Keyboard Shortcuts
Likes
- Ucinet
- Messages
Search
|
finding triads with clique analysis
Hello, ? For our project, we are trying to trace triads inside networks. Because 'triad census' doesn’t specify the nodes involved in triads (and we need to know them), we also use 'clique analysis' (minimum = 3). However, clique analysis doesn’t seem to have an option to put in a ‘maximum’. We are interested in cliques with 3 nodes, but in the output are also cliques with 4 or 5 nodes, and the 3-node structures within this nodes have to be dissected manually. Also, if a node is involved in a 5-node clique, it is also in a 4-node clique so that overlaps.? Is there a better way to dissect triads (with ‘node labels’), or can a maximum of 3-noded cliques be rendered in clique analysis somehow? ? Kind regards,
T. Cinjee & M. Martens Utrecht University |
|
Re: structural holes
开云体育Just go to Network|Ego networks|Structural holes. ? Most people use constraint. ? steve ? From: ucinet@... Sent: Monday, June 11, 2018 15:21 ? ? Dear Please, I have a project that includes hypotheses about the relationship between structural holes and performance performance. There are many ways to calculate structural holes such as effective size and constraint, which type is appropriate for the above hypothesis and why? Is there an application example showing how to calculate structural holes using UCINET? ? Thanks a lot for the help. ? ? .............................................................................. Dr. Amer ? |
|
structural holes
Dear Please, I have a project that includes hypotheses about the relationship between structural holes and performance performance. There are many ways to calculate structural holes such as effective size and constraint, which type is appropriate for the above hypothesis and why? Is there an application example showing how to calculate structural holes using UCINET? Thanks a lot for the help. ? .............................................................................. Dr. Amer |
|
Reminder: LINKS workshop on social network analysis, June 4-8, 2018 -- registration closing MAY 15
开云体育Apologies for coss-postings! ? The LINKS Center for social network analysis at the University of Kentucky is offering its annual 1-week summer workshop on social network analysis again this June 4-8, 2018 on the University of Kentucky campus.?Registration will be closing on May 15th. To sign up, visit?.? Students pay less than half-price. ? The workshop features LINKS Center resident faculty?Steve Borgatti, Dan Brass, Joe Ferrare, Eric Gladstone, Dan Halgin, Joe Labianca, Ajay Mehra, Andy Pilny?and?Scott Soltis, as well as guest instructors?Filip Agneessens, Ron Burt, Alan Daly, Rich DeJordy, Jeff Johnson, David Krackhardt, Brea Perry, Scott Poole?and?Tom Valente. In addition, all of the sessions are assisted by a cadre of students and recent graduates in order to ensure that participants are able to obtain personalized attention. We encourage participants to bring their own data and work with our 30+ instructors and TAs to analyze the data. ? The workshop features five major, concurrent tracks for the first four days, followed by four choices of area-specific sessions (communication, education, health, organizations) on the final day. At the end of each day, we also offer multiple 1.5 hour short modules on specialized topics, as well as a data lab for analyzing your own data. Finally, we offer a number of 45-minute 1-on-1 consultations with LINKS Center faculty to discuss your research. ? ? MAJOR TRACKS (4 days each; Monday-Thursday) ? Introduction to Social Network Analysis. Led by?Dan Halgin?and?Dan Brass?(U. of Kentucky). Includes a theoretical and empirical overview of the field, followed by a comprehensive survey of the concepts and methods of social network analysis, including data collection, data management, centrality, social capital, cohesion, and hypothesis testing. In addition, participants participate in lab sessions to learn how to use network analysis software, including UCINET and NetDraw. ? Analyzing Social Network Data. Led by?Rich DeJordy?(Fresno State) with special guest?David Krackhardt?