Versions
SCROLL DOWN TO SEE THE LATEST VERSIONS
Note
To find revision information for really old versions of UCINET look in these two places:
Related Links
To download the latest version, click download.
FYI. Google Chrome's version 86 (released 6 Oct 2020) introduced a security feature which blocks certain downloads. (Click here for more information.) This means that, as a user, you click on a download link, and nothing happens. No error message; nothing. It turns out they are doing this with exe files (like UCINET's installer packages) when the exe is stored on a non-secure website (such as we used to use).
This does not affect downloads of recent versions of UCINET, since these are now located on a secure server. However, it does affect downloads of old versions. If you want to download an old version, copy the link, paste it into the address bar of your browser, and change "http" to "https".
Version 6.791 | 17 Apr 2024
Changes
Fixed Excel button on the toolbar so that it can open not only UCINET system files but also Excel files
Version 6.790 | 10 Apr 2024
Changes
Changed CLI density command to include average degree on output.
Version 6.789 | 8 Apr 2024
Changes
Added additional measures to Network|Ego networks|Ego-alter Similarity|Categorical. New measures in include odds ratio and log odds ratio. Also, ego's category is now printed alongside the measures.
Added new CLI routine Zegers, which calculates four measures of association: identity
Version 6.788 | 3 Apr 2024
Changes
Added CLI routine Moca (measures of clustering adequacy). Good for assessing the quality of community detection partitions. This implementation does not have as many measures as the menu option Tools|Clustering|Measures of Cluster Adequacy.
Syntax: moca(<matrix> <partition>)
Usage:
->f = factions(campnet 3)
->dsp moca(campnet f)
Version 6.787 | 1 Apr 2024
Changes
Fixed recently-introduced bug in Network|Mixing/Density Tables, which made it impossible to use the "identity" and "complete" partition buttons on the dialog box
Updated fastgreedy algorithm to allow user to choose a clique-based starting partition
Added InitialPartition routine to the CLI. It runs a clique analysis and generates a partition from it. Intended as a starting place for agglomerative clustering algorithms such as fastgreedy. Syntax is: InitialPart(<network>), as in :
->ip = initialpart(drugnet)Added factor analysis based clustering algorithm to menu under Network|Community Detection|Factor Analysis Clustering. This already existed in the CLI as Facd.
Added routine for polishing community detection algorithms. Basically, it ensures that each node has more ties to its own kind than to any other group. In menu under Network|Community Detection|Polish Partitions.
Version 6.786 | 26 Mar 2024
Changes
Updated Network|Centrality|Induced Centrality to include more information on output
Fixed bug in Data|Header Browser. Previously, when changing a dataset's data type to "byte" the data was written incorrectly.
Version 6.785 | 11 Mar 2024
Changes
Added new procedure Tools|Cultural Consonance model. It implements Bill Dressler's cultural consonance model, and is used in combination with the Romney et al consensus analysis.
Version 6.784b | 8 Mar 2024
Changes
Replaced "pweakcomp" measure in Network|Egonets|Egonet Basic Measures with component ratio. Note: this routine still runs only on the 32-bit version
Fixed bug in NetDraw's egonet viewer that was causing up/down button to not work right
Version 6.784 | 4 Mar 2024
Changes
Changed the implementation of the FastGreedy approach to community detection. It is now faster with large datasets.
Also, added FastGreedy to the menu system under Network|Communities/Subgroups|FastGreedy. Previously, was only available in the CLI
Version 6.783 | 2 Mar 2024
Changes
Fixed bug in CLI routine LabelProp which is a clustering / community detection algorithm. The bug was causing crashes
Added Label Propagation routine to the menu as Network|Communities/Subgroups|Label Propagation. Note that this routine differs from the CLI version in that it automatically symmetrizes the network
Version 6.782 | 28 Feb 2024
Changes
Fixed bug in Data|Filter/Extract|Extract Matrices. Previously, it was mislabeling the output matrices (but the data was correct).
Fixed bug in Network|Communities/Subgroups|FastGreedy. It had been outputing the wrong Q score (but this did not affect the partition obtained).
Version 6.781 | 26 Feb 2024
Changes
Fixed bug in Network|Communities/Subgroups: the browse button ("...") wasn't working.
