Re: Questions on node-weighted centrality or hybrid centrality

Dear Prof. Everett (and all),

?

After reading the sections recommended again, I agree with that it would be clearer in concept if we only weight either edges or nodes rather than both of them. Otherwise, if we get a large score, we cannot determine whether this is because of the strong tie strength or a high level of node attribute (e.g., wealth). In addition to this, I have three follow-up questions.

?

1. From the perspective mentioned above, it seems that we had better use binary network when calculating PN centrality, though Everett and Borgatti (2014) indicates that the algorithm can be applied to directed valued network and the xPNCentrality also supports. Is this understanding right?

?

2. For measures of alter composition, the current content in the book is mainly about degree, which focuses on alters directly connected with the ego. I think maybe it is not easy to depict the composition of alters which the ego indirectly connected with. This is because we cannot confirm the scope of the indirect ties that we need to include in. After all, we can reach everyone in the world through six steps (Milgram, 1967). As such, if we include the nodes which the ego indirectly connected with when calculating composition of alter, probably the variance of this variable would be small because every ego can reach all the others ultimately. Therefore, I think we cannot test idea like: Beta centrality × Composition of alters accounted for calculating Beta → Dependent variable. And this conclusion can be applied to all the centrality measures accounting for indirect ties like Eigenvector, Beta, k-step reach, Beta reach. Is this understanding right?

?

3. However, I am not sure whether we can calculate the composition of alters involved in calculating closeness and betweenness, which do not involve indirect ties (if my understanding is correct). I am especially interested in the latter because closeness “is of limited use as a centrality measure” (Borgatti et al., 2022). And if yes, how shall we determine the alters and then calculate the composition of them?

?

I look forward to hearing from you. Thanks again in advance!

?

Best,

Chuding

meverett61 via <meverett61=[email protected]> 于2024年8月11日周日 19:19写道：

Show quoted text

Hi
See my answers below.

Martin

On Saturday 10 August 2024 at 07:31:23 BST, CHUDING LING <lingchuding@...> wrote:

Dear Prof. Borgatti and all,

?

I am a reader of your book “Analyzing social networks using R” (Borgatti et al., 2022). Recently I notice that there is a stream of literature called node-weighted centrality or hybrid centrality (Singh et al., 2020; Singh, 2022). This literature points out that previous measures of centrality account for weight on edges (i.e., valued network) or length (e.g., eigenvector or beta centrality). However, these previous measures treat nodes as uniform. But it is very likely that different alters which an ego connects with are different in terms of a particular attribute (i.e., has different levels of importance). Let me use an example from Singh et al. (2020) to further illustrate:

?

For a friendship network, nodes represent persons and edges represent the friendship relationship between the considered set of persons. Here, weights on the nodes can be understood as a mapping of wealth, power, education level, or some other attribute of persons. It is notable that existence of two persons with identical attributes is highly unlikely and therefore all person’s attributes can be mapped to different real values based on the application specific mapping.

?

Therefore, my first question is whether we can account for such weights of nodes when calculating different types of centrality. Of note, I have quickly gone through the literature on node-weighted centrality or hybrid centrality, including 18 methodological or statistical research cited by Singh and colleagues (Singh et al., 2020; Singh, 2022), though I do not fully understand the mathematical parts. However, it seems that all of them use another type of centrality as the weight when calculate a particular type of centrality. For example, they use betweenness centrality as the weight when calculating degree centrality. From this perspective, your work on PN centrality might be also understood as a hybrid centrality (Everett & Borgatti, 2014). Particularly, in Everett and Borgatti (2014), you state:

?

Note that this equation weights everyone the same, but in many cases we might prefer a measure in which actors with high centrality affect the centrality score of the actors they are connected to more than those with a low score. In other words, we might want to capture the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.

?

Nevertheless, all the literature aforementioned do not fully capture the notion in the example from Singh et al. (2020) that I cited. That is, they use relational attributes to obtain the weight. Therefore, my second question is whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

?

Related to the PN centrality in Everett and Borgatti (2014), I have the third question. That is, whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality. According to your example in Chapter 2 of Borgatti et al. (2022), maybe we can indicate the first matrix as the negative network and the second matrix as the positive network, then the product of the two networks indicates the extent to which an ego is negatively connected to alters who are positively connected to others. Therefore, conceptually it may also capture “the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.” However, after experimenting with the example data provided in PN centrally section, I find the results are significantly different. Nevertheless, my fourth question is whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

?

To recap, I list the four questions as follows:

?

1. Whether we can account for weights of nodes when calculating different types of centrality, especially in population software like R.

