This page shows changes to UCINET between versions. Just scroll down to see release notes in reverse chronological order.
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


  • Week-long workshop on SNA The LINKS Center at the University of Kentucky is offering its annual 1-week summer workshop on social network analysis June 6-10, 2016 on the University of Kentucky campus ...
    Posted Mar 15, 2016, 12:54 PM by Steve Borgatti
Showing posts 1 - 1 of 9. View more »

Version 6.625 | 16 Nov 2016

posted Nov 16, 2016, 6:32 PM by Steve Borgatti   [ updated Nov 16, 2016, 7:04 PM ]

  • Fixed bug in the DL editor. When importing data in matrix form in which there were duplicate labels (e.g., two nodes named "steve"), the program was incorrectly assigning all ties to the first of the duplicate nodes. 

Version 6.624 | 15 Oct 2016

posted Oct 16, 2016, 5:29 AM by Steve Borgatti   [ updated Oct 16, 2016, 5:30 AM ]

  • Added subgraph command to the CLI. Given a square matrix representing a network and a vector indicating membership of nodes in groups, the subgraph command allows you to extract a subgraph consisting only of nodes belonging to one or more groups. The syntax is <outfile> = subgraph(<input matrix> <attribute vector> <relational operator> <value>). For example:
    ->gender = cols(campattr 1)
    ->boys = subgraph(campnet gender eq 1)
    ->girls = subgraph(campnet gender eq 1)

    If the operator and value are omitted, the program assumes "GT" and "0". For example, if you have a vector of 1s and 0s indicating who is a member of a given component, you could write:

    ->mainweb = subgraph(campnet maincomp)

    This would select all nodes with a value greater than zero on maincomp.

Version 6.622 | 8 Sept 2016

posted Sep 8, 2016, 1:22 PM by Steve Borgatti   [ updated Sep 11, 2016, 11:46 AM ]

  • Fixed discrepancy in how Networks|Compare Densities|Paired and Networks|Compare Densities|Theoretical parameter worked. Now, given the same random seed, they produce the same density standard errors

Version 6.621 | 5 Sept 2016

posted Sep 5, 2016, 3:19 PM by Steve Borgatti

  • Added cosine similarity (aka Tucker's congruence coefficient) to Tools|Similarities & Differences. 
  • Fixed bug in matrix editor which was adding rows and columns gratuitously

Version 6.620 | 5 July, 2016

posted Jul 5, 2016, 7:56 AM by Steve Borgatti   [ updated Jul 5, 2016, 9:32 AM ]

  • Added exact keyplayer routine to the CLI. The procedure finds the best set of k nodes which, if removed from the network, would maximize the extent of fragmentation (as measured by the number of pairs of nodes that can't reach each other by any path). The algorithm is brute force: it checks all possible subsets of size k. As a result, it is only appropriate for small networks and small k. For example, it can solve N=50 and K=3 in a few seconds, but N=100 and K=3 takes several minutes. Syntax is exhaustivekp(<dataset> <k>), such as:

    ->dsp exhaustivekp(zackar 3)

  • Installer package contains version 2.160 of NetDraw, which changes the way the Width checkbox on the Rels tab works. Before, the width of the line between two nodes was determined by the average tie strength of all tie types currently active. Now, the width is determined by the largest tie strength of any currently active types of tie between that pair of nodes. 

UCINET 6.620 | 13 June, 2016

posted Jun 13, 2016, 11:11 AM by Steve Borgatti   [ updated Jun 13, 2016, 11:12 AM ]

  • Fixed important bug in Data|Affiliations|Mode=Rows. When given a multi-matrix dataset as input, it was writing out each 1-mode projection twice.
  • Replaced k-plex routine in Network|Groups|K-Plexes

UCINET Version 6.619 | 10 June 2016

posted Jun 10, 2016, 4:49 AM by Steve Borgatti   [ updated Jun 10, 2016, 4:51 AM ]

  • Made change in CLI to force use of courier new font. Previously, it was using the default Windows system font, which was almost always courier, but occasionally not. Using courier ensures that columns of numbers line up properly
[ download

Version 6.618 | 4 June 2016

posted Jun 5, 2016, 9:23 AM by Steve Borgatti   [ updated Jun 5, 2016, 9:24 AM ]

  • Added new command Data|Filter/extract|Extract Columns. Use it to pull out one or more columns from a dataset and put them in a new dataset.

Version 6.617 | 3 June 2016

posted Jun 3, 2016, 3:40 PM by Steve Borgatti

  • Fixed bug in Tools|Testing hypotheses|Node level|Regression that was causing the program to always choose the first column of the dependent variable dataset
  • Fixed bug in the matrix editor which was causing crashes on pressing the New button.

Version 6.616 | 30 May 2016

posted May 30, 2016, 3:14 PM by Steve Borgatti   [ updated Jun 1, 2016, 5:45 PM ]

  • Fixed bug in the CLI's Forfiles command. Previously, commands of the form ->forfiles samplk? draw were failing, while commands like ->forfiles samplk* $-cent = centrality($) were succeeding.
  • Note: bug discovered in this version in Tools|Testing hypotheses|Node-level|Regression. No matter which column you select as dependent variable, it always runs the regression on column 1

1-10 of 208