Version 6.780 | 23 Feb 2024
Changes
Added Data|Extract Matrices routine. Allows you to pull out selected matrices from a multi-relational dataset and save as a new dataset
Version 6.779 | 19 Feb 2024
Changes
Updated Netdraw (now version 2.188) so that the "attributes as coordinates" layout now automatically puts larger Y values at the top of the screen instead of the bottom (where Windows thinks they should be)
Fixed Data|Header Browser. It was screwing up when reading older ucinet files saved in "byte" format.
Version 6.778c | 14 Feb 2024
Changes
Updated Netdraw (now version 2.187) so that the Circle layout no longer moves the isolates
Version 6.778 | 4 Feb 2024
Changes
Fixed Network|Mixing/Density Tables. Previously was failing to label the rows and columns of the density matrix, including when the user provided them manually
Version 6.777 | 28 Jan 2024
Additions
Added Fast Greedy algorithm for community detection to the main menu (previously only available in the CLI). See Network|Communities/Subgroups|FastGreedy.
Added DivBy routine to the CLI. Use this to divide a matrix by another matrix, a vector, or a constant. Syntax is: divby(<matrix> <matrix>|<vector>|constant). Usage examples:
->events = xtx(davis) //creates event by event matrix
->normed = divby(events getdiag(events)) //divides each cell in events by the diagonal, which is largest value
->dsp normed // normed(i,j) gives proportion of event i's attendees to who also attended event j
Changes
Fixed display of datasets containing matrix names (as when there are multiple matrices in same dataset). A previous change had caused column labels to shift over.
Version 6.776 | 23 Jan 2024
Changes
Data|Random|Matrix changed to include row and column labels on output
Tools|Testing Hypotheses|Node level|Regression changed to accept datasets that don't have column labels
Visualize|Netdraw (quick) now asks you for the name of the dataset you want to draw
[bug] display of datasets containing matrix names (as when there are multiple matrices in same dataset) now includes matrix name above row labels column. unfortunately, this can cause column labels to shift over. will be corrected in next version
Version 6.775 | 2 Dec 2023
Changes
This installer includes an update to NetDraw, which includes changes like
printing to pdf no longer moves isolates to the side
rim thickness is respected when printing
resize/reposition button no longer move isolates to the side
Changed Data|Header Browser to allow user to type in filename instead of pressing "load" button
Version 6.774 | 28 Oct 2023
Changes
Made a temporary fix to Network|Community Detection - Subgroups|NCD. The way it was calculating which communities to merge was not quite in line with Newman's 2003 article "Fast algorithm for detecting community structure in networks". The new version is slow, but will be optimized later.
Version 6.773b | 2 Sep 2023
Changes
Added Properties|Nodes|Symbols|Shade option to NetDraw 1.83. This allows you to represent an ordinal attribute as shades of grey
Fixed bug in Properties|Lines|Size|Tie Strength which could sometimes yield incorrect results when the number of edges exceeded the number of nodes
Version 6.773 | 17 Aug 2023
Changes
Added classic multidimensional scaling routine to the menu. See Tools|Scaling/Decomposition|Classic MDS.
Version 6.772 | 13 Aug 2023
Changes
Added additional option to Tools|Scaling/Decomposition|Metric MDS. Now, you can choose a "classic MDS" solution (sometimes known as Torgerson-Gower scaling), as well as minimizing stress.
Added ClassicMDS command to the CLI to provide 2-dimensional classic scaling.
Description. Performs classic (Torgerson-Gower) multidimensional scaling. The input is a valued dissimilarity matrix. The output is a set of coordinates in 2-dimensional space. The ADD option adjusts the diagonal of the matrix so that it conforms to Euclidean distance. The METRIC option causes the program to iterate the classic solution to minimize stress in two dimensions.
Syntax: classicmds(<dissim matrix> [Add] [Metric])
Usage
->coord = classic(cities) //coord can then be viewed in tools|scatterplot
Version 6.771 | 30 July 2023
Changes
Replaced old Network|Centrality|Katz-Hubbell routine with new Network|Centrality|Katz family measures routine. It has more options than the old one, including the ability to generate Katz, Hubbell, Bonacich, Taylor and Friedkin scores. Also, the old one sometimes gave unnecessary "cannot invert" error messages.