First I refer to the R book so you can read what we say. I suggest you look at section 8.4.2. In the book we deal with degree and weighted attributes and these are in the chapter about node level measures. We do not (mainly through lack of space) discuss other centrality measures. There would be a vast array of possible measures. If the edges do not have weights the simplest way is to use the node attribute weight on the edge. If the edges do have weights then some method of combining node attribute weight and edge weight would need to be done. What you do needs to depend on exactly what is being measured. For example if you are in an organisation and the edges are number of times an advice interaction takes place in a week and the node attributes are tenure you could take the product but this assumes this is meaningful in some way as these are two very different quantities.
In summary this is about research questions and data so yes it is possible but personally to enable easyy interpretation I would stick with the degree type example we have in the book.??

?

2. Whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

Yes this is done all the time

?

3. Whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality.

Yes this is possible but again it is about interpretation. See section 2.6. After multiplication the network you have is a count of composite relations. Suppose P is friend and N is enemy then the entries of PN(i,j) count the number of i's friends who have j as an enemy. If on PN we ran weighted degree centrality then the raw score would be a count of the total number of friends who have enemies relation. Note NP is different to PN.

?

4. Whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

No this is really covered in the PN answer above as this is the same thing in essence.

?

I look forward to any suggestions from all of you in the list. Thanks in advance!

?

Best,

Chuding

?

References

?

Borgatti, S., Everett, M., Johnson, J., & Agneessens, F. (2022). Analyzing social networks using R. SAGE Publications.

?

Everett, M. G., & Borgatti, S. P. (2014). Networks containing negative ties. Social Networks, 38, 111–120.

?

Singh, A., Singh, R. R., & Iyengar, S. R. S. (2020). Node-weighted centrality: A new way of centrality hybridization. Computational Social Networks, 7, 6.

?

Singh, R. R. (2022). Centrality measures: A tool to identify key actors in social networks. In A. Biswas, R. Patgiri, & B. Biswas (Eds.), Principles of Social Networking: The New Horizon and Emerging Challenges (pp. 1–27). Springer.

Re: Questions on node-weighted centrality or hybrid centrality

#7063

Dear Prof. Everett,

Thanks for your prompt and detailed responses! I will read the sections you recommend again.

?

Best,

Chuding

meverett61 via <meverett61=[email protected]> 于2024年8月11日周日 19:19写道：

Show quoted text

Hi
See my answers below.

Martin

On Saturday 10 August 2024 at 07:31:23 BST, CHUDING LING <lingchuding@...> wrote:

Dear Prof. Borgatti and all,

?

I am a reader of your book “Analyzing social networks using R” (Borgatti et al., 2022). Recently I notice that there is a stream of literature called node-weighted centrality or hybrid centrality (Singh et al., 2020; Singh, 2022). This literature points out that previous measures of centrality account for weight on edges (i.e., valued network) or length (e.g., eigenvector or beta centrality). However, these previous measures treat nodes as uniform. But it is very likely that different alters which an ego connects with are different in terms of a particular attribute (i.e., has different levels of importance). Let me use an example from Singh et al. (2020) to further illustrate:

?

For a friendship network, nodes represent persons and edges represent the friendship relationship between the considered set of persons. Here, weights on the nodes can be understood as a mapping of wealth, power, education level, or some other attribute of persons. It is notable that existence of two persons with identical attributes is highly unlikely and therefore all person’s attributes can be mapped to different real values based on the application specific mapping.

?

Therefore, my first question is whether we can account for such weights of nodes when calculating different types of centrality. Of note, I have quickly gone through the literature on node-weighted centrality or hybrid centrality, including 18 methodological or statistical research cited by Singh and colleagues (Singh et al., 2020; Singh, 2022), though I do not fully understand the mathematical parts. However, it seems that all of them use another type of centrality as the weight when calculate a particular type of centrality. For example, they use betweenness centrality as the weight when calculating degree centrality. From this perspective, your work on PN centrality might be also understood as a hybrid centrality (Everett & Borgatti, 2014). Particularly, in Everett and Borgatti (2014), you state:

?

Note that this equation weights everyone the same, but in many cases we might prefer a measure in which actors with high centrality affect the centrality score of the actors they are connected to more than those with a low score. In other words, we might want to capture the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.

?

Nevertheless, all the literature aforementioned do not fully capture the notion in the example from Singh et al. (2020) that I cited. That is, they use relational attributes to obtain the weight. Therefore, my second question is whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

?

Related to the PN centrality in Everett and Borgatti (2014), I have the third question. That is, whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality. According to your example in Chapter 2 of Borgatti et al. (2022), maybe we can indicate the first matrix as the negative network and the second matrix as the positive network, then the product of the two networks indicates the extent to which an ego is negatively connected to alters who are positively connected to others. Therefore, conceptually it may also capture “the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.” However, after experimenting with the example data provided in PN centrally section, I find the results are significantly different. Nevertheless, my fourth question is whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

?

To recap, I list the four questions as follows:

?

1. Whether we can account for weights of nodes when calculating different types of centrality, especially in population software like R.

