Download


Files
You can download either the 32-bit or 64-bit version of UCINET. The 32-bit version is the standard version and runs on both 32-bit and 64-bit Windows. The 64-bit version is limited in that it does not have all of the functions of the 32-bit version. It often crashes. Therefore, it is best used in tandem with the 32-bit version.

32-bit Installation Package. This installs the 32-bit version of UCINET along with several helper programs (such as NetDraw), and puts a copy of all the standard datasets in a folder called Ucinet Data under your Documents folder. It runs on 64-bit and 32-bit Windows. The installation program is new as of version 6.531 and unfortunately can't automatically uninstall versions of UCINET prior to 6.531, so you should do that yourself prior to installing current version.

64-bit installation package. This is NOT recommended. If you have a 64-bit version of Windows, you can try this 64-bit version of UCINET. It lets you analyze much larger datasets, if you have the RAM memory for it. However, it is also flaky. Older parts of UCINET are not compatible with 64-bit execution, so these parts will crash. 

Trial vs "real" version. Actually, there is no separate trial version. If you download the program, it will run on your computer for 60 days without having to enter a registration code. Buying the program gets you that registration code. So whether you buy or just try, you download the same program.

Virus/malware issues. There was a time when the download files were stored in a DropBox location. For some reason, this caused Norton AntiVirus to flag the program as dangerous. So we don't use DropBox for this anymore. But you might still run into issues with Norton or some other anti-virus program when installing UCINET. One thing to be aware of is that Windows 8 will issue a warning that the publisher is unknown and probably dangerous. This just means we haven't sought (read: paid for) Microsoft certification.


Installation Notes

Windows

Installation works best if you right-click on the installation file (typically called something like setup32UCI6534.exe) and choose 'Run as Administrator'. Similarly, it is also helpful to use 'Run as Administrator' when running UCINET itself for the first time, as this allows you to register the program for all users of your machine. See the FAQ on this.

MAC

The best way to run UCINET on a Mac is to use a Windows emulator such as Parallels (or, of course, Bootcamp). However, it is (often) possible to run UCINET on a Mac using Wine. For more information, see this FAQ

Version Info

  • Version 6.614 | 22 May 2016
    • Changed Network|Compare aggregate proximity matrices|partition to be able to handle missing values
    • Changed the CLI's IPF routine to default to treating diagonal values as valid. If you want the diagonal ignored you must add the word IGNORE to the command. E.g.,
      ->dsp ipf(padgm ignore)
     download 

    Posted May 22, 2016, 1:51 PM by Steve Borgatti
  • Version 6.613 | 18 May 2016
    • Added ability to sort columns in matrix editor via double-clicking column header
    • Added new version of node level regression under tools|testing hypotheses|node level|regression
    • Fixed bug in CLI version of regression which was multiplying classical p-values by df/(df-1). 
    • Fixed recently added bug in tools|testing hypotheses|dyadic|relational cross tabs which was printing incorrect correlation values
     download 

    Posted May 18, 2016, 7:38 PM by Steve Borgatti
  • Version 6.611 | 8 May 2016
    • Adds countcombinations routine to the CLI. Given a matrix of 1s and 0s, it counts up the number of combinations of column items found across all rows of the matrix. For example, if applied to the 2-mode Davis data, it would tell you how often each different combination of events occurs in the data. Syntax is CountComb(<dataset>). Example:

      ->dsp countcomb(davis)
    Frequencies
                                             1     2 
                                          Freq  Prop 
                                         ----- ----- 
        1     E2, E3, E4, E5, E6, E8, E9     1 0.056 
        2         E1, E3, E5, E6, E7, E8     2 0.111 
        3     E2, E4, E5, E6, E7, E8, E9     1 0.056 
        4                     E3, E4, E7     1 0.056 
        5                     E3, E5, E8     1 0.056 
        6                     E5, E6, E8     1 0.056 
        7                         E6, E9     1 0.056 
        8                     E5, E7, E8     1 0.056 
        9                E7, E8, E9, E12     1 0.056 
       10               E8, E9, E10, E12     1 0.056 
       11          E8, E9, E10, E13, E14     1 0.056 
       12      E7, E8, E9, E10, E12, E14     1 0.056 
       13 E6, E7, E9, E10, E11, E12, E13     1 0.056 
       14          E7, E8, E10, E11, E12     1 0.056 
       15                         E8, E9     1 0.056 
       16                        E9, E11     2 0.111 

    In these data, only two combinations of events occur more than once. 