(Carnegie Mellon). A more software- and data-oriented version of Introduction to SNA that covers most of the same concepts as Intro but focuses more on using the software and devotes a bit more time to interpreting the equations & formulas that define many network concepts. Includes lab sessions in which participants work through analysis exercises using the UCINET and NetDraw software. Participants are encouraged to use their own data in the exercises. ? Intermediate Network Analysis. Led by?Steve Borgatti?(U. of Kentucky). For people interested in both a deeper and broader look at network concepts and methods. Topics include advanced centrality methods?(including group centrality and edge centrality),?analyzing?network change, analyzing negative ties,?new clustering methods, working with multiple relations, and integrating node attributes with network measures. This module also introduces participants to UCINET's command-line and batch processing capabilities. ? Egocentric Network Analysis. Led by?Brea Perry?(Indiana U). An in-depth tutorial on collecting and analyzing personal networks. The course begins with an introduction to the foundational concepts of egocentric SNA, and moves on to research design considerations and data collection techniques. The rest of the course explains how to use standard statistical software to analyze egocentric network data using multi-level models. ? Stochastic Network Models. Led by?Filip Agneessens?(U. of Surrey). This course provides an introduction to exponential random graph models (ERGMs) and stochastic actor-based models for network dynamics (as implemented in rSIENA). The course will focus on hands-on use of MPNet and rSiena and on the interpretation of output. An introduction to the R analysis language (needed for RSiena) will also be provided during the workshop. ? ? AREA-SPECIFIC TRACKS (1 day each; Friday) ? Networks and Communication. Led by?Scott Poole?(Illinois) and?Andy Pilny?(U. of Kentucky). Applications of social network analysis?to the field of communication.? ? Networks and Education. Led by?Alan Daly?(UC-San Diego) and?Joe Ferrare?(U. of Kentucky). Applying the social network perspective to education research and policy. ? Networks and Health. Led by?Tom Valente?(USC). Social network theory and method in the context of understanding health-related behaviors, interventions and disease epidemiology. ? Networks and Organizations. Led by?Ron Burt?(Chicago). The session covers the latest research on networks and organizations. ? ? MINI-MODULES (1.5 hours each, Monday-Thursday after 4pm) ? The mini-modules are short sessions on specialized topics. Some focus on research design topics, others on using specialized software, and still others on handling particular kinds of data. Past offerings have included: managing your IRB, managing your research site, entering network data, working with various SNA packages in R, handling archival data, working with cognitive social structure data, working with 2-mode data, analyzing negative ties, egocentric data analysis, scaling of 1- and 2-mode data, ERGM for dummies, and more.? ? In addition, each day we offer a data lab where people can bring their data and have someone experienced with the software help them with the analysis. ? ? 1-ON-1 CONSULTATIONS (45-minute slots available Tuesday-Friday) ? We provide opportunities for participants to discuss their research in depth with?Joe Labianca,?Jeff Johnson,?Ajay Mehra, and?Scott Soltis. ? ________________________________ ? For more information, please visit the workshop website at?. ? Steve. ? Stephen P. Borgatti Paul Chellgren Endowed Chair and Professor Dept. of Management Gatton College of Business and Economics University of Kentucky ? |
|
Re: Missing data being used in analyses
Hi Allan, how did you create the new matrices without the non-participating nodes? I hardly managed to do this in Excel (my data format is a nodelist). Is there an elegant way in UCINET? Regards, Josh
On Friday, April 13, 2018, 1:03:58 AM GMT+2, 'Steve Borgatti' steve.borgatti@... [ucinet] wrote:
?
Hmm. You should be able to exclude missings but I see you cannot. I’ll have to fix that. In the meantime, yes, I would extract submatrices. ? From: ucinet@...
