Jan Vegelius
Biographic Data
| ID | 8867397 |
|---|---|
| NAME | Jan Vegelius |
| GIVEN NAMES | Jan |
| FAMILY NAME | Vegelius |
| SIGNATURE | VEGELIUS J |
| AFFILIATIONS | Uppsala University |
| VERIFIED | No |
| TOTAL WORKS | 29 |
| TOTAL CITATIONS | 0 |
| WORKS AS AUTHOR | 29 |
| WORKS AS EDITOR | 0 |
| FIRST PUBLICATION YEAR | 1975 |
| LATEST PUBLICATION YEAR | 1987 |
| H-INDEX | 0 |
Corall-2
This new version of the CORALL program contains around 50 various measures of similarity and correlation-as well as some other elementary statistical measures
Efam
A program for computing correlation coefficients of variables from different scale types is presented. The computations are based on the concept of the E-correlation family of coefficients. Such a general coefficient is invariant over permitted transformations of the variables for their respective scale types. The program EFAM is able to consider the E-correlations from interval, ordinal, and nominal scales
A Q-Analysis for Nominal Data
The possibility of using a Q-analysis also for nominal data is discussed. The J-index is suggested as a suitable measure of similarity between the persons in this case. An example is given where ten persons should sort sixteen playing cards into as many groups as they wished. A Q-analysis was made on these data giving a natural two-dimensional structure
Hcla
The computer program HCLA for hierarchical cluster analysis is described. It gives the user various possibilities both for strategies and distance and similarity measures
Correlation Coefficients For More Than One Scale Type
The problem of correlating variables from different scale types is discussed. A general correlation coefficient, based on symmetrization theory, is derived. This coefficient is invariant over permitted transformations of the variables for their respective (possibly nonequivalent) scale types. Such a general coefficient is called an E-correlation family. One such family (based on the product-moment correlation coefficient and a nominal scale corre…
An Enquiry About Religion, Analyzed By a Nominal Scale Truncated Component Analysis
The data from an enquiry among 118 theological students with 24 items is analyzed by a truncated component analysis, where only nominal information is used. Tschuprow's T 2 is applied as the correlation coefficient. Eight easily interpreted factors are given by the method
Significance Tests for the G-Index
Some tests for the G-index are described. For each case an example is included
WHIDD, a Fortran IV Program for Whidd Analysis
The computer program WHIDD makes the WHIDD analysis available for computer users. The input and output of the program are described. A presentation of the details of the method will be found elsewhere
The Relationship Between the Phi Coefficient and the G Index1
The value of the phi coefficient (Φ) cannot, in general, be uniquely determined from the G index. Its value is shown to obey the inequalities G-1/3+ G ≤ Φ ≤ 1+ G/3- G if -1 When G equals -1 or 1 so does Φ. As a consequence of this Φ is too unstable over item reflection when used as a measure of similarity between individuals. G is a more suitable choice
Is the G Index a Correlation Coefficient
For three different definitions of the correlation coefficient concept is examined whether the G index is a correlation coefficient or not. According to the Pearson definition it is not. However, it is equal to the Pearson coefficient, obtained from double scoring. The G index is shown to be a correlation coefficient according to the definition by Kendall. It is also shown to be an E-correlation coefficient
A General Discriminant Analysis Based On E-Correlation Coefficients
A general discriminant analysis method, based on the E-correla tion coefficient concept is introduced. The method is applicable to data on any scale type, even to nominal data. It separates the vari ous subgroups by using Euclidean spaces. Several criteria for how to separate between the subgroups are suggested. The G-analysis, the weighted G analysis and the WHIDD analysis are special cases of the method
On Whidd Analysis
A discriminant analysis method, called the WHIDD-analysis is presented. It is rather similar to Holley's G-analysis. However, it does not utilize any component analysis and is thus cheaper and simpler to use. The WHIDD-analysis uses the weighted H-index as the E-correlation coefficient. In two different sets of data only one person out of 32 were erroneously classified. This was the same accuracy as the more complicated weighted G-analysis had sh…
On the Use of the G Index in O-, P-, Q-, R-, S-, and T-Analyses
The possibility of using the G index and its generalizations in the six basic factor analytic designs is discussed. For O- and Q-analyses G should be used if there is no mutual direction of all the variables, as G is invariant over variable reflection. In all the six designs G should also be used if the scale centre is more suitable as a reference point than the mean value