Version 6.770 | 15 July 2023
Changes
Added quickclus routine to Network|Communities/Subgroups. Fast routine for optimizing q modularity. Allows user to input a starting partition
Added a routine to extract specific rows from matrix. Data|Filter/Extract|Extract rows
Version 6.769 | 2 June 2023
Changes
Fixed bug in Network|Mixing Table (same as Network|Whole Networks|Density|Density by Groups where, if you chose 'Match by labels' but had no labels, the program produced no error and no output
Version 6.768 | 28 May 2023
Changes
Added Options|Get Standard Datasets to the menu. Not needed in a normal installation, but useful in a few special situations. Normally, when you install UCINET, a folder called '\Ucinet Data' is created under the user's Documents folder to hold standard datasets often used in teaching network analysis. When this doesn't happen (or the user has deleted the folder), the Options|Get Standard Datasets procedure recreates the data folder from a copy the program keeps under \program files, where UCINET is installed.
Version 6.767 | 16 May 2023
Changes
Fixed the DL Editor. Previously, when entering an edgelist dataset, the program was recoding edge values of NA, even when the user typed them. Now, if you type a value of NA, it stays that way.
Previously, in Transform|Timestack, if you stacked matrices with different sets of nodes, the program would add rows and columns with missing data so that all matrices had the same rows and columns. Now, there is also an option to select only the rows/cols that were present in all time periods.
Version 6.766 | 8 Apr 2023
Changes
changed the default setting for the diagonal option in the CLI's IPF function to YES. This means the diagonal will be included in the calculations unless you turn it off (e.g., ->dsp ipf(campnet diag:no)
updated Data|Reshape|Simple command to work in 64-bit mode
added new Reshape menu option (Data|Reshape|Guided). It can do the following reshaping procedures:
matrices as additional rows. If a dataset contains multiple matrices, it can concatenate the matrices to create additional rows. So a dataset with 10 rows 3 columns and 2 matrices is converted into a dataset with 20 rows and 3 columns
each matrix becomes a row. If a dataset contains multiple matrices, it can string out each matrix into one long row. So a dataset with 10 rows 3 columns and 2 matrices is converted into a dataset with 2 rows and 30 columns
one long column. Converts to the dataset into a one long column. So a dataset with 10 rows 3 columns and 2 matrices is converted into a dataset with 60 rows and 3 columns
transpose rows and columns. Interchanges rows and columns for each matrix in a dataset. So a dataset with 10 rows 3 columns and 2 matrices is converted into a dataset with 3 rows, 10 columns and 2 matrices
transpose rows and levels. If a dataset contains multiple matrices, it interchanges the rows with the matrices (levels). So a dataset with 10 rows 3 columns and 2 matrices becomes a dataset with 2 rows, 3 columns and 10 matrices
transpose rows and columns. If a dataset contains multiple matrices, it interchanges the columns with the matrices (levels). So a dataset with 10 rows 3 columns and 2 matrices becomes a dataset with 10 rows, 2 columns and 3 matrices
(Netdraw 2.182). Fixed obscure bug in NetDraw. Previously, if you had a dataset with purely numeric node labels and used the Transform|Attribute editor, it would add spaces in front of the labels, making them appear far away from the node symbols
Version 6.765 | 06 Mar 2023
Changes
added nodelist capability to dl exporter
fixed bug in transform|transpose -- was crashing when trying to print the transposed matrix
added cyclicality function to menu -- Network|Whole networks|Cyclicality. This measures the tendency for networks to have tie k-->i when i-->j and j-->k are present.
(10 Mar 2023) fixed bug in DL Editor. The edgearray format was failing to symmetrize data even when "undirected" option had been selected
Version 6.764 | 12 Feb 2023
Changes
Fixed Data|Export|DL so that now, when it shows you the contents of the output file it created, it displays unicode characters correctly
Version 6.762 | 6 Feb 2023
Changes
Added 'copy multiple ucinet datasets' to the File menu. Simplifies copying ucinet datasets to a different folder
Version 6.761 | 28 Jan 2023
Changes
Added NetDraw version 2.180 to the installation package. This version of NetDraw fixes a bug in which VNA files did not record whether arrowheads should be visible or not
Added Mixing Tables option to Network menu. It is the same procedure that you can find under Network|Whole network|Density|Density by Groups. This makes it easier to find, especially if you are not familiar with the older term "density table"
Changed menu DL Editor to make clear that you can import CSV, Excel, and Txt files
Version 6.759 | 18 Nov 2022
Changes
Added option to Network|Egonets|Alter composition|Categorical. Now, it is possible to ignore ego's own category when calculating heterogeneity of alters.
Version 6.758 (updated) | 17 Oct 2022
Changes
Previously, there were some toolbar functions that were not available in the menu. Now, all are available. Version number was not changed.