First I refer to the R book so you can read what we say. I suggest you look at section 8.4.2. In the book we deal with degree and weighted attributes and these are in the chapter about node level measures. We do not (mainly through lack of space) discuss other centrality measures. There would be a vast array of possible measures. If the edges do not have weights the simplest way is to use the node attribute weight on the edge. If the edges do have weights then some method of combining node attribute weight and edge weight would need to be done. What you do needs to depend on exactly what is being measured. For example if you are in an organisation and the edges are number of times an advice interaction takes place in a week and the node attributes are tenure you could take the product but this assumes this is meaningful in some way as these are two very different quantities.
In summary this is about research questions and data so yes it is possible but personally to enable easyy interpretation I would stick with the degree type example we have in the book.??

?

2. Whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

Yes this is done all the time

?

3. Whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality.

Yes this is possible but again it is about interpretation. See section 2.6. After multiplication the network you have is a count of composite relations. Suppose P is friend and N is enemy then the entries of PN(i,j) count the number of i's friends who have j as an enemy. If on PN we ran weighted degree centrality then the raw score would be a count of the total number of friends who have enemies relation. Note NP is different to PN.

?

4. Whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

No this is really covered in the PN answer above as this is the same thing in essence.

?

I look forward to any suggestions from all of you in the list. Thanks in advance!

?

Best,

Chuding

?

References

?

Borgatti, S., Everett, M., Johnson, J., & Agneessens, F. (2022). Analyzing social networks using R. SAGE Publications.

?

Everett, M. G., & Borgatti, S. P. (2014). Networks containing negative ties. Social Networks, 38, 111–120.

?

Singh, A., Singh, R. R., & Iyengar, S. R. S. (2020). Node-weighted centrality: A new way of centrality hybridization. Computational Social Networks, 7, 6.

?

Singh, R. R. (2022). Centrality measures: A tool to identify key actors in social networks. In A. Biswas, R. Patgiri, & B. Biswas (Eds.), Principles of Social Networking: The New Horizon and Emerging Challenges (pp. 1–27). Springer.

Re: Questions on node-weighted centrality or hybrid centrality

#7062

Hi
See my answers below.

Martin

On Saturday 10 August 2024 at 07:31:23 BST, CHUDING LING <lingchuding@...> wrote:

Dear Prof. Borgatti and all,

?

I am a reader of your book “Analyzing social networks using R” (Borgatti et al., 2022). Recently I notice that there is a stream of literature called node-weighted centrality or hybrid centrality (Singh et al., 2020; Singh, 2022). This literature points out that previous measures of centrality account for weight on edges (i.e., valued network) or length (e.g., eigenvector or beta centrality). However, these previous measures treat nodes as uniform. But it is very likely that different alters which an ego connects with are different in terms of a particular attribute (i.e., has different levels of importance). Let me use an example from Singh et al. (2020) to further illustrate:

?

For a friendship network, nodes represent persons and edges represent the friendship relationship between the considered set of persons. Here, weights on the nodes can be understood as a mapping of wealth, power, education level, or some other attribute of persons. It is notable that existence of two persons with identical attributes is highly unlikely and therefore all person’s attributes can be mapped to different real values based on the application specific mapping.

?

Therefore, my first question is whether we can account for such weights of nodes when calculating different types of centrality. Of note, I have quickly gone through the literature on node-weighted centrality or hybrid centrality, including 18 methodological or statistical research cited by Singh and colleagues (Singh et al., 2020; Singh, 2022), though I do not fully understand the mathematical parts. However, it seems that all of them use another type of centrality as the weight when calculate a particular type of centrality. For example, they use betweenness centrality as the weight when calculating degree centrality. From this perspective, your work on PN centrality might be also understood as a hybrid centrality (Everett & Borgatti, 2014). Particularly, in Everett and Borgatti (2014), you state:

?

Note that this equation weights everyone the same, but in many cases we might prefer a measure in which actors with high centrality affect the centrality score of the actors they are connected to more than those with a low score. In other words, we might want to capture the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.

?

Nevertheless, all the literature aforementioned do not fully capture the notion in the example from Singh et al. (2020) that I cited. That is, they use relational attributes to obtain the weight. Therefore, my second question is whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

?

Related to the PN centrality in Everett and Borgatti (2014), I have the third question. That is, whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality. According to your example in Chapter 2 of Borgatti et al. (2022), maybe we can indicate the first matrix as the negative network and the second matrix as the positive network, then the product of the two networks indicates the extent to which an ego is negatively connected to alters who are positively connected to others. Therefore, conceptually it may also capture “the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.” However, after experimenting with the example data provided in PN centrally section, I find the results are significantly different. Nevertheless, my fourth question is whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

?

To recap, I list the four questions as follows:

?

