Keyboard Shortcuts
Likes
- Ucinet
- Messages
Search
|
Re: network density
Many thanks for your useful answer.? .............................................................................. Dr. Amer Ali Al-Atwi
On Sunday, May 12, 2019, 12:37:26 AM GMT+3, 'Steve Borgatti' steve.borgatti@... [ucinet] wrote:
?
There are two ways one could define density with valued data. One way is to average all values. In this case, density is really average tie strength, and the values would not, in your case, be between 0 and 1. The other way is to normalize this average by dividing by the largest tie strength (in your case, 4). That would guarantee that the values would be between 0 and 1, and reflect the idea that in principle, everyone in the group could be connected to every other at a strength level of 4. Ucinet only does the first way, but you can take the result and divide by 4. ? From: ucinet@...
Sent: Friday, May 10, 2019 10:42 To: ucinet@... Subject: [UCINET] network density ? ? Dear UCINET group, ? How can I calculate network density when group members are rated on scale : 4 = always, 3=very often, 2=sometimes, 1=seldom, and 0 = never ? Does network density remain its range between 0 and 1??? ? For example, when a group includes 4 members (A, B, C, and D) and their rating are : A= 0, B= 2, C= 3, D= 4.? ? network density = Actual Connection/?Potential Connection? ? What will be the density value according to example above? ? Kind regards, ? Amer? |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: network density
开云体育There are two ways one could define density with valued data. One way is to average all values. In this case, density is really average tie strength, and the values would not, in your case, be between 0 and 1. The other way is to normalize this average by dividing by the largest tie strength (in your case, 4). That would guarantee that the values would be between 0 and 1, and reflect the idea that in principle, everyone in the group could be connected to every other at a strength level of 4. Ucinet only does the first way, but you can take the result and divide by 4. ? From: ucinet@...
Sent: Friday, May 10, 2019 10:42 To: ucinet@... Subject: [UCINET] network density ? ? Dear UCINET group, ? How can I calculate network density when group members are rated on scale : 4 = always, 3=very often, 2=sometimes, 1=seldom, and 0 = never ? Does network density remain its range between 0 and 1??? ? For example, when a group includes 4 members (A, B, C, and D) and their rating are : A= 0, B= 2, C= 3, D= 4.? ? network density = Actual Connection/?Potential Connection? ? What will be the density value according to example above? ? Kind regards, ? Amer? |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
network density
Dear UCINET group, How can I calculate network density when group members are rated on scale : 4 = always, 3=very often, 2=sometimes, 1=seldom, and 0 = never Does network density remain its range between 0 and 1??? For example, when a group includes 4 members (A, B, C, and D) and their rating are : A= 0, B= 2, C= 3, D= 4.? network density = Actual Connection/?Potential Connection? What will be the density value according to example above? Kind regards, Amer? |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Chi-Square Test for more than two groups with numerical variance less than 5
Hello everyone, I would like to know if there is a Ucinet alternative to Chi-Square to compare data from three groups of countries. ?? Some of the numerical values in the data are less than 5.? Data points range from 700 to 1800.? I would also like some advice on how many data permutations the computation process would support with such number of data points. ? The recommended one thousand might be hard on the equipment in this case. Regards, Xanat V. Meza Ph.D. Kansei, Behavioral and Brain Sciences University of Tsukuba M.A. Media and Communication Yeungnam University B.D. Graphic Communication Design Universidad Autonoma Metropolitana |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: 2-Step Centrality
开云体育Maybe there is a setting on your gmail that would allow you to receive images. ? In the current version of ucinet (6.677, and possibly earlier – I seem to have failed to record the change), you can go to Data|Select/filter|Egonet and tell it which node or nodes you want to be ego, and how many steps out from ego you want to allow, and it will produce a reduced network that is just ego(s) and everyone within k steps of these egos. ? So for campnet I chose node 2 (Brazey) as the ego and chose 2 steps as the distance from ego. I got the below. Note that x(u,v) = 1 means u can reach v in 2 steps. Note also that brazey can reach russ in 2, but not the other way around, ? CONSTRUCT NEIGHBORHOOD OF EGO -------------------------------------------------------------------------------- ? Input dataset:????????????????????????? campnet (C:\Users\Steve Borgatti\Documents\UCINET data\campnet) Focal nodes:??????????????????????????? 2 Number of steps from ego:?????????????? 2 Include focal nodes???????????????????? YES Output dataset:???????????????????????? campnet-Ego ? This egonet has 5 nodes. It comprises 27.8% of all nodes in the network. ? ?????????????????? 2??? 11??? 16??? 17??? 18 ?????????????? BRAZE?? LEE STEVE? BERT? RUSS ?????????????? ----- ----- ----- ----- ----- ??? 2 ?BRAZEY? 1.000 1.000 1.000 1.000 1.000 ?? 11???? LEE? 1.000 1.000 1.000 1.000 1.000 ?? 16?? STEVE? 1.000 1.000 1.000 1.000 1.000 ?? 17??? BERT? 1.000 1.000 1.000 1.000 1.000 ?? 18??? RUSS? 0.000 1.000 1.000 1.000 1.000 ? Output egonet saved as dataset campnet-Ego ? ---------------------------------------- Running time:? 00:00:01 Output generated:? 06 May 19 10:58:18 ? ? From: ucinet@...