Sent: Wednesday, April 11, 2018 13:22 To: ucinet@... Subject: [UCINET] Missing data being used in analyses ? ? Hello- I'm working with a matrix of 360 people, of whom 290 provided data. The others did not provide data, but could be a target of ties from the participants. ? I'm running an Anova Density Model for categorical homophily of race (white vs. non-white). However, the nodes that did not participate are being included in the analysis, and their missing information for race is being recoded as 1 for use in the analysis. ? Is there a way to specify that missing data should not be included in the analysis at all? Or do I simply need to create new matrices without the non-participating nodes to get an accurate measure of homophily? ? Thanks for any advice on this. ? Best, Allan Clifton |
|
Re: Constraint scores in UCINET 6.649
开云体育HiYou are wrong the maximum for constraint is 1.125 and can only occur for nodes of degree 2. On 2 May 2018, at 21:20, Kamal Badar kamal.badar1980@... [ucinet] <ucinet@...> wrote:
|
|
Constraint scores in UCINET 6.649
Dear All, When I calculate structural holes measures using UCINET (Network>Ego Networks>Structural Holes), I observe constraint scores of more than 1. More specifically, for nodes of degree 2, whose alters are themselves connected, I find a score of score for constraint as 1.125. Where as I think it should be 1 (100%).? How can we use 1-constraint as our measure of structural holes in that case? Let me know where am I thinking wrong? Regards?? Dr. Kamal Badar Assistant Professor (TTS) HEC Approved Ph.D. Supervisor Institute of Management Sciences University of Balochistan Quetta, Pakistan.? |
|
Request for information
Good day Sir/Ma, Top of the day to you. I am a PhD study carrying out a research on effect of construction stakeholder's relationship on time and cost performance. I will like to know what to analyse output using UCINET 6....Characteristics such as desnsity of the network structure and what? I will very much appreciate responses to help me as I prepare for collection of network data soonest... Opeyemi Oyeyipo Research student? |
|
Re: Missing data being used in analyses
开云体育Hmm. You should be able to exclude missings but I see you cannot. I’ll have to fix that. In the meantime, yes, I would extract submatrices. ? From: ucinet@...
Sent: Wednesday, April 11, 2018 13:22 To: ucinet@... Subject: [UCINET] Missing data being used in analyses ? ? Hello- I'm working with a matrix of 360 people, of whom 290 provided data. The others did not provide data, but could be a target of ties from the participants. ? I'm running an Anova Density Model for categorical homophily of race (white vs. non-white). However, the nodes that did not participate are being included in the analysis, and their missing information for race is being recoded as 1 for use in the analysis. ? Is there a way to specify that missing data should not be included in the analysis at all? Or do I simply need to create new matrices without the non-participating nodes to get an accurate measure of homophily? ? Thanks for any advice on this. ? Best, Allan Clifton |
|
Re: Creating a list of pairs from spreadsheet
开云体育If I understand you right, this is very easy to do. Here’s a short video demonstrating: ? ? steve ? From: ucinet@...
Sent: Tuesday, April 10, 2018 1:47 To: ucinet@... Subject: [UCINET] Creating a list of pairs from spreadsheet ? ? Hi all -- ? this is a basic question, but I'm flummoxed and hope someone can connect me to a resource that will show me how to solve this problem: ? ?I have a spreadsheet.? Each line in the spreadsheet is a group of people who work together.? There are about 170 people.? The spreadsheet looks like ? ABCDEF AFG ADRMOX MOCB ASDBMPRW ? I want to know how many pairs are in here, because I think that is the best way to get UCINet/NetDraw to display the network.? (If I am wrong and I don't need this pairs list, I'd be happy to learn that, otherwise....)? There are too many names for me to count the pairs by hand.? Is there a better way than having a friend write a piece of code that will calculate this for me?? Or is there a simple way for me to generate a pair list on my own.. ? Also -- is there a resource other than this group that can help answer data questions such as these?? :) ? With kind regards, Christine |
|
Re: Comparing average strength of ties in ego networks of different size
开云体育Kamal, I’m not seeing a problem with 6 being the value you calculate for the ego with just one alter. It is the correct answer. However, consider this. Suppose you are going to regress an ego outcome of social support and other variables. For example, ? H = b0 + b1A + b2F + b3S ? Where A is age of ego, F is gender of ego, and S is the average social support of ego. When you run the regression in Stata or Spss, you can easily include network size as a case weight. This means that the regression will put weight on egos with more alters. ? You could also simply exclude egos with a single alter. ? Steve. ? From: ucinet@...