1. Whether we can account for weights of nodes when calculating different types of centrality, especially in population software like R.

First I refer to the R book so you can read what we say. I suggest you look at section 8.4.2. In the book we deal with degree and weighted attributes and these are in the chapter about node level measures. We do not (mainly through lack of space) discuss other centrality measures. There would be a vast array of possible measures. If the edges do not have weights the simplest way is to use the node attribute weight on the edge. If the edges do have weights then some method of combining node attribute weight and edge weight would need to be done. What you do needs to depend on exactly what is being measured. For example if you are in an organisation and the edges are number of times an advice interaction takes place in a week and the node attributes are tenure you could take the product but this assumes this is meaningful in some way as these are two very different quantities.

In summary this is about research questions and data so yes it is possible but personally to enable easyy interpretation I would stick with the degree type example we have in the book.??

?

2. Whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

Yes this is done all the time

?

3. Whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality.

Yes this is possible but again it is about interpretation. See section 2.6. After multiplication the network you have is a count of composite relations. Suppose P is friend and N is enemy then the entries of PN(i,j) count the number of i's friends who have j as an enemy. If on PN we ran weighted degree centrality then the raw score would be a count of the total number of friends who have enemies relation. Note NP is different to PN.

?

4. Whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

No this is really covered in the PN answer above as this is the same thing in essence.

?

I look forward to any suggestions from all of you in the list. Thanks in advance!

?

Best,

Chuding

?

References

?

Borgatti, S., Everett, M., Johnson, J., & Agneessens, F. (2022). Analyzing social networks using R. SAGE Publications.

?

Everett, M. G., & Borgatti, S. P. (2014). Networks containing negative ties. Social Networks, 38, 111–120.

?

Singh, A., Singh, R. R., & Iyengar, S. R. S. (2020). Node-weighted centrality: A new way of centrality hybridization. Computational Social Networks, 7, 6.

?

Singh, R. R. (2022). Centrality measures: A tool to identify key actors in social networks. In A. Biswas, R. Patgiri, & B. Biswas (Eds.), Principles of Social Networking: The New Horizon and Emerging Challenges (pp. 1–27). Springer.

Questions on node-weighted centrality or hybrid centrality

#7061

Dear Prof. Borgatti and all,

?

I am a reader of your book “Analyzing social networks using R” (Borgatti et al., 2022). Recently I notice that there is a stream of literature called node-weighted centrality or hybrid centrality (Singh et al., 2020; Singh, 2022). This literature points out that previous measures of centrality account for weight on edges (i.e., valued network) or length (e.g., eigenvector or beta centrality). However, these previous measures treat nodes as uniform. But it is very likely that different alters which an ego connects with are different in terms of a particular attribute (i.e., has different levels of importance). Let me use an example from Singh et al. (2020) to further illustrate:

?

For a friendship network, nodes represent persons and edges represent the friendship relationship between the considered set of persons. Here, weights on the nodes can be understood as a mapping of wealth, power, education level, or some other attribute of persons. It is notable that existence of two persons with identical attributes is highly unlikely and therefore all person’s attributes can be mapped to different real values based on the application specific mapping.

?

Therefore, my first question is whether we can account for such weights of nodes when calculating different types of centrality. Of note, I have quickly gone through the literature on node-weighted centrality or hybrid centrality, including 18 methodological or statistical research cited by Singh and colleagues (Singh et al., 2020; Singh, 2022), though I do not fully understand the mathematical parts. However, it seems that all of them use another type of centrality as the weight when calculate a particular type of centrality. For example, they use betweenness centrality as the weight when calculating degree centrality. From this perspective, your work on PN centrality might be also understood as a hybrid centrality (Everett & Borgatti, 2014). Particularly, in Everett and Borgatti (2014), you state:

?

Note that this equation weights everyone the same, but in many cases we might prefer a measure in which actors with high centrality affect the centrality score of the actors they are connected to more than those with a low score. In other words, we might want to capture the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.

?

Nevertheless, all the literature aforementioned do not fully capture the notion in the example from Singh et al. (2020) that I cited. That is, they use relational attributes to obtain the weight. Therefore, my second question is whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

?

Related to the PN centrality in Everett and Borgatti (2014), I have the third question. That is, whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality. According to your example in Chapter 2 of Borgatti et al. (2022), maybe we can indicate the first matrix as the negative network and the second matrix as the positive network, then the product of the two networks indicates the extent to which an ego is negatively connected to alters who are positively connected to others. Therefore, conceptually it may also capture “the notion that to have negative connections with actors who have a low centrality score is better than having negative connections to actors who have a high score.” However, after experimenting with the example data provided in PN centrally section, I find the results are significantly different. Nevertheless, my fourth question is whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

?

To recap, I list the four questions as follows:

?

1. Whether we can account for weights of nodes when calculating different types of centrality, especially in population software like R.