Sent: Tuesday, April 30, 2019 17:31 To: ucinet@... Subject: RE: [UCINET] 2-Step Centrality ? ? Yes, that worked well and was very clear although the example image did not appear in my browser, the text sufficed. ? However, my challenge is that I have 167 nodes with 26,438 ties (yes, it is dense and highly connected) and picking one ego and using at a value of 2 did not reduce the cloud. ? ? |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: Is permutations t-test in UCINET affected by the size of the categories
开云体育Well, it is not an artifact of group size. The t-test is about means, so it doesn’t matter whether one group is larger than the other – the smaller group could have the larger mean. BUT … there could be sociological confounds. Suppose people are homophilous, so that positive people send ties to positive people, and negative to negative. Then of course the large positive group will be more central than the small negative group on most centrality measures. ? steve ? From: ucinet@...
Sent: Tuesday, April 30, 2019 10:28 To: ucinet@... Subject: [UCINET] Is permutations t-test in UCINET affected by the size of the categories ? ? Hi, ? I have a network where 75% of nodes have a certain attribute (let's say positive) and 25% of nodes are negative. If I run permutations t-tests in UCINET to test the difference in the means of various centrality measures (ex: closeness centrality). Could the results of the t-tests be an artifact because of the difference in size of the two categories. In other words, could positive nodes statistically have higher mean closeness centrality because they are more prevalent than the negative nodes? ? Help is much appreciated Best, Deena? ? ? ? Deena Abul Fottouh, PhD. Postdoctoral Research Fellow Social Media Lab Ryerson University |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: 2-Step Centrality
Yes, that worked well and was very clear although the example image did not appear in my browser, the text sufficed. However, my challenge is that I have 167 nodes with 26,438 ties (yes, it is dense and highly connected) and picking one ego and using at a value of 2 did not reduce the cloud. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Is permutations t-test in UCINET affected by the size of the categories
Hi, I have a network where 75% of nodes have a certain attribute (let's say positive) and 25% of nodes are negative. If I run permutations t-tests in UCINET to test the difference in the means of various centrality measures (ex: closeness centrality). Could the results of the t-tests be an artifact because of the difference in size of the two categories. In other words, could positive nodes statistically have higher mean closeness centrality because they are more prevalent than the negative nodes? Help is much appreciated Best, Deena? Postdoctoral Research Fellow Social Media Lab Ryerson University |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: 2-Step Cenrality
开云体育One way is via netdraw. For example, open the campnet dataset in netdraw, and then choose “ego” from the toolbar. Select Holly as the node, and then set the distances to and from Holly to 2. Like this: ? ? From: ucinet@...
Sent: Saturday, April 27, 2019 18:14 To: ucinet@... Subject: [UCINET] 2-Step Cenrality ? ? In one of my datasets I have an actor that does not score high on eigenvector or degree, but very high on 2-step centrality. What would be the best way to quickly identify their 1 and 2-step connections or am I asking the wrong question? Thanks. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Best Way to Import Ego-Alter-Alter Data
开云体育Hello All: ? While I know I have done it before, for the life of me I am forgetting to how to important – or what the best import strategy would be for data that is structured as EGO-ALTERA-ALTERB-ALTERAB Relationship. Initially, I was thinking it would be an edgelist but because of the Alter Relationship Data this does not seem to be the best option, and creates some rather interesting measurement outputs. ? The data is structured as: ?