On the Item Orthogonality Assumption of the G Index and Other Person Similarity Measures
A new measure of similarity between persons applicable in Q-analysis is proposed. It allows assumptions of non-orthogonality (dependence) between the items, across which the similarity is computed. The similarity measure may also be applied in an R-analysis, when the person orthogonality assumption may be considered as doubtful. Finally an artificial example is given
A G Index Generalization for Trichotomized Data
Two G index generalizations for trichotomized data are presented, one which permits the use of various weights for the different items and one which does not. As both indices satisfy the requirements of a scalar product between normalized vectors, a Q component analysis may be based on any of them. The indices are also invariant over item reflection
A Generalized Procedure for the Rotation of Three or More Factors in G Analysis
Presents a generalized technique for the rotation of factors in G analysis in which marker variables are brought into alignment with their adjoining reference axes. The method is a further extension of an earlier system and is applicable for the rotation of any number of factors
Weigan
The computer program WEIGAN makes the weighted G analysis available for computer users. The input and output of the program is described. A presentation of the details of the method will be found elsewhere
On Generalizations Of The G Index And The Phi Coefficient To Nominal Scales
If the same categories are used for two nominal scale variables, this information should be used in similarity measures between those variables. Two such similarity measures, one proposed by Goodman & Kruskal and one (kappa) by Cohen, are examined. Two alternative coefficients, called C and S are further proposed. They are found to be generalizations of the G index and the phi coefficient respectively. Both of them seem to have many desirable cha…
On the Utility of the E-Correlation Coefficient Concept in Psychological Research
The term correlation coefficient has been defined in various ways. In this article the E-(correlation) coefficient concept is considered. Six characteristics of an E-coefficient are mentioned. For 23 similarity measures of interval, ordinal, dichotomous, and nominal data is considered whether they are E-coefficients or not. Finally, the importance of the concept is discussed
Contin, a Fortran IV Program for Nominal Scale Correlation Coefficients
The computer program described contains various coefficients for nominal scales, such as the contingency coefficient, Tschuprow's T 2 , Cramer's V, and the J-index. It may be used to generate a correlation matrix with Tschuprow's T 2 , for example, as the correlation coefficient. This matrix may later be utilized in a truncated component analysis such as the one provided by the BMDO8M program
A Note on the G Analysis for Non-Dichotomized Data
A generalization of the G analysis to cases where the variables are on an interval or ordinal scale is described. It can be based on various transformations which are consistent with the usual G index generalizations for those scales. The placement system used in the description is distance-based, but other systems may also be applied
The Problem of Varying Scales for G Index Generalizations
The usual G index generalizations assume that the response alternatives are the same for all items. Here an attempt is made to overcome that limitation without resorting to dichotomization of the items for cases where the number of response alternatives may differ. This requires the assumption that the influence of the verbal labels of the alternatives are of minor importance
On the Weighted G Index
The G index of agreement does not permit the use of various weights for its various items. The weighted G index described here, makes it possible to use unequal weights. The weighted G index satisfies the requirements of a scalar product between normalized vectors, and a Q component analysis may, therefore, be based on it. It is also invariant over item reflection
An Ordinal Scale Generalization of the G Index Invariant Over Item Reflection
A new type of scale, called the bipolar ordinal scale, is defined with assumptions slightly stronger than those of the normal ordinal scale. A similarity index, called r BP , is defined between persons with scores on this scale. It is invariant over item reflection. It is also an E-correlation coefficient and can thus be used in a Q-analysis
On Weighted G Analysis
An analytic method for discriminating between clinical subgroups is introduced. It is based on a weighted G index and is therefore called weighted G analysis. In this analysis group weights for the various items are determined, after which a Q analysis is run on the analysis group. Scores in the factors for imagined delegates of the clinical subgroups are computed, after which scores of the individuals in a validation group are computed. Using a …
No prominent works on this page.