Sent: Saturday, April 7, 2018 19:51 To: ucinet@... Subject: [UCINET] Comparing average strength of ties in ego networks of different size ? ? Dear All, ? I am looking for a measure which compares ego networks of different size. Here is the scenario postulated step by step: ? 1)? The egos rate their ties (social support) with their alters on a 0-6 scale.? 2)? We are not interested to know the ties/strength of ties among alters. 3)? We then want to know the average strength of each egos' ties. Here is a problem: So if one ego is connected to to 3 alters with valued ties of say 3, 4, 6 which sums up to 13 and 13/3= 4.333 is the average.. Then there is another ego connected to only one alter with a valued tie of 6 and then 6/1 is 6 which doesn't make sense.? ? ? How to resolve this problem? Do we have a measure in UCINET VI to address this problem?? ? Regards? ?? ? -- ? Dr. Kamal Badar Assistant Professor (TTS) HEC Approved Ph.D. Supervisor Institute of Management Sciences Quetta, Pakistan.? |
|
2-day workshop on network analysis using ucinet April 27-28 in Philadelphia
开云体育This is geared to complete beginners. Offered under the auspices of www.statisticalhorizons.com. ? ? Introduction to Social Network Analysis taught by? April 27-28, 9:00 AM-5:00 PM ? This course provides an introduction to doing research on social networks. The course is very hands-on, emphasizing mastering the software and using the concepts and methods to answer research questions. It is very how-would-we-use-this oriented.? ? For more information,? ? Steve. ? Stephen P. Borgatti Paul Chellgren Endowed Chair and Professor Dept. of Management Gatton College of Business and Economics University of Kentucky steve.borgatti@... ? |
|
Missing data being used in analyses
Hello- I'm working with a matrix of 360 people, of whom 290 provided data. The others did not provide data, but could be a target of ties from the participants. I'm running an Anova Density Model for categorical homophily of race (white vs. non-white). However, the nodes that did not participate are being included in the analysis, and their missing information for race is being recoded as 1 for use in the analysis. Is there a way to specify that missing data should not be included in the analysis at all? Or do I simply need to create new matrices without the non-participating nodes to get an accurate measure of homophily? Thanks for any advice on this. Best, Allan Clifton |
|
Creating a list of pairs from spreadsheet
christine
Hi all -- this is a basic question, but I'm flummoxed and hope someone can connect me to a resource that will show me how to solve this problem: ?I have a spreadsheet.? Each line in the spreadsheet is a group of people who work together.? There are about 170 people.? The spreadsheet looks like ABCDEF AFG ADRMOX MOCB ASDBMPRW I want to know how many pairs are in here, because I think that is the best way to get UCINet/NetDraw to display the network.? (If I am wrong and I don't need this pairs list, I'd be happy to learn that, otherwise....)? There are too many names for me to count the pairs by hand.? Is there a better way than having a friend write a piece of code that will calculate this for me?? Or is there a simple way for me to generate a pair list on my own.. Also -- is there a resource other than this group that can help answer data questions such as these?? :) With kind regards, Christine |
|
Comparing average strength of ties in ego networks of different size
Dear All, I am looking for a measure which compares ego networks of different size. Here is the scenario postulated step by step: 1)? The egos rate their ties (social support) with their alters on a 0-6 scale.? 2)? We are not interested to know the ties/strength of ties among alters. 3)? We then want to know the average strength of each egos' ties. Here is a problem: So if one ego is connected to to 3 alters with valued ties of say 3, 4, 6 which sums up to 13 and 13/3= 4.333 is the average. Then there is another ego connected to only one alter with a valued tie of 6 and then 6/1 is 6 which doesn't make sense.? ? How to resolve this problem? Do we have a measure in UCINET VI to address this problem?? Regards? ??