Version 6.758 | 26 Sep 2022
Changes
The CLI removecols() function has been fixed -- previously, it was doing nothing. Note that unlike removenodes() it still does not accept column names, only numbers. However, it does accept ranges, as in 2-5. Usage examples:
->x = removecols(myattributes 2-7)
Version 6.757 | 15 Sep 2022
Changes
Previously, UCINET opened NetDraw in a maximized state. Now, it opens NetDraw in a "normal" state. However, there remains a bug where if you maximize the NetDraw window, the drawing window does not expand. Instead, you have stretch the window manually.
In the CLI, the aggbycols() function has been changed so that it sorts by values of the categorical attribute. Usage:
->x = aggbycols(
Version 6.756 | 3 Sep 2022
Fixes
Fixed the various UCINET procedures that take a UCINET file and open it in Excel. In cases were the UCINET had a blank or invalid label for the matrix name (think of the padgett dataset which contains two matrices, one labeled padgm and one labeled padgb), UCINET was creating an invalid Excel sheet name, causing a crash with a message about "OLE error".
Changes
The "savecsv' command in the CLI was changed to automatically add a column header for the row labels. As a result, typing:
->savecsv campattr
now yields this:
"ID","Gender","Role","leader"
"HOLLY",1,1,1
"BRAZEY",1,1,2
"CAROL",1,1,2
"PAM",1,1,2
"PAT",1,1,2
etc.
Whereas before, there was "" in the first line, instead of "ID". Note that the savcsv command automatically substitutes "_" for spaces in row and column labels. If this is not to your liking use the menu version instead: File|Clip to clipboard.
Version 6.755 | 14 Aug 2022
Changes
Made several changes to Data|Filter|Subset data by node/row list
The program now works on datasets containing 1-mode networks, and on datasets containing node attributes (where the rows are nodes). The idea is that if you want your network datasets and attribute datasets to all reflect just a certain subset of nodes, you can just cut and paste the node names into the box, and the input dataset will be filtered to only contain data on those nodes
Small changes in the user interface
Fixes
In the CLI, combined the previously separate zscore and stdize routines and fixed errors in how it read the command line syntax, which is zscore(<dataset> [dim:MATRIX|row|col] [diag:NO|ok]. Now you can type things like this:
->zwomen = zscore(women) //standardize entire dataset (except diagonals) to have mean 0 and sd 1
->zattribs = stdize(attr col) //stdize columns of attr. Note it is not necessary to right stdize(attr dim:col)
Version 6.753 | 20 June 2022
Changes
Added new routine Network|Whole-network|Homophily/Influence|Continuous to measure overall network autocorrelation for continuous attributes. Essentially, it measures the correlation between having a tie, and having similar attribute values
Added reciprocity routine to the CLI. Both arc and dyad reciprocity are computed, and for arc reciprocity, a comparison with density is also provided, as (in random graphs) the expected value of reciprocity is density. Syntax is reciprocity(<dataset>).
->dsp reciprocity(sampson)Added density comparison to the triplet transitivity routine in the CLI, as in a random graph, the expected value of triplet transitivity is density
Added access to the DL Editor from the CLI. Syntax: DLEDIT [<datasetname>]
->DLEDIT //opens the dl editor without reading in a file
->DLEDIT campnet //opens the dl editor and offers to open campnet in a variety of formats
Fixes
Fixed Network|Dyadic measures|Reachability routine. It was failing to run with larger datasets. In addition, it now runs in 64-bit mode as well as 32-bit.
Version 6.752 | 6 June 2022
Changes
Improved transitivity calculation in both the menu and the CLI. It is now much faster. Also, for very large networks (5000 nodes and 5 million edges), it was giving the wrong answers due to overflow. Now it is correct.
Version 6.751 | 5 June 2022
Changes
Added option in Tools|Testing Hypotheses|Dyadic|Logistic Regression to choose 1-tailed vs 2-tailed significance. Previously, only 1-tailed was available. Now, 2-tailed is the default
Version 6.750 | 1 June 2022
Changes
The windows for the DL Editor and the CLI have been modified so that the user cannot minimize them by pressing the minimize icon (it was hard to find them again once minimized). Instead of minimizing, the user should close the windows instead.
