Conventional Statistical Techniques Research Articles

Improving data utilization: the case-wise alternative

Both research and practical experience suggest that alternative forms of data analysis and presentation may affect their use in the policy arena. The variable-wise alternative produces estimates of population parameters through conventional statistical techniques. The case-wise alternative defines and identifies qualitatively distinct types of cases through techniques of numerical taxonomy. In the experiment reported here, variable-wise and case-wise presentations of Census data on poverty were used to compare the effects of the alternatives on policy makers in a priority-setting task. The results show that the case-wise alternative can improve utilization, assimilation, recall, and recognition of the data. Under the circumstances examined here, the case-wise alternative is a practical means of increasing the return on the public's investment in data resources to improve policy decisions.

The Stability of Precipitation Samples under Field Conditions

Abstract Eight wet-only samplers. placed in one location were used to sample precipitation. Subsamples were taken from the sample bottles directly after the precipitation event and at regular intervals up to a period of 2100 hours and analysed, while the samples stayed in the samplers under field conditions. Outliers were deleted by means of conventional statistical techniques. The resulting data base was tested by means of the F-test for consistency and for changes in the chemical composition of precipitation samples directly after sampling or after an extended stay under field conditions. Eighty-one percent of the samples showed consistent results and no clear indication of chemical changes was found. Regression analysis indicated that the uncertainties for samples analysed after 1 week are less than 5% generally for the bulk elements and less than 14% for the trace elements. After a period of 400 hours the uncertainties for the buld elements increase to a level of 5–10% and 5–30% for the trace elements...

Near infra‐red analysis of grass silage by principal component analysis of transformed reflectance data

AbstractMost of the observed variation between n.i.r. spectra in any sample collection is due to particle size variation; in addition, absorbance readings are highly inter‐correlated. Application of multiple linear regression techniques to such spectral data produces a multiplicity of solutions which vary in the wavelengths selected and in predictive accuracy. Evaluation of all or most of these calibrations requires a significant amount of time. In the present study, alternative data treatment methods are described which reduce the effect of particle size variations, select only those wavelengths which contain significant information, and overcome the problem of inter‐correlation by means of principal component analysis. These methods were applied to the analysis of dried silage for crude protein (CP) and in vitro dry matter digestibility (IVDMD); accuracy of the novel data treatments (best standard error of prediction of CP and IVDMD equal to 0.63 and 2.7, respectively) was better than that generally achieved by more conventional statistical techniques.

