical profession should recognize that such heterogeneity exists. [ABSTRACT FROM AUTHOR]

The aim of this research was to increase understanding of the relationship between family variables and measures of child welfare by using the conceptual framework which is being developed by Frank W. Young and Ruth C. Young (1969). The concept of differentiation was measured by ten household-level and nine individual-level indicators, and the concept of integration by twelve indicators. These were factor analyzed and the significant measures in each set used in a multiple regression with four measures of child welfare. Conclusions were that welfare status is not explained by the use of individual measures on their own: integration is an important family dimension to consider along with differentiation, and the wife's level of differentiation has a significant independent effect. [ABSTRACT FROM AUTHOR]

This research explores the consistency of cognitive style. It was hypothesized that a "closed" cognitive style would be associated with a consensual type of religiosity and an "open" cognitive style would be associated with a committed type of religiosity. Instruments used to measure the major variables were: Rokeach's Dogmatism Scale and Spilka's Religious Viewpoints Scale. The sample consisted of 264 college students. Factor analysis and multiple regression were employed in the analysis. The findings indicate that the Religious Viewpoints Scale did not differentiate well between committed and consensual types of religiosity, but alpha factor analysis showed that the scale seemed to be measuring these two dimensions of religiosity. "Closed" (as opposed to "open") cognitive style was more positively associated with consensual religiosity than with committed religiosity; this finding was significant at the .001 level. [ABSTRACT FROM AUTHOR]

Logan M.I. (1970) The spatial dimensions of economic development: The case of the upper Midwest, Reg. Studies 4, 117–125. Few theoretical notions have been developed with reference to regional patterns of economic development. Three hypotheses, however, may be formulated: (1) regional economic development occurs on a number of interpretable dimensions, (2) a region's level of development is associated with its functional integration into the wider economic system and (3) regional development is associated with distance from large urban centres. The study area consists of the states of Minnesota, Wisconsin, and Michigan, the data is for counties and the methodology is principal components analysis and multiple regression analysis. The first hypothesis is verified, the second appears to be supported, but the third, the most important from a spatial viewpoint, is not confirmed. [ABSTRACT FROM PUBLISHER]

This article discusses the application of multiple regression analysis to expense allocation in the insurance industry. The purpose of the approach is to analyze statistically relations which may exist among various quantitative characteristics. It requires that there be data from a reasonable sample of concerns, and, to minimize measurement error. Given the use of a percentage expense factor by rating bureaus in establishing premiums, the technique can be of assistance to departments of insurance and other regulatory bodies.

Examines voice quality changes in depression. Application of voice spectra to track daily shifts in mood in disturbed psychiatric patients; Multiple regression of voice spectra; Measurement of the voice spectrum.

The effect of stimulus-parameters on word-recognition thresholds was investigated. Fifty words were tachistoscopically administered to 24 subjects with an average age of 16.4 years. Of the 40 parameters analyzed, the following were the best predictors of the thresholds and were included in a multiple regression equation: Classification of words into concrete nouns vs. all others; Classification of words with vs. without prefixes; Logarithms of word-frequencies; and Number of letters.

Large variations in the felting rate of loose wool are attributed, principally, to crimp characteristics of the single fibers. Crimp form alone accounts for 77.4% of the variation and, in combination with the directional frictional effect, for 83.3%. Crimp frequency accounts for 32.5% of the variations and, in combination with diameter, for 70.4%. The product of these latter two variables also accounts for a considerable amount of variation (67.1%). [ABSTRACT FROM PUBLISHER]

This article discusses the tools to be used in measuring change. Partial correlations or multiple regression analysis should be used with final status as the criterion and initial status as one of the covariates. Multiple regression analysis is preferable because of its greater flexibility. Change scores, residuals, or base-free measures of change should not be used in statistical analysis. Residuals can be used to compare the performance of individuals or groups while holding the initial status variable.

This article focuses on the relationship between sales and advertising. The models developed herein and tested via multiple regression analysis, are outgrowths of two concepts: one, the notion that advertising expenditures are similar in many ways to capital investments and second, a closely related one that the effects of advertising tend to persist beyond the current period. The models were formulated by focusing attention on the relevant economic parameters motivating business decisions regarding advertising outlays. First, the concept of diminishing returns to advertising is not permitted to be incorporated. Second, it was assumed that the depreciation rate of goodwill was constant. As most capital goods, advertising goodwill probably depredates faster initially and its rate subsequently slows. Last, the relationship between advertising expenditures and price is ignored, an omission economists would be quick to attack. Inclusion of these factors would undoubtedly enhance the predictability of our models or at least improve their veracity.