    Posted May 8, 2016, 1:22 PM by Steve Borgatti
  • Version 6.610 | 25 April 2016
    • Fixed bug in matrix editor that was causing it to save matrices with more rows and columns than were actually present
    • (Also, a bug was fixed a few versions ago that was causing "argument out of range" errors, but the fix was not previously reported)
    Update 7 May 
    • Updated the installer to include latest version of NetDraw (2.159), which fixes a bug in reading 2-mode data

     download 

    Posted May 7, 2016, 1:39 PM by Steve Borgatti
  • Version 6.609 | 18 Apr 2016
    • Made change in Tools|Cluster|Hierarchical Clustering. Now, when the user chooses "Newman community detection", the program symmetrizes the data via the maximum method, and also dichotomizes. 

    Posted Apr 18, 2016, 2:14 PM by Steve Borgatti
  • Version 6.608 | 16 Apr 2016
    • Fixed bug in Network|Cohesion|Density|Density by Groups, which was failing to match the attribute dataset with the network dataset when the nodes in the two datasets were in different order. 
     download 

    Posted Apr 16, 2016, 8:04 AM by Steve Borgatti
  • Version 6.607 | 11 Mar 2016
    • Fixed bug in CLI's ForFiles command to allow handling files that are not ucinet system files. For example, suppose you want to convert several excel files to ucinet. If the excel files are all named with the same prefix (such as 'team1.xlsx', 'team2.xlsx', etc, you can type:

      ->forfiles team*.xlsx $$ = loadexcel($$)
      ->dsp centrality(team1)

      Previously, UCINET assumed that the files represented by $$ were ucinet files. 
     download 

    Posted Apr 11, 2016, 4:09 PM by Steve Borgatti
  • Version 6.606 | 4 Apr 2016
    • Added option to the Regress routine the CLI. Now, you can add YPERM as the last parameter; the program will generate significance using a Y permutation test on the t-statistics. Syntax is ->regress <yvar> <xvar1> <xvar2> ... [yperm]. For example,

      ->regress betweenness(campnet) cols(campattr "gender" "role"), yperm

      Regression statistics for 1

                             Value 
                         --------- 
                    Nobs        18 
                R-Square   0.08894 
            Adj R-square  -0.03254 
                  F(0,0)   0.73213 
                Prob > F   0.49730 


      Regression coefficients

                           Coef        SE         T       Sig 
                      --------- --------- --------- --------- 
            Intercept   9.62500  21.48261   0.44804   0.70440 
               Gender  -7.62500  13.39676  -0.56917   0.29040 
                 Role  19.37500  16.01219   1.21002   0.13500 


      Note, the results are not very different from the classical significance test (recalling that permutation tests are conducted 1-tailed, so the classical p-values should be halved for comparison):

      ->regress betweenness(campnet) cols(campattr "gender" "role")

      Regression statistics for 1

                             Value 
                         --------- 
                    Nobs        18 
                R-Square   0.08894 
            Adj R-square  -0.03254 
                  F(0,0)   0.73213 
                Prob > F   0.49730 


      Regression coefficients

                           Coef        SE         T       Sig 
                      --------- --------- --------- --------- 
            Intercept   9.62500  21.48261   0.44804   0.65978 
               Gender  -7.62500  13.39676  -0.56917   0.57669 
                 Role  19.37500  16.01219   1.21002   0.24283 




    Posted Apr 4, 2016, 5:28 PM by Steve Borgatti
Showing posts 1 - 8 of 197. View more »

Comments