The Measurement of Values in Surveys: A Comparison of Ratings and Rankings

Social values are most commonly measured using ranking techniques, but there is scarcity of systematic comparisons between rankings and other approaches to measuring values in survey research. On the basis of data from the 1980 General Social Survey, this article evaluates the comparability of results obtained using rankings and ratings of valued qualities. The comparison focuses on (1) the ordering of aggregate value preferences and (2) the measurement of individual differences in latent value preferences. The two methods are judged to be similar with-respect to ordering the aggregate preferences of the sample, but dissimilar with regard to the latent variable structure underlying the measures. Duane F. Alwin is Research Scientist and Professor of Sociology, Survey Research Center, Institute for Social Research, The University of Michigan. Jon A. Krosnick is Assistant Professor of Psychology, The Ohio State University. This article is revised version of paper presented at the annual meetings of the American Association for Public Opinion Research, May 17-20, 1984. An earlier version of the paper was reprinted as Technical Report No. 40, General Social Survey, NORC, Chicago. The research was supported by grant from the National Institute of Mental Health (MH-37289). An important debt is owed NORC and the staff of the General Social Survey for cooperation in conducting this research. The authors wish to thank Frank Andrews, David Jackson, Melvin Kohn, and Tom Smith for helpful comments on earlier drafts and Lynn Dielman and Marion Wirick for help in manuscript preparation. Address all correspondence to Duane F. Alwin, Survey Research Center, Institute for Social Research, P.O. Box 1248, Ann Arbor, MI 48106. Public Opinion Quarterly Vol. 49:535-552 C) by the American Association for Public Opinion Research Published by Elsevier Science Publishing Co, Inc. 0033-362X/85/0049-535/$2.50 This content downloaded from 157.55.39.116 on Sat, 11 Jun 2016 05:15:19 UTC All use subject to http://about.jstor.org/terms 536 DUANE F. ALWIN AND JON A. KROSNICK have number of problems tied to their degree of difficulty, and some have suggested that other techniques, such as ratings, could serve as effective alternatives. In this article, we compare the rating and ranking approaches to measuring values using data on parental orientations toward children from randomized split-ballot experiment carried out in the 1980 General Social Survey. Before reporting our findings, we place the work in context by reviewing the literature that has discussed the relative advantages and disadvantages of rating and ranking techniques for measuring values. The Measurement of Values: Ratings vs. Rankings Value researchers have consistently argued on theoretical grounds that ranking techniques provide the most appropriate conceptual mapping to conceptions of values. For example, Rokeach (1973:6) points out that values are often thought to be inherently comparative and competitive, and thus the choice nature of the ranking task fits nicely with this conceptualization. Also, Kohn (1977:19) observes that the ranking approach to measuring values is demanded by their very nature, in that a central manifestation of value is to be found in choice. This point of view is validated to some extent by the sheer prevalence of the use of rankings to measure values (e.g., Allport et al., 1960; Kluckholn and Strodtbeck, 1961; Lenski, 1961; Bengston, 1975). Despite these arguments regarding the conceptual mapping of ranking techniques to the concept of values, this approach has number of significant practical drawbacks when used in survey research. First, rankings are often difficult and taxing for respondents, demanding considerable cognitive sophistication and concentration. This is particularly problematic when the list of concepts to be ranked is lengthy (Rokeach, 1973:28; Feather, 1973:228). Second, the use of ranking techniques is time-consuming and may therefore be more expensive to administer (Munson and McIntyre, 1979:49). Third, since rankings often require the use of visual aids, or show-cards, it is difficult to gather such information using telephone methods of data collection (Groves and Kahn, 1979:122-33). And finally, the sum of the ranks for any individual respondent equals constant, so there is linear dependency among the set of ranked items (Clemans, 1966; Jackson and Alwin, 1980). Consequently, it may not always be possible to employ conventional statistical techniques in the analysis of the latent content of ranked preference data. In contrast to rankings, rating scales are easy to present to respondents. Munson and McIntyre (1979:49) estimate that ranking tasks take three times longer than similar rating tasks and involve considerable decrease in respondent burden. Also, ratings can readily be administered over the This content downloaded from 157.55.39.116 on Sat, 11 Jun 2016 05:15:19 UTC All use subject to http://about.jstor.org/terms RATINGS AND RANKINGS 537 telephone and do not involve the linear dependency problem inherent in rankings. So ratings have none of the major disadvantages of rankings and might therefore be good substitute for survey measurements of

Tensor biometrics for changes in cranial shape.

D'Arcy Thompson's method of Cartesian transformations can be adapted to the biometric study of group differences and trends in shape. The key to this extension is the symmetric tensor, a mathematical representation of the linear deformation taking one triangle onto another. These tensors may easily be computed from data that come in the form of Cartesian coordinates of homologous landmarks, such as are customary in roentgenographic cephalometrics. This article shows how the tensors make possible the rigorous statistical analysis of populations of shapes without requiring the specification of particular shape measures in advance. Instead, the findings of a study include the extraction of particular shape measures most clearly manifesting the group difference or treatment effect under study. Only conventional multivariate statistical techniques, such as Hotelling's T2, are involved. Configurations of many landmarks can be analysed via the display of their mean difference as a biorthogonal grid pair, perusal of which leads to the selection of triangles of landmarks for more rigorous analysis. The techniques are exemplified in two cephalometric studies: the description and prediction of normal craniofacial growth, and a comparison of the craniofacial deformities associated with Apert's and Crouzon's syndromes.

A note on the use of directional statistics in weighted euclidean distances multidimensional scaling models

The weighted euclidean distances model in multidimensional scaling (WMDS) represents individual differences as dimension saliences which can be interpreted as the orientations of vectors in a subject space. It has recently been suggested that the statistics of directions would be appropriate for carrying out tests of location with such data. The nature of the directional representation in WMDS is reviewed and it is argued that since dimension saliences are almost always positive, the directional representations will usually be confined to the positive orthant. Conventional statistical techniques are appropriate to angular representations of the individual differences which will yield angles in the interval (0, 90) so long as dimension saliences are nonnegative, a restriction which can be imposed. Ordinary statistical methods are also appropriate with several linear indices which can be derived from WMDS results. Directional statistics may be applied more fruitfully to vector representations of preferences.