? ? The issue I am running into is how to best import and maintain the alter relationships. Would it be through a matrix with multiple sheets? ? Any assistance would be appreciated! I am just blanking on the best method right now. ? ? M. Aaron Guest ? |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
UCINET Workshop: One week left to pre-register
开云体育
California State University, Fresno is offering a 3.5 day workshop, "Introduction to Analyzing Social Networks Using UCINET" from May? 28-31 (the week after Memorial Day).
In addition to a Keynote address by Steve Borgatti, the workshop covers:
Each topic pairs a content module with a hands-on Lab exercise to allow participants to develop both the analytic and software skills to perform social network analysis.
Reserve your spot, by pre-registering before May 1.
More details are available at:?
Rich DeJordy
Department of Management
Craig School of Business
California State University, Fresno
dejody@...
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Relational Contingency Table Analysis Null Hypothesis
Hello Fellow SNAers! I'm a humanities scholar who is working with networks (and indeed statistics) for the first time, so please forgive what are likely to be silly questions. I am tracking the transnational networks of memory activists. I've run a RCT analysis on my networks using UCInet (Tools>Testing Hypotheses>Mixed Dyadic/Nodal>Categorical Attributes>Relational Contingency Table Analysis) using attribute data for regional location of the actor. I want to test if particular regions group in a particular way. I'm unsure how to formulate the null hypothesis for the test; would it be something along the lines of: "The regional location of the actor shows no correlation with the regional location of the actors with which it forms ties"? Presuming that I get a significant result (p<0.05) and can reject the null hypothesis (which is the case), I can then interpret what the effect is from the Observed/Expected tables - i.e., I can show where actors are grouped more often than would be expected in a random distribution. Is that right? Thanks in advance! Sara |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: Matrix value recoding
开云体育As I understand it, you have an adjacency matrix that 2k by 2k. You have physical distance matrix that is 2k by 2k. You want to create a copy of the adjacency matrix with one change: all pairs of nodes within 50 units of distance of each other will be assigned adjacency 1. To do this, you can use transform|matrix ops|between|Boolean combinations. See this video ? ? steve ? From: ucinet@...
Sent: Wednesday, April 17, 2019 16:32 To: ucinet@... Subject: [UCINET] Matrix value recoding ? ? Dear Ucinet users, |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Larger Dataset Crashes
Most of my datasets so far have been in the 50-100 node range. However, when trying a larger dataset consisting of 2147 cases and 849 affiliations (all binary) UCINET crashes with different error messages. The procedure is Network->2-Mode Networks->2-Mode Centrality (both rows and columns). With UCINET 6.672 the program goes about 80% (based on green bar) fairly quickly and does create the files (1K) but they are empty due to the crash which is: Problem signature: After installing UICNET 6.676 I received a different error message with no files being created, i.e., Access violation at address 3130302E in module 'Uci6.exe'. Read of address 3130302E. The only thing I can think of is the number of affiliations that are positive for only one node. Any other ideas would be appreciated. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Re: Export to Pajek Net File
开云体育I’m not sure what the problem was, but I think I fixed it. In version 6.676, on the web now. ? steve ? From: ucinet@... <ucinet@...>
Sent: Saturday, April 6, 2019 17:14 To: ucinet@... Subject: [UCINET] Export to Pajek Net File ? ? I'm having problems with exporting my data (116 symmetric) file to Pajek net format. I can easily do the task with Southern Women dataset or Knokburr, but UCINET crashes whenever it tries my dataset. Displaying the data shows it is a well structured file in UCINET. Any ideas wuld be welcome.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Matrix value recoding
Dear Ucinet users,
I currently have a 2k*2k network. I wanted to change the adjacent matrix value based on the distance between any two nodes. For example, replace e(I, J)=1 if d(i, j)<50, where e(I,j) is the edge between nodes I and j, d(I,j) is the geographical distance between two nodes I and j. One method I could use is to creat another adjacent matrix based on the distance, then take a matrix operation (add). The current problem is dL matrix edit only allows 30k rows, where the days distance has 2k*2k/2 rows. I can do all matrix operations using r or matlab. However, how I can import a full matrix back to Ucinet for network metric measure? Ideally,it would great if I could complete all operations in Ucinet without import/export data back and forth. I would appreciate any assistance along this lines. Thank you in advance and look forward to hearing from you! YIng |