?

2. Whether we can use individual-level attribute like wealth, educational level or personality variables to indicate the weight.

?

3. Whether we can multiply the positive network and negative network using the matrix multiplication function in R first, and then calculate a particular type of centrality to obtain the weighted centrality.

?

4. Whether we can also understand the centrality based on the product of two matrix as one form of hybrid centrality.

?

I look forward to any suggestions from all of you in the list. Thanks in advance!

?

Best,

Chuding

?

References

?

Borgatti, S., Everett, M., Johnson, J., & Agneessens, F. (2022). Analyzing social networks using R. SAGE Publications.

?

Everett, M. G., & Borgatti, S. P. (2014). Networks containing negative ties. Social Networks, 38, 111–120.

?

Singh, A., Singh, R. R., & Iyengar, S. R. S. (2020). Node-weighted centrality: A new way of centrality hybridization. Computational Social Networks, 7, 6.

?

Singh, R. R. (2022). Centrality measures: A tool to identify key actors in social networks. In A. Biswas, R. Patgiri, & B. Biswas (Eds.), Principles of Social Networking: The New Horizon and Emerging Challenges (pp. 1–27). Springer.

Re: About xQAPLogisticRegression()

#7060

Tnx Tomas. (-:

I'll get in contact with Robert.

Call for contributions honoring Harrison White

#7059

开云体育

Hi all,

Below is a call from Jan Fuhse for contributions honoring the work of Harrison White. Papers will be published in the INSNA journal of CONNECTIONS.

The primary focus is on short, personal pieces and reflections rather than full-blown research papers (even though these might also be of interest). Ideally, we would have a wide range of contributions reflecting the different sides of Harrison's work and his far-reaching impact on a number of fields and trends.

If in doubt or if you're undecided what to write, please contact Jan Fuhse <jan.fuhse@...>, who is serving as guest editor for this series of contributions.

steve

---------------------------------------------------------------

?Call for short pieces in honor of Harrison White for Connections (official journal of the International Network of Social Network Analysis; )

Eminent networks scholar and 1984 Simmel awardee Harrison Colyar White passed away on May 18, 2024, in his Tucson home aged 94. As an early convert from physics to sociology, White has played an outsized role in the development of social network analysis between 1960 and 2010. His early contributions from the 1960s and 1970s have been hailed as the “Harvard thrust” (Freeman) or the “Harvard breakthrough” (Scott). They center around the notions of vacancy chains and structural equivalence, and the method of blockmodeling analysis. While at Harvard, he shaped two extraordinary generations of graduate students, including Simmel awardees Phillip Bonacich, Ronald Breiger, Kathleen Carley, Bonnie Erickson, Mark Granovetter, Philippa Pattison, and Barry Wellman. His ideas shaped the new economic sociology and the sociology of art. Since 1990, White was the driving force behind “relational sociology” (Emirbayer), offering a theoretical account of social networks and their role in the social world. This approach combines network structuralism with a focus on meaning and culture in networks, as well as on the dynamics of social processes in them.

To honor these far-reaching contributions to social network research, Connections invites short pieces of two to five pages. These may center on personal memories, as well as on reflections of White’s impact on personal intellectual trajectories and on developments in the field. The contributions will be selected on the basis of their relevance. Proper articles of up 40 pages will also be considered.

Please submit your contributions until October 31, 2024, via the Connections and make sure to specify “Special Series for Harrison White” in the drop-down menu “Select Article Type”. Connections publishes accepted articles online, open access under . For any questions, please contact Jan Fuhse (jan.fuhse@...).

?

Re: About xQAPLogisticRegression()

#7058

Hi Paulo,

The newer version of xUCINET can be downloaded here: As for your question, the number of permutations depends on how much time you have - the more permutations you run, the more precise your p-values will be, but at the expense of computational time (as you observed yourself).

Robert Krause (University of Kentucky) is currently working on extending QAP estimation procedure to all Generalised Linear Models (Poisson regression etc.) which also includes the possibility to parallelise the procedure on multiple processor cores which can speed things up considerably. If you are interested in this, let Robert know and I am quite sure he will share his script and thoughts with you.

Best,

Tomas Diviak

Re: About xQAPLogisticRegression()

#7057

Computational cost
NPerm >1000 10-minute calculation
NPerm >1500 16-minute calculation

Results on betas are quite the same, on p-value some refinement - see the attached files.

Looks we don't have to consider any high permutations numbers.

NPERM-1000.txt

NPERM-1500.txt

Comparing two different groups #help

#7056

开云体育

Hey, grad student here. I was wondering if there’s a way in UCINET 6 to do a MR-QAP test using two different matrices and sets of nodes. Specifically, I have two different classes with two different groups of students. One class has 19 nodes and the other 22. I have the matrices made for each class individually for the question “Who do you turn to for advice?” What I’m hoping to answer is whether it’s possible to do these tests. I’m adept at doing the tests with two matrices from the same group of nodes, but I’m wondering if/how to get the r^2 and p value with two different groups.