Level crossing statistics and Hausdorff‐Besicovitch similarity dimension of generalized rough surfaces

Generalized rough surfaces can exhibit a rich variety of relief that cannot be easily described as either periodic or Gaussian. For such surfaces, one can explore the spatial distribution per unit length of occurrence of the mean height (=0 by definition), the rms heights, or the crossing of the relief with a chosen level. An extension of this concept is also possible with joint probability distribution characterization of the distribution of successive level crossing known as Palm distributions. Evaluation of these measures for some test functions and stochastically generated surface relief is presented and the benefits of their use is contrasted with conventional statistical and correlation function techniques. Surfaces that contain a possible infinite hierarchy of scales, e.g., fractal surfaces, are shown to best be characterized by a parameter known as the Hausdorff‐Besicovitch similarity (HBS) dimension. The use of this parameter in describing generalized rough surfaces possessing self‐similar characteristics or multiple scales of roughness is described and the reflection characteristics of such surfaces in optical and acoustical applications are discussed.

A comparative study of the multivariate structure of towns

Conventional multivariate statistical techniques are inhibiting in their analysis of structure. This paper introduces the ‘language structure’ or ‘Q-analysis’ proposed by R H Atkin, and compares its results with widely adopted multivariate statistical techniques within the context of the socioeconomic structure of towns in England and Wales.

FOCAS - Faint Object Classification and Analysis System

view Abstract Citations (324) References (18) Co-Reads Similar Papers Volume Content Graphics Metrics Export Citation NASA/ADS Focas: faint object classification and analysis system. Jarvis, J. F. ; Tyson, J. A. Abstract The FOCAS (Faint Object Classification and Analysis System) apparatus for the detection and classification of images on astronomical plates which is intended to be used for the automatic assembly of catalogs of faint objects, is presented. The system has been used to process plates obtained by the KPNO 40-m telescope under good seeing conditions for 12 selected areas, which were digitized in 20-micron steps with a 20-sq micron pixel in a 6000 x 6000 raster corresponding to a field 37 arcmin on a side. The FOCAS acts to reduce the digitized image to a list including the magnitude, position and classification data for objects in the field as faint as the 24th magnitude, by a process including the segmentation of objects, object area assembly, and object assembly. Objects are then classified as noise, stars or galaxies on the basis of conventional nonparametric statistical pattern recognition techniques, and catalogs of objects meeting given criteria can be obtained by various programs. System tests have shown detection efficiencies of 97% and classification capabilities for 90 and 80% of the galaxies and stars, respectively, for apparent magnitudes down to 24, although the optimality and accuracy of the classifier have not been established. FOCAS has been used to assemble catalog of 86,000 objects which is 3 to 4 mag deeper than existing surveys. Publication: The Astronomical Journal Pub Date: March 1981 DOI: 10.1086/112907 Bibcode: 1981AJ.....86..476J Keywords: Astronomical Catalogs; Astronomical Photography; Celestial Bodies; Image Resolution; Systems Analysis; Classifications; Digital Systems; Galaxies; Magnitude; Pattern Recognition; Photographic Plates; Stars; Astronomy full text sources ADS | data products SIMBAD (7)

DATA ANALYSIS FOR SIMULATION EXPERIMENTS: APPLICATION OF A DISTRIBUTION‐FREE MULTIPLE COMPARISONS PROCEDURE

AbstractThe analysis of simulation data in the evaluation of competing alternatives presents a problem for the analyst using conventional statistical techniques. The assumptions of normality and common variances generally cause difficulties since many classes of simulation experiments typically violate both of these assumptions. In addition, the analyst is usually interested in comparing competing alternatives using environments that are as close to identical as possible. In these situations, blocking is desirable and can usually be accomplished by using common random number streams. This paper discusses a distribution‐free method based on Friedman's rank sums test that can be used to analyze simulation results exhibiting the above characteristics. The procedure requires only independence within treatments.

Evaluation of U. S. Pharmacopeia Sampling Plans for Dissolution

The United States Pharmacopeia presents sampling plans and specifications for pharmaceutical products which are used in judging the quality of these products. Previously accepted and revised sampling plans for product dissolution are evaluated and compared using conventional statistical techniques and computer simulation. This paper presents operating characteristic curves and related graphs for these dissolution sampling plans that form a basis upon which lots of dosage units are accepted or rejected. These plans have also been evaluated for their effectiveness in assuring the quality of drug products.