Corall, a Fortran Iv Program for Correlation Measures
The program can compute a great number of different correlation and other statistical measures. The user is free to select among the measures and also among the variables that are read by the program. When a particular set of variables has been treated in the prescribed way, a new set may follow together with new measure definitions
Siegel, a Fortran Iv Program for Nonparametrical Methods
The program can perform any one of those nonparametric statistical methods that is mentioned in Sidney Siegel's classical work on the subject. The user is free to select among the methods and also among the samples that are read by the program. When a particular set of samples has been treated in the prescribed way, a new set may follow together with new method definitions
On Generalizations of the G Index
The G* index of agreement, introduced by Lienert as a generalization of the G index is shown to have some deficiencies. It cannot, for example, be used in a Q-factor analysis. An alternative index, r oz , is proposed for ordinal scales with fixed neutral (zero-) points. The r oz index can be used for measuring both person similarity and item similarity. It fulfills the demands of a normalized scalar product and can thus be applied in a Q- or an R…
Polari, a Fortran IV Program for Measures of Attitudinal Polarization
The computer program described contains various measures of attitudinal polarizations, vis. K, introduced by McCall and Simmons (1966), and V 1 , V 2 and V 3 , suggested by Edvardsson and Vegelius (1974, 1975). In the latter three measures, two parameters, C and D, may be chosen by the user. The choice C= D= 0.5 is recommended. The two major purposes of this paper were: (1) to describe the theoretical background underlying the mathematical develo…
The Problem of Varying Scales for G Index Generalizations
The usual G index generalizations assume that the response alternatives are the same for all items. Here an attempt is made to overcome that limitation without resorting to dichotomization of the items for cases where the number of response alternatives may differ. This requires the assumption that the influence of the verbal labels of the alternatives are of minor importance
On the Weighted G Index
The G index of agreement does not permit the use of various weights for its various items. The weighted G index described here, makes it possible to use unequal weights. The weighted G index satisfies the requirements of a scalar product between normalized vectors, and a Q component analysis may, therefore, be based on it. It is also invariant over item reflection
An Ordinal Scale Generalization of the G Index Invariant Over Item Reflection
A new type of scale, called the bipolar ordinal scale, is defined with assumptions slightly stronger than those of the normal ordinal scale. A similarity index, called r BP , is defined between persons with scores on this scale. It is invariant over item reflection. It is also an E-correlation coefficient and can thus be used in a Q-analysis
On Weighted G Analysis
An analytic method for discriminating between clinical subgroups is introduced. It is based on a weighted G index and is therefore called weighted G analysis. In this analysis group weights for the various items are determined, after which a Q analysis is run on the analysis group. Scores in the factors for imagined delegates of the clinical subgroups are computed, after which scores of the individuals in a validation group are computed. Using a …
On the Utility of the E-Correlation Coefficient Concept in Psychological Research
The term correlation coefficient has been defined in various ways. In this article the E-(correlation) coefficient concept is considered. Six characteristics of an E-coefficient are mentioned. For 23 similarity measures of interval, ordinal, dichotomous, and nominal data is considered whether they are E-coefficients or not. Finally, the importance of the concept is discussed
Contin, a Fortran IV Program for Nominal Scale Correlation Coefficients
The computer program described contains various coefficients for nominal scales, such as the contingency coefficient, Tschuprow's T 2 , Cramer's V, and the J-index. It may be used to generate a correlation matrix with Tschuprow's T 2 , for example, as the correlation coefficient. This matrix may later be utilized in a truncated component analysis such as the one provided by the BMDO8M program
A Note on the G Analysis for Non-Dichotomized Data
A generalization of the G analysis to cases where the variables are on an interval or ordinal scale is described. It can be based on various transformations which are consistent with the usual G index generalizations for those scales. The placement system used in the description is distance-based, but other systems may also be applied
On the Use of the G Index in O-, P-, Q-, R-, S-, and T-Analyses
The possibility of using the G index and its generalizations in the six basic factor analytic designs is discussed. For O- and Q-analyses G should be used if there is no mutual direction of all the variables, as G is invariant over variable reflection. In all the six designs G should also be used if the scale centre is more suitable as a reference point than the mean value
