Skip to main content


Welcome, the Hub