I suppose I could combine *all* nodes to one advice network and use an attribute for which class they are in, but I’m guessing there’s a better way.

Thanks!

Re: About xQAPLogisticRegression()

#7055

Tnx Professor (-:

In the meantime, I have another question about LR QAP. From the HELP of xQAPCorrelation(). I guess NPERM is also present at LR QAP () - taking UCINET as an example, and the description says "If NPERM (number of permutations) is set to a value higher than 0 (default is 1000), a quadratic assignment permutations-based (QAP) significance is also produced, which will provide information about the regression coefficients for the permuted variables. A higher value than 1000 is generally advisable." Especially at the last period, it says higher than 1000. What criteria should I use to define how high NPERM should be? 10000, 5000, 30000...

p

Re: Comparing two different groups #help

#7054

开云体育

Having written the below (because it was fun to do), I am guessing that you have a different research question in mind. You write

If my hypothesis is that the two sections (and therefore two different student groups) had significantly different networks (with alpha set at 0.05), could I do this in UCINet? With 20,000 permutations? If I try to use the QAP tool, UCINet won't let me compare since both matricies are different with a different set of nodes.

This suggests you want to know whether the advice networks in the two sections are similar in structure. For example, are both clumpy? Are both centralized? etc. To me this is best answered qualitatively, but you could also consider ERGM to determine whether you can fit the same model to the two datasets.?

steve

Stephen P. Borgatti

Carol Martin Gatton Chair of Management

Gatton College of Business and Economics

University of Kentucky

?

Show quoted text

From: [email protected] <[email protected]> on behalf of Steve Borgatti via groups.io <steve.borgatti@...>
Sent: Thursday, July 11, 2024 3:27 PM
To: [email protected] <[email protected]>
Subject: [ucinet] Comparing two different groups #help

?

Hi Clark. Suppose you have two datasets, advice1 and advice2, as shown here:

?

->dsp advice1

?

? ? ? ? ? ?1 ?2 ?3 ?4 ?5 ?6 ?7 ?8 ?9 10

? ? ? ? ? -- -- -- -- -- -- -- -- -- --

? ? ? ? 1 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0

? ? ? ? 2 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0

? ? ? ? 3 ?0 ?0 ?0 ?1 ?0 ?0 ?1 ?0 ?0 ?0

? ? ? ? 4 ?0 ?0 ?1 ?0 ?1 ?0 ?0 ?0 ?1 ?0

? ? ? ? 5 ?0 ?0 ?0 ?1 ?0 ?0 ?1 ?0 ?0 ?0

? ? ? ? 6 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?0 ?0

? ? ? ? 7 ?0 ?0 ?1 ?0 ?1 ?1 ?0 ?0 ?0 ?1

? ? ? ? 8 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0

? ? ? ? 9 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?1 ?0 ?0

? ? ? ?10 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?0 ?0

?

10 rows, 10 columns, 1 levels.

?

->dsp advice2

?

? ? ? ? ? ?1 ?2 ?3 ?4 ?5 ?6 ?7 ?8 ?9 10 11 12

? ? ? ? ? -- -- -- -- -- -- -- -- -- -- -- --

? ? ? ? 1 ?0 ?1 ?0 ?1 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?0

? ? ? ? 2 ?1 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?1

? ? ? ? 3 ?0 ?1 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?1 ?0

? ? ? ? 4 ?1 ?0 ?0 ?0 ?1 ?0 ?0 ?1 ?0 ?1 ?0 ?0

? ? ? ? 5 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0

? ? ? ? 6 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?1 ?1 ?0 ?0

? ? ? ? 7 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?1 ?1

? ? ? ? 8 ?1 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0

? ? ? ? 9 ?0 ?0 ?0 ?0 ?0 ?1 ?1 ?0 ?0 ?1 ?0 ?1

? ? ? ?10 ?0 ?1 ?0 ?1 ?0 ?1 ?0 ?0 ?1 ?0 ?0 ?0

? ? ? ?11 ?0 ?0 ?1 ?0 ?1 ?0 ?1 ?0 ?0 ?0 ?0 ?1

? ? ? ?12 ?0 ?1 ?0 ?0 ?0 ?0 ?1 ?0 ?1 ?0 ?1 ?0

?

12 rows, 12 columns, 1 levels.

?

Now you want to mush them together into a single network that has 22 nodes. In the CLI you would type:

?

->advice = union(advice1 advice2)

->dsp advice

?