On the Item Orthogonality Assumption of the G Index and Other Person Similarity Measures
A new measure of similarity between persons applicable in Q-analysis is proposed. It allows assumptions of non-orthogonality (dependence) between the items, across which the similarity is computed. The similarity measure may also be applied in an R-analysis, when the person orthogonality assumption may be considered as doubtful. Finally an artificial example is given
A G Index Generalization for Trichotomized Data
Two G index generalizations for trichotomized data are presented, one which permits the use of various weights for the different items and one which does not. As both indices satisfy the requirements of a scalar product between normalized vectors, a Q component analysis may be based on any of them. The indices are also invariant over item reflection
A Generalized Procedure for the Rotation of Three or More Factors in G Analysis
Presents a generalized technique for the rotation of factors in G analysis in which marker variables are brought into alignment with their adjoining reference axes. The method is a further extension of an earlier system and is applicable for the rotation of any number of factors
Weigan
The computer program WEIGAN makes the weighted G analysis available for computer users. The input and output of the program is described. A presentation of the details of the method will be found elsewhere
On Generalizations Of The G Index And The Phi Coefficient To Nominal Scales
If the same categories are used for two nominal scale variables, this information should be used in similarity measures between those variables. Two such similarity measures, one proposed by Goodman & Kruskal and one (kappa) by Cohen, are examined. Two alternative coefficients, called C and S are further proposed. They are found to be generalizations of the G index and the phi coefficient respectively. Both of them seem to have many desirable cha…
WHIDD, a Fortran IV Program for Whidd Analysis
The computer program WHIDD makes the WHIDD analysis available for computer users. The input and output of the program are described. A presentation of the details of the method will be found elsewhere
The Relationship Between the Phi Coefficient and the G Index1
The value of the phi coefficient (Φ) cannot, in general, be uniquely determined from the G index. Its value is shown to obey the inequalities G-1/3+ G ≤ Φ ≤ 1+ G/3- G if -1 When G equals -1 or 1 so does Φ. As a consequence of this Φ is too unstable over item reflection when used as a measure of similarity between individuals. G is a more suitable choice
Is the G Index a Correlation Coefficient
For three different definitions of the correlation coefficient concept is examined whether the G index is a correlation coefficient or not. According to the Pearson definition it is not. However, it is equal to the Pearson coefficient, obtained from double scoring. The G index is shown to be a correlation coefficient according to the definition by Kendall. It is also shown to be an E-correlation coefficient
A General Discriminant Analysis Based On E-Correlation Coefficients
A general discriminant analysis method, based on the E-correla tion coefficient concept is introduced. The method is applicable to data on any scale type, even to nominal data. It separates the vari ous subgroups by using Euclidean spaces. Several criteria for how to separate between the subgroups are suggested. The G-analysis, the weighted G analysis and the WHIDD analysis are special cases of the method
On Whidd Analysis
A discriminant analysis method, called the WHIDD-analysis is presented. It is rather similar to Holley's G-analysis. However, it does not utilize any component analysis and is thus cheaper and simpler to use. The WHIDD-analysis uses the weighted H-index as the E-correlation coefficient. In two different sets of data only one person out of 32 were erroneously classified. This was the same accuracy as the more complicated weighted G-analysis had sh…
An Enquiry About Religion, Analyzed By a Nominal Scale Truncated Component Analysis
The data from an enquiry among 118 theological students with 24 items is analyzed by a truncated component analysis, where only nominal information is used. Tschuprow's T 2 is applied as the correlation coefficient. Eight easily interpreted factors are given by the method
Significance Tests for the G-Index
Some tests for the G-index are described. For each case an example is included
A Q-Analysis for Nominal Data
The possibility of using a Q-analysis also for nominal data is discussed. The J-index is suggested as a suitable measure of similarity between the persons in this case. An example is given where ten persons should sort sixteen playing cards into as many groups as they wished. A Q-analysis was made on these data giving a natural two-dimensional structure
Computer Science (27 works) · Mathematics (27 works) · Statistics (27 works) · Index (typography) (14 works) · Artificial Intelligence (10 works) · Artificial Intelligence (10 works) · Correlation (10 works) · Multi-Criteria Decision Making (9 works) · Programming language (9 works) · Psychology (9 works)