Added an option to Transform|Egotize so that alter-alter ties can be distinguished from ego-alter ties
Version 6.748 | 22 May 2022
Fixes
Two-mode centrality. The eigenvector routine in Network|Two-mode|Centrality was giving poor answers when the largest eigenvalue was repeated -- e.g., if the largest eigenvalue was not unique. However, even though the routine has been corrected to get the same results as R and Stata, we still recommend ignoring eigenvector centrality when the principal eigenvalue is repeated.
SVD. The SVD procedure in the CLI was updated to match the 2-mode centrality routine discussed above
Eigenvectors. The eigenvectors() function of the CLI was updated to give better answers in the case of repeated eigenvalues.
Hubs and authorities. The Network|Centrality|Hubs and Authorities routine was updated to give better answers when the principal singular value is repeated. In addition, the singular values are now printed.
Version 6.747 | 18 May 2022
Fixes
Data|Remove|Isolates. Fixed bug which caused it to treat missing values as ties. Fixed another bug which was causing it to print the original matrix, not the one with isolates removed.
Additions
Data for Learning Ucinet.xlsx. This excel file, which is automatically installed with UCINET, has been updated
New datasets Kracknet and Krackattr, and Karatenet and Karateattr added. They are just copies of krack-high-tec, high-tec-attributes, zackar and zachattr, respectively
Version 6.746 | 18 Apr 2022
Additions
Fragmentation centrality. Added fragcent function to the CLI. It is somewhat faster than other fragmentation centrality routines in ucinet, especially for undirected data.
Syntax: fragcent(<dataset> [method:delete|isolate])
->mycent = fragcent(campnet)
The method parameter chooses between isolating the node whose centrality is being measured (keeping the total number of nodes the same), or deleting the node (which means we compare a fragmentation score for a network with N nodes with a fragmentation score for a network with N-1) nodes. The default is delete.
Version 6.745 | 16 Apr 2022
Changes
Geodesic Distances. Upgraded geodesic distance routine used in several parts of UCINET to have limited parallel-computing capability.
Additions
CountValues. Added new CountValues function to the CLI, which counts the number of ties in a network. For valued data, you can specify what counts as a tie. Currently limited to 1-mode matrices. Syntax: Countv(<network> [op:GT|ge|eq|le|lt|ne] [<cutoff value>]). Examples:
->dsp countv(campnet) //counts # of ties in campnet
->dsp countv(newfrat op:le 3) //counts # of ties w/ tie strength <= 3 for each net in newfratSave to CSV. Added fast new SaveToCSV procedure to the CLI and also to the main menu. Should facilitate exporting files to R. Datasets without row/column labels are automatically provided with labels. Also, spaces within labels are recoded to underscores. The menu version can be found at Data|Export|CSV. The syntax of the CLI version is: ->SaveToCSV <ucinet dataset> [<filename.csv>]. If no output file name is provided, the program uses the dataset name together with a .csv extension. Example:
->savetocsv campnet
Version 6.744 | 8 Apr 2022
Additions
Clip function. Added new routine to both CLI and menu to simplify transferring UCINET data to other programs.
In the menu, choose 'Files|Copy UCINET dataset to windows clipboard'. It will ask the name of a UCINET dataset. When you press OK, the contents of the dataset will be copied to the Windows clipboard. You can then paste the data into Excel, Stata, or other programs.
In the CLI, the syntax is clip <dataset>, as in:
->clip padgett
QuickClus. Added new community detection routine to the CLI. It's a greedy algorithm for maximizing Q modularity ,and is intended to replace the slower Factions routine. An especially nice feature is that, optionally, the user can specify a starting partition from which to work from, allowing you to take the output of another community detection algorithm and improve upon it. The syntax is quickclus(<dataset> <# of clusters desired> [<starting partition>]), as in:
->clus = qclus(pv960 3)
->draw pv960 clus
->lou = louvain(zache)
->clus = qclus(zache 4 lou)K-Means. Added a k-means clustering routine to the CLI. Input is any valued proximity matrix (similarities or dissimilarities) along with a choice of number of clusters. The routine is suitable for large datasets. The syntax is kmeans(<matrix> <# of clusters desired> <sim|dis>). If 'sim' is entered, the program tries to put pairs of nodes with high values into the same cluster. If 'dis' is entered, the program puts nodes with low values in the same cluster.
Version 6.743 | 30 Mar 2022
Additions
Added new routine Network|Egonets|Egonet tie composition. Given multiple relations, the routine counts (for each node) how many ties the node has of each type. Results are reported both as frequencies and proportions. Heterogeneity scores are also provided.