Dr. Kamal Badar Assistant Professor (TTS) HEC Approved Ph.D. Supervisor Institute of Management Sciences University of Balochistan Quetta, Pakistan.? |
|
LINKS Center summer workshop on social network analysis, June 4-8, 2018
开云体育The LINKS Center at the University of Kentucky is offering its annual 1-week summer workshop on social network analysis again this June 4-8, 2018 on the University of Kentucky campus.?Registration is now open and will be closing on May 15th. To sign up, visit?.? Students pay half-price. The workshop features LINKS Center resident faculty?Steve Borgatti, Dan Brass, Joe Ferrare, Eric Gladstone, Dan Halgin, Joe Labianca, Ajay Mehra, Andy Pilny?and?Scott Soltis, as well as guest instructors?Filip Agneessens, Ron Burt, Alan Daly, Rich DeJordy, Jeff Johnson, David Krackhardt, Brea Perry, Scott Poole?and?Tom Valente. In addition, all of the sessions are assisted by a cadre of students and recent graduates in order to ensure that participants are able to obtain personalized attention. We encourage participants to bring their own data and work with our 20+ instructors and TAs to analyze the data. ? The workshop features five major, concurrent tracks for the first four days, followed by four choices of area-specific sessions (communication, education, health, organizations) on the final day. At the end of each day, we also offer multiple 1.5 hour short modules on specialized topics, as well as a data lab for analyzing your own data. Finally, we offer a number of 45-minute 1-on-1 consultations with LINKS Center faculty to discuss your research. ? ? MAJOR TRACKS (4 days each; Monday-Thursday) ? Introduction to Social Network Analysis. Led by?Dan Halgin?and?Dan Brass?(U. of Kentucky). Includes a theoretical and empirical overview of the field, followed by a comprehensive survey of the concepts and methods of social network analysis, including data collection, data management, centrality, social capital, cohesion, and hypothesis testing. In addition, participants participate in lab sessions to learn how to use network analysis software, including UCINET and NetDraw. ? Analyzing Social Network Data. Led by?Rich DeJordy?(Fresno State) with special guest?David Krackhardt?(Carnegie Mellon). A more software- and data-oriented version of Introduction to SNA that covers most of the same concepts as Intro but focuses more on using the software and devotes a bit more time to interpreting the equations & formulas that define many network concepts. Includes lab sessions in which participants work through analysis exercises using the UCINET and NetDraw software. Participants are encouraged to use their own data in the exercises. ? Intermediate Network Analysis. Led by?Steve Borgatti?(U. of Kentucky). For people interested in both a deeper and broader look at network concepts and methods. Topics include advanced centrality methods?(including group centrality and edge centrality),?analyzing?network change, analyzing negative ties,?new clustering methods, working with multiple relations, and integrating node attributes with network measures. This module also introduces participants to UCINET's command-line and batch processing capabilities. ? Egocentric Network Analysis. Led by?Brea Perry?(Indiana U). An in-depth tutorial on collecting and analyzing personal networks. The course begins with an introduction to the foundational concepts of egocentric SNA, and moves on to research design considerations and data collection techniques. The rest of the course explains how to use standard statistical software to analyze egocentric network data using multi-level models. ? Stochastic Network Models. Led by?Filip Agneessens?(U. of Surrey). This course provides an introduction to exponential random graph models (ERGMs) and stochastic actor-based models for network dynamics (as implemented in rSIENA). The course will focus on hands-on use of MPNet and rSiena and on the interpretation of output. An introduction to the R analysis language (needed for RSiena) will also be provided during the workshop. ? ? AREA-SPECIFIC TRACKS (1 day each; Friday) ? Networks and Communication. Led by?Scott Poole?(Illinois) and?Andy Pilny?(U. of Kentucky). Applications of social network analysis?to the field of communication.? ? Networks and Education. Led by?Alan Daly?(UC-San Diego) and?Joe Ferrare?(U. of Kentucky). Applying the social network perspective to education research and policy. ? Networks and Health. Led by?Tom Valente?(USC). Social network theory and method in the context of understanding health-related behaviors, interventions and disease epidemiology. ? Networks and Organizations. Led by?Ron Burt?(Chicago). The session covers the latest research on networks and organizations. ? ? MINI-MODULES (1.5 hours each, Monday-Thursday after 4pm) ? The mini-modules are short sessions on specialized topics. Some focus on research design topics, others on using specialized software, and still others on handling particular kinds of data. Past offerings have included: managing your IRB, managing your research site, entering network data, working with various SNA packages in R, handling archival data, working with cognitive social structure data, working with 2-mode data, analyzing negative ties, egocentric data analysis, scaling of 1- and 2-mode data, ERGM for dummies, and more.? ? In addition, each day we offer a data lab where people can bring their data and have someone experienced with the software help them with the analysis. ? ? 1-ON-1 CONSULTATIONS (45-minute slots available Tuesday-Friday) ? We provide opportunities for participants to discuss their research in depth with?Joe Labianca,?Jeff Johnson,?Ajay Mehra, and?Scott Soltis. ? ________________________________ ? For more information, please visit the workshop website at?. ? Steve. ? Stephen P. Borgatti Paul Chellgren Endowed Chair and Professor Dept. of Management Gatton College of Business and Economics University of Kentucky ? |
|
Re: Please help, Problem with my Ucinet
Anthony Ashiedu
Hello Paul i did just as you said and yet am still having some issues, can you please guide me through the steps you use in analyzing your network, thank you.