This article presents a study which examines the hypothesis that there are uniform patterns of change in the structure of production as income levels rise. This study consists of three parts: reestimation of multiple regressions describing intercountry growth patterns for major sectors and country groups, which provide a more satisfactory treatment of the effects of differences in income level, scale and trade patterns; comparison of postwar changes in each group of countries to the intercountry regressions and to the historical patterns of the advanced countries; and analysis of twelve industry sectors designed to provide a disaggregated view of production patterns. A development pattern may be defined for a given country by the time paths of variables describing production, domestic use, international trade, and resource allocation in each sector. Large countries are expected to industrialize earlier than small ones because economies of scale shift their comparative advantage toward industry.

Results are reported from an investigation of the predictive power of three interviewer social status and five personality factors in a multiple regression analysis with four interviewer performance criteria. Personality factors are found to be more predictive of performance than social status. Clearly, one can improve the probability of hiring quality interviewers by collecting and evaluating social status and personality characteristics of applicants. [ABSTRACT FROM AUTHOR]

This study compares two statistical techniques-multiple re gression and con figural analysis-used in developing parole pre- diction tables, according to their ability to (1) differentiate between offenders who violate parole and those who do not, (2) predict violators from among a new group of parolees, and (3) assist administrators and researchers. First, experience tables had to be developed and tested for prediction ability. Once their accuracy in predicting had been demonstrated,, they could be used as base expectancies because they had the quality of being "expected." As such, they could be used as a yardstick to evaluate correctional program? ability to reduce these "expected" violation rates. The two methods were applied to the same body of data and the results were coin pared. The data consist of Wisconsin adult male offenders paroled in 1954-57 and in 1958-59. All were followed up for two years while they were on parole. The first group was used to develop, the experience tables; the second group was used to test prediction ability. The tables were compared for accuracy in predicting through use of the Dc index and for accuracy in differentiating by the J index, measures developed by H. Richard John using Daniel Glaser's data gathered for federal parolees. Both configural and regression methods worked about equally well in differentiating between violators and nonviolators and in estimating the proportion who will become violators in later groups. A combination, of approaches is suggested. [ABSTRACT FROM AUTHOR]

The purpose of this study was to investigate whether the prediction of suggestibility with the autokinetic phenomenon could be made from a weighted composite of selected personality variables, as measured by the Berhreuter Personality Inventory and the MMPI. After showing that a change in perceived distance of light movement could be induced by suggestion, two personality tests were administered to 20 subjects, to get measures of the traits, Introversion-Extraversion, Ascendance-Submission, Self-Sufficiency, and Hysteria. From the five factors (the above four plus suggestibility), a multiple regression was arrived at predicting suggestibility from the remaining four. Twenty other subjects were then tested with the personality inventories. No suggestibility measure was made at that time. Utilizing the multiple regression equation derived earlier, two groups of 10 subjects each were then set up: (a) those prédicted to be most suggestible, and (b) those predicted to be least suggestible. These groups were then measured for suggestibility. Significant differences between the two groups were found in what was here defined as suggestibility. This supported our major hypothesis. [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]

This article focuses on the correlations between membership and social class in voluntary associations in the U.S. The influence of voluntary associations in the U.S. society has been the subject of many inquiries, for the most part, empirical investigations directed at distinguishing individuals who participate in such associations from those who do not participate, have been limited to analyses of the relationship between membership and social class. These social class studies have been conducted in a number of communities, and there is little doubt about the generality of the finding that members of voluntary associations cluster in middle-class families. This article presents a study which employs complementary quantitative methods, multiple regression analysis and factor analysis, in an attempt to discover variables, in addition to social class, that distinguish those who join voluntary associations from those who do not. The logic of the analysis is analogous to that employed in contemporary prediction research, especially in the field of criminology.

The problem of 'prediction of future hospital bed needs' cannot be adequately solved by usual prediction techniques in case of high occupancy rates of a hospital over the past years, since trends are not recognized in this situation. The following method is proposed if the usual prediction techniques do not seem adequate (1) Relate the number of patients per month to factors that can be predicted by using multiple regression techniques (2) Predict the number of patients per month with the help of multiple regression equations (3) Predict the number of beds from the predicted number of patients per month, by multiplication of the predicted number of patients per month by the fraction. [ABSTRACT FROM AUTHOR]