Version 6.742 | 2 Mar 2022
Fixes
The Network|Whole network|Density|Density by Groups routine was ignoring the 'utilize diagonal' checkbox, so it always ignored the diagonal. (fixed earlier, bnly being reported now)
When printing the average geodesic distance statistic, the Transform|Rewire routine had been doubling the value.
When 'sum of cross-products' was chosen in the Data|Affiliations routine, it was delivering missing values when the sum of cross-products was zero.
Version 6.741 | 24 Feb 2022
Fixes
The Network|Roles & Positions|Regular Equivalence|Catrege routine was failing to read in a starting partition correctly. Now fixed. Note that Catrege is an older program that wants the input partition to be a row vector rather than a column.
Version 6.740 | 11 Feb 2022
Additions
(I believe it was this version that introduced this). Added new routine 'Network|Egonets|Egonet alter diversity'. The routine accepts multiple attributes, both categorical and continuous, and computes appropriate heterogeneity measures (IQV for categorical variables, and standard deviation for continuous variables).
Fixes
In the CLI, the divconst function (divide by constant) has been changed so that it will accept both a numeric value and a dataset containing a value. For example, you can do this:
->numties = sum(lltab)
->newmat = divconst(lltab numties)
In the past, you would have had to put in a numeric value (such as '70').
Version 6.739 | 02 Feb 2022
Additions
Fixes
includes updated NetDraw. Now you can highlight nodes and then turn labels on/off just for those nodes
Version 6.738 | 9 Jan 2022
Additions
Added 'bowtie' edge overlap measure to the CLI. Syntax is bowtie(<network> meth:binun|bindir|wtd), where binun indicates the network is binary and undirected, bindir signals binary directed, and wtd indicates a symmetric valued network. The outcome is a valued symmetric network where Xij = 0 for non-adjacent nodes (or adjacent nodes that have no friends in common). Example:
->bt = bowtie(campnet meth:bindir)
->dsp bt
Reference: https://arxiv.org/pdf/1712.07110.pdf
Version 6.737 | 6 Jan 2022
Fixes
The routine Network|Centrality|Closeness was giving incorrect values for the Valente-Forman closeness measure. The results given were a linear combination of the correct values, so would not affect a regression
Version 6.736 | 2 Jan 2022
Fixes
The routine Network|Roles & positions|Structural Equivalence|Concor|Standard previously did not print the blocked matrix if the network had more than 50 nodes. This has been upped to 250.
Version 6.735 | 2 Dec 2021
Fixes
The density by groups procedure (Network|Whole networks|Density|Density by groups) had a button that allowed the user to input a text file of group names (e.g., department names when creating a dept-by-dept density matrix). But this had a problem which was that after doing it once, you couldn't get rid of the group names. This was especially problematic when running a new density table that had more categories than the previously-used group names. This has now been fixed.
Other
The installer for this version includes an updated version of keyplayer1.exe, which corrects a problem with the printed fragmentation scores.
Version 6.734 - 16 Nov 2021
Fixes & Changes
In the DL editor, there had been a problem where clicking on the options on the right-hand side would sometimes turn a cell in the spreadsheet blue, and then it was impossible to paste into the cell. This no longer happens.
In the DL editor, a new data format has been added which is called '1-mode matrix'. This option assumes that the rows and columns are the same nodes. This is particularly useful when you have people who did not respond to a survey, so their row is missing. Since their name is present in the columns, the importer automatically creates a row of missing values for that node, ensuring that the matrix remains square.
Version 6.733 - 25 Sept 2021
Fixes
LR-QAP had been (mistakenly) treating missing values in the dependent variable as 1s. Now they are properly accounted for.
LR-QAP no longer gives a p-value for the overall model when the data have missing values
Additions
Changed the pseudo r-squared statistic in LR-QAP to the McFadden r-squared
Added additional columns of data to the history dataset that LR-QAP automatically creates
Version 6.732 | 1 September 2021
New features
Added more capability to the CLI recode command. Previously, you could only recode one value, such as changing 99 to Missing. Now you can make multiple changes. Syntax: recode(<dataset> oldvalue:newvalue [oldvalue2:newvalue2 ... ]. Note: instead of the colon you can use "=" or just a space. Example:
g = geo(symmet(campnet))
rg = recode(g 5:15 3:13) // recodes 5s to 15s and 3s to 13s
Download