Note: To find revision information for really old versions of UCINET look in these two places:  Related Links
 To see a list of known bugs and user wish list items, check the fix list
 To download the latest version, click download.

News

2day UCINET workshop
I will be offering a 2day workshop providing an introduction to social network analysis using UCINET in Philadelphia Friday April 29th30th, 2016. Find the details here:http://statisticalhorizons ...
Posted Dec 10, 2015, 1:27 PM by Steve Borgatti

posted Jan 22, 2016, 6:13 PM by Steve Borgatti
[
updated Jan 22, 2016, 6:21 PM
]
 Changed netdraw button on toolbar to immediately ask for a file name before starting netdraw
 Added cutpoints routine to CLI (aka matrix algebra). Given a network, it outputs a vector V in which v(i) = 1 if node i is a cutpoint. Unlike menu version, works for both undirected and directed graphs. Syntax and example are as follows:
><outfile> = cutpoints(<dataset>)
>cp = cut(campnet) >dsp cp

posted Jan 16, 2016, 12:29 PM by Steve Borgatti
[
updated Jan 16, 2016, 12:35 PM
]
 Fixed bug in Double Dekker QAP regression. When coefficients were sufficiently small, the program was setting them to zero. As a result, when it compared the observed coefficient (zero) with the coefficient from a random permutation (also zero), it was always finding that the permutation value was as large as the observed, yielding a pvalue of 1.0.
 Fixed bug in TransformSymmetrize which was only symmetrizing the first matrix in a dataset

posted Dec 16, 2015, 10:32 AM by Steve Borgatti
 Yet another install package for the 64bit version  previously, after installation the program would not start up
 Change in NetworksCohesionAlters in Common so that valued data would automatically be dichotomized. In addition, missing values are now ignored
 Bug fixed in TransformAggregateCSS. Previously, the labels Pa and pA were reversed. Now, the value labeled Pa correctly gives the number of ties Perceived but not Actually there, while pA gives the number of Actual ties that the respondent failed to Perceive

posted Dec 9, 2015, 12:22 PM by Steve Borgatti
 Uploaded new install package for the 64bit version
 Added option to Gould & Fernandez brokerage to allow closed triads to count as brokerage instances as well as open triads
 Reverted minres routine in continuous coreperiphery procedure to old version that seems to handle negative values and nonsymmetric data better

posted Nov 11, 2015, 1:53 PM by Steve Borgatti
[
updated Nov 13, 2015, 10:32 AM
]
 Minor change in TransformAggregateCSS to make default output file names match the method of creating the "true" network
 Fixed bug in NetworksCohesionDensityDensity by Groups which was causing crashes when the user selected the Identity partition instead of a categorical attribute
 Fixed another bug in NetworksCohesionDensityDensity by Groups which was failing to retain node labels when Identity partition was used
 Added CLI command for aggregating columns of a matrix, much like Density by Groups. It is called AggCol and is used like this:
>gender = col(campattr "gender") >dsp aggcol(campnet gender)
The result, in this case, would be counts of the number of male and female alters that each node has:
1 2 1 2   1 1 2 1 2 2 0 3 3 3 3 0 4 4 3 0 5 5 3 0 6 6 3 0 7 7 3 0 8 8 3 0 9 9 1 2 10 10 0 3 11 11 1 2 12 12 1 2 13 13 1 2 14 14 1 2 15 15 0 3 16 16 0 3 17 17 0 3 18 18 0 3

posted Oct 25, 2015, 6:11 PM by Steve Borgatti
[
updated Oct 25, 2015, 6:15 PM
]
 Added routine in CLI to calculate Goodman and Kruskal Gamma coefficients (measures of rankorder similarity) on columns of a matrix. Example usage is:
>coefs = ordinalassoc(campnet) >dsp coefs

posted Sep 6, 2015, 8:31 AM by Steve Borgatti
[
updated Sep 6, 2015, 8:57 AM
]
 Added att2mat command to CLI. This converts an attribute to a matrix. The syntax is <outfile> = att2mat(<vector> [<method>]) where the choices for method are: 'match', 'diff', 'absdiff', 'sqrdiff', 'prod', 'sum', 'identity', 'rows', and 'cols'. The default is 'match'. Example:
>gen = col(campattr gender) >samegender = att2mat(gen match)  Added mat2vec command to CLI. This vectorizes a matrix so that it looks like one long column. If the input dataset contains multiple relations, then each becomes a column in the vectorized dataset. The syntax is <outfile> = mat2vec(<dataset> [<method>] [diagok]) where the choices for method are: 'whole', 'lowerhalf' and 'upperhalf'. The default is 'whole'. By default, the diagonal is omitted, but if want the diagonal included, add 'diagok' to the list of parameters. Example:
>pvec = mat2vec(padgett lowerhalf) >svec = mat2vec(sampson whole diagok)
posted Jul 23, 2015, 6:33 PM by Steve Borgatti
 Fixed bug in Matrix editor. It had been crashing when the user saved a dataset after having cleared the editor using the New button
posted Jul 20, 2015, 11:23 AM by Steve Borgatti
[
updated Jul 20, 2015, 11:55 AM
]
 Fixed bug in TransformSymmetrize in which it was failing to heed the user's choice of method (for the methods lower down in the list).
 Fixed bug in ToolsUnivariate which was failing to save the output statistics for all but the first matrix in a dataset.
 Fixed bug in ToolsHypothesis TestingNode levelRegression which was putting in a nonsense number for the standardized beta coefficient of the intercept (which is supposed to be 0). Nothing else was affected by this.
 [64bit version only] Have added warnings to most (but not yet all) routines which do run in 64bit mode. This is better than simply crashing ...
posted Jul 16, 2015, 7:05 PM by Steve Borgatti
 Fixed bug in Network2modeCategorical Core/Periphery Model that was causing crashes. Note: this is a routine that does not run in the 64bit version.