On Tuesday, February 27, 2018, 8:39:49 AM PST, Anthony Ashiedu wrote:
Hello Martin, Thanks for your reply, I have check this out and noticed I am still having the same issue, sure i have on the data some board members sitting in more than one firm I even have some board members sitting in (8) firms, Please Martin how do i get to draw this connection between this board so it can work well on my Ucinet, I have attached a copy of the data showing the connection and also list of members in more than one firm and the firms they occupy, also the other data 1 and data 2 shows what i have that is noting working well for me. Thanks for your input so far it been helpful, please use few minute to play around the data to see if am on the right track, I sincerely need some help here.
On Saturday, February 17, 2018, 7:10:04 PM PST, Martin meverett61@... [ucinet]
?
Hi I had a quick look at your data. The real issue is that there is no real network. You have 200 cols and 1,800 rows (roughly). But each of the cols has a unique set of rows. Ie there are no interlocking directorates. No row is attached to more than one column. To see this take your data and run tools univariate stats and just selected row sums. They are all one. The picture you would get in netdraw (and why it is tricky to get is they all look the same) is just each column as a node with the rows attached. IE a collection of stars with no connections between the cols hence no interlock. In short none of your board members are on more than one board so there is no real network. ? Martin ? Sent from for Windows 10 ? From: Anthony Ashiedu anthonyashiedu@... [ucinet]
Sent: 14 February 2018 05:02 To: ucinet@... Cc: Ucinet Moderator Subject: [UCINET] Please help, Problem with my Ucinet ? ? Hello Group, I am Anthony, am new in SNA I would like to apply the SNA to my project “Interlocking Directorate of board members” sitting in one or two firms, I need help to compute the number of directors interlocked in these firms by using company-director, director-director using 1- mode or 2-mode matrix form. I want to be able to determine the geodesic distance, the connectivity of board members, network centrality, and density. I have tried several times in using my data, but I keep getting wrong response from Ucinet especially on the graph layout and also it freezes without me seeing anything, am not sure if the problem is coming from my data, please I need help in this area since I am new with SNA. I have attached a copy of my raw data and data I uploaded on ucinet to this email if anyone could be of help here, many thanks in advance for your time. ? Kind Regards, ? Anthony.? ? ? |
|
Research help needed
Top of the day to you Sir/Ma,
My name is Opeyemi Oyeyipo, a PhD student from Covenant University, Ota, Nigeria. I am currently at the stage of eliciting network data for my research. I wanted to see an example of a network communication checklist to be used to elicit network data from construction project stakeholders in building project as well as how the the strength of network relationship can be determined and if the network can have impact on project performance.... These are needed in order to further my dissertation. I will be looking forward to hearing from you Sir.. Opeyemi Oyeyipo Research student |
|
Re: Bonacich - Memory Issue
开云体育You need to use Pajek for large networks. ? Martin ? Sent from for Windows 10 ? From: johnharvey86@... [ucinet]
Sent: 23 February 2018 23:37 To: ucinet@... Subject: [UCINET] Bonacich - Memory Issue ? ? Hi, ? I'm trying to calculate the Bonacich Power for nodes in a large directed network (~9k nodes + 54k edges).? ? It's a food sharing network that people often use when in poverty, so I'm in interested in trying to calculate dependencies (and therefore vulnerability) that may exist between people.? ? When I try to run the Bonacich algorithm in Ucinet my computer quickly runs out of memory.? As an alternative I tried on a high spec machine with 221GB, but still encountered?memory issues.? ? Has anyone managed to run the algorithm for a network of this size? And if so, do you have any advice? ? I'm relatively new to network analysis, so any help (or alternative recommendations and/or network measures) would be greatly appreciated.? ? Cheers John ? |