? ? ? ? ? ? ?1 ?2 ?3 ?4 ?5 ?6 ?7 ?8 ?9 10 11 12 13 14 15 16 17 18 19 20 21 22

? ? ? ? ? ? 1- 1- 1- 1- 1- 1- 1- 1- 1- 1- 2- 2- 2- 2- 2- 2- 2- 2- 2- 2- 2- 2-

? ? ? ? ? ? ?1 ?2 ?3 ?4 ?5 ?6 ?7 ?8 ?9 10 ?1 ?2 ?3 ?4 ?5 ?6 ?7 ?8 ?9 10 11 12

? ? ? ? ? ? -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

? ? ?1 ?1-1 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?2 ?1-2 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?3 ?1-3 ?0 ?0 ?0 ?1 ?0 ?0 ?1 ?0 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?4 ?1-4 ?0 ?0 ?1 ?0 ?1 ?0 ?0 ?0 ?1 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?5 ?1-5 ?0 ?0 ?0 ?1 ?0 ?0 ?1 ?0 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?6 ?1-6 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?7 ?1-7 ?0 ?0 ?1 ?0 ?1 ?1 ?0 ?0 ?0 ?1 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?8 ?1-8 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?9 ?1-9 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?1 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? 10 1-10 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? 11 ?2-1 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?1 ?0 ?1 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?0

? ? 12 ?2-2 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?1 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?1

? ? 13 ?2-3 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?1 ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?1 ?0

? ? 14 ?2-4 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?1 ?0 ?0 ?0 ?1 ?0 ?0 ?1 ?0 ?1 ?0 ?0

? ? 15 ?2-5 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0

? ? 16 ?2-6 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?1 ?1 ?0 ?0

? ? 17 ?2-7 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?1 ?0 ?1 ?1

? ? 18 ?2-8 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?1 ?0 ?0 ?1 ?0 ?0 ?0 ?0 ?0 ?0 ?0 ?0

? ? 19 ?2-9 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?0 ?0 ?0 ?0 ?1 ?1 ?0 ?0 ?1 ?0 ?1

? ? 20 2-10 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?1 ?0 ?1 ?0 ?1 ?0 ?0 ?1 ?0 ?0 ?0

? ? 21 2-11 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?0 ?1 ?0 ?1 ?0 ?1 ?0 ?0 ?0 ?0 ?1

? ? 22 2-12 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?0 ?1 ?0 ?0 ?0 ?0 ?1 ?0 ?1 ?0 ?1 ?0

?

22 rows, 22 columns, 1 levels.

?

When you run the Union function, it will automatically create an attribute dataset called "unionid":?

?

->dsp unionid

?

? ? ? ? ? ? 1

? ? ? ? ? ? F

? ? ? ? ? ? i

? ? ? ? ? ? l

? ? ? ? ? ? e

? ? ? ? ? ? -

? ? ?1 ?1-1 1

? ? ?2 ?1-2 1

? ? ?3 ?1-3 1

? ? ?4 ?1-4 1

? ? ?5 ?1-5 1

? ? ?6 ?1-6 1

? ? ?7 ?1-7 1

? ? ?8 ?1-8 1

? ? ?9 ?1-9 1

? ? 10 1-10 1

? ? 11 ?2-1 2

? ? 12 ?2-2 2

? ? 13 ?2-3 2

? ? 14 ?2-4 2

? ? 15 ?2-5 2

? ? 16 ?2-6 2

? ? 17 ?2-7 2

? ? 18 ?2-8 2

? ? 19 ?2-9 2

? ? 20 2-10 2

? ? 21 2-11 2

? ? 22 2-12 2

?

This identifies which original network (in your terms, section) a node belongs to.?(It will come in handy later).

?

Let's suppose you also have an independent variable called "simil" that you will use the predict advice:

?

->dsp simil

?

? ? ? ? ? ? ? ? 1 ? ? 2 ? ? 3 ? ? 4 ? ? 5 ? ? 6 ? ? 7 ? ? 8 ? ? 9 ? ?10 ? ?11 ? ?12 ? ?13 ? ?14 ? ?15 ? ?16 ? ?17 ? ?18 ? ?19 ? ?20 ? ?21 ? ?22

? ? ? ? ? ? ? 1-1 ? 1-2 ? 1-3 ? 1-4 ? 1-5 ? 1-6 ? 1-7 ? 1-8 ? 1-9 ?1-10 ? 2-1 ? 2-2 ? 2-3 ? 2-4 ? 2-5 ? 2-6 ? 2-7 ? 2-8 ? 2-9 ?2-10 ?2-11 ?2-12

? ? ? ? ? ? ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -----

? ? ?1 ?1-1 ? ? 0 0.521 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?2 ?1-2 0.521 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?3 ?1-3 ? ? 0 ? ? 0 ? ? 0 0.670 ? ? 0 ? ? 0 0.381 ? ? 0 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?4 ?1-4 ? ? 0 ? ? 0 0.670 ? ? 0 0.183 ? ? 0 ? ? 0 ? ? 0 0.500 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?5 ?1-5 ? ? 0 ? ? 0 ? ? 0 0.183 ? ? 0 ? ? 0 0.932 ? ? 0 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?6 ?1-6 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 0.326 ? ? 0 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?7 ?1-7 ? ? 0 ? ? 0 0.381 ? ? 0 0.932 0.326 ? ? 0 ? ? 0 ? ? 0 0.876 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?8 ?1-8 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 0.885 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? ?9 ?1-9 ? ? 0 ? ? 0 ? ? 0 0.500 ? ? 0 ? ? 0 ? ? 0 0.885 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? 10 1-10 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 ? ? 0 0.876 ? ? 0 ? ? 0 ? ? 0 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?