Detection of low-contrast images in film-grain noise

When low contrast photographic images are digitized by a very small aperture, extreme film-grain noise almost completely obliterates the image information. Using a large aperture to average out the noise destroys the fine details of the image. In these situations conventional statistical restoration techniques have little effect, and well chosen heuristic algorithms have yielded better results. In this paper we analyze the noisecheating algorithm of Zweig et al. [J. Opt. Soc. Am. 65, 1347 (1975)] and show that it can be justified by classical maximum-likelihood detection theory. A more general algorithm applicable to a broader class of images is then developed by considering the signal-dependent nature of film-grain noise. Finally, a Bayesian detection algorithm with improved performance is presented.

Using regression techniques in cost analysis Part I†

Abstract Computer time-sharing is having a powerful effect on managerial decision-making in the cost analysis and estimating area, through the easy availability of multiple regression analysis. Unfortunately, the regression technique is still obscured by factitious technicalities, while the very nature of cost problems limits the value of conventional statistical techniques in evaluating the usefulness of an analysis. This article attempts to clarify the situation through a review of selected cases in which regression is used, and by suggesting evaluation techniques developed during extensive applications. Serious inadequacies in standard regression calculations and computer programmes are pointed out, new measures of formula adequacy are described and we conclude with a discussion of the criteria by which a cost formula may be judged to be ' reasonable '.

Put Engineering Efforts Back in Reliability Techniques

Conventional statistical reliability techniques are a necessary condition of reliability. They are, however, insufficient and inadequate. Other reliability techniques are needed to turn a feasibly designed system into a completely successful one. The implementation of these techniques is lagging. They have not as yet been unified on an industry-wide basis. They must be implemented and unified, however, to attain the reliability goals desired. The majority of specific reliabilityfunded efforts since World War II have had as their goal the design and fabrication of parts. To this end, large sums of money have been spent and progress has been made. It is not suggested, of course, that the effort to improve parts be abandoned. The method of investigating, improving and using these parts should be the prime technique, and should be implemented. The state of the equipment art has been considered a limiting factor in implementing ever higher goals of reliability. The limitations are, in reality, reliability semantics, and our inability to properly utilize the techniques that exist in the field of reliability. Another large portion of the reliability dollar has been spent in improving by corrective action, resulting in after-the-fact know-how. Reliability personnel must engineer before-thefact, explore new physical, chemical, and manufacturing concepts, and communicate these before-the-fact findings to the personnel concerned. We do not require better definition of consistency in reliability requirements. We do need reliability personnel who are concerned not only with the strict quantitative interpretation-- but who live in the technical climate of today's business world. Most importantly, perhaps, these people must be capable of communicating with the customer and the designer/producer to the mutual satisfaction of both. Reliability techniques which embody engineering principle s are presented which are not necessarily new to personnel in small companies, to users of equipment of uncomplex design, or to reliability groups who have been actively doing instead of agreeing to do. However, these methods should be reviewed by the newcomers in reliability organizations and the newcomers not in reliability organizations but who tend to think that reliability groups are intended to be statistics oriented.

Statistical analysis of body weight data

Statistical methods, currently employed for examining body weights collected in toxicologic studies, were reviewed and illustrated. The use of nonparametric analogs of conventional parametric statistical techniques was described. Nonparametric techniques have not been widely employed for evaluating weight data. The results obtained by parametric techniques were compared with those obtained by using their nonparametric analogs. It was concluded that nonparametric techniques are of value in toxicologic investigations if a rapid, preliminary survey of the effects of treatment on body weight is desired.

An Extension to 100,000 Feet of the U.S.A.F. Standard Cold and Hot Atmospheres

For the purpose of setting up engineering standards the U. S. A. F. was requested to determine standard cold and hot atmospheres from 60,000 feet to 100,000 feet. Data were too scarce to use conventional statistical techniques and an alternate method is presented herein. Included are graphs of the cold and hot atmospheres for each of four latitudinal zones from sea level to 100,000 feet, and a table of heights and temperatures at critical points along straight-line envelopes enclosing all zonal curves.