? ? 11 ?2-1 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 1 0.829 0.380 0.608 0.473 0.814 0.948 0.611 0.868 0.679 0.950 0.887

? ? 12 ?2-2 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.829 ? ? 1 0.967 0.533 0.799 0.739 0.788 0.864 0.877 0.673 0.928 0.888

? ? 13 ?2-3 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.380 0.967 ? ? 1 0.952 0.975 0.659 0.898 0.837 0.319 0.460 0.407 0.865

? ? 14 ?2-4 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.608 0.533 0.952 ? ? 1 0.351 0.757 0.881 0.991 0.418 0.898 0.802 0.842

? ? 15 ?2-5 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.473 0.799 0.975 0.351 ? ? 1 0.942 0.834 0.983 0.702 0.611 0.863 0.603

? ? 16 ?2-6 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.814 0.739 0.659 0.757 0.942 ? ? 1 0.970 0.445 0.734 0.404 0.349 0.541

? ? 17 ?2-7 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.948 0.788 0.898 0.881 0.834 0.970 ? ? 1 0.817 0.569 0.295 0.909 0.915

? ? 18 ?2-8 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.611 0.864 0.837 0.991 0.983 0.445 0.817 ? ? 1 0.521 0.886 0.564 0.144

? ? 19 ?2-9 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.868 0.877 0.319 0.418 0.702 0.734 0.569 0.521 ? ? 1 0.659 0.704 0.999

? ? 20 2-10 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.679 0.673 0.460 0.898 0.611 0.404 0.295 0.886 0.659 ? ? 1 0.419 0.425

? ? 21 2-11 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.950 0.928 0.407 0.802 0.863 0.349 0.909 0.564 0.704 0.419 ? ? 1 0.544

? ? 22 2-12 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 0.887 0.888 0.865 0.842 0.603 0.541 0.915 0.144 0.999 0.425 0.544 ? ? 1

?

Now you are ready to run QAP Regression (Ctrl-R in the UCINET main menu). Fill out the dialog box as follows:

?

Note the line under "Optional partition" says "unionid col 1". This tells it that the permutations should be conducted only within section.?

?

steve.

?

Stephen P. Borgatti

Carol Martin Gatton Chair of Management

Gatton College of Business and Economics

University of Kentucky

?

From: [email protected] <[email protected]> on behalf of Andersen, Clark Isaac via groups.io <andersenc21@...>
Sent: Tuesday, July 9, 2024 3:57 PM
To: [email protected] <[email protected]>
Subject: Re: [ucinet] Comparing two different groups #help

?

Thanks Joe!?

1. So I did run the same survey in both classes. We'll call them "Section 1" and "Section 2". Section 1 was a different learning style than section 2 (i.e. group work vs. lecture) so the students weren't the same. I have two separate matrices with different nodes but having asked the same question. Just as an example, let's say that section 1 had only 5 students indicate that they seek out others for advice, whereas section 2 had 25. This obviously leads to much different densities and avg. in-and-out degrees. If my hypothesis is that the two sections (and therefore two different student groups) had significantly different networks (with alpha set at 0.05), could I do this in UCINet? With 20,000 permutations? If I try to use the QAP tool, UCINet won't let me compare since both matricies are different with a different set of nodes.

Does that make sense?

2. I'll look into this as well. If you know of an online resource that uses this strategy, could you send it my way?
Cheers!
Clark

Re: Comparing two different groups #help

#7053

Thanks Dr. Borgatti! This is perfect.

Re: About xQAPLogisticRegression()

#7052

开云体育

Hi Paulo, I’ve forwarded your message to Filip Agneessens.

?

steve

?

Show quoted text

From: [email protected] <[email protected]> On Behalf Of Paulo Matui via groups.io
Sent: Tuesday, July 9, 2024 09:33
To: [email protected]
Subject: [ucinet] About xQAPLogisticRegression()

?

Hello,

About xQAPLogisticRegression(), I can't find it in xUCINET version 0.0.2.0020. However, in the 2022 edition of the book 'Analyzing Social Networks Using R', there are even exercises.?
Are there any patches or scripts I should consider?

matui

Comparing two different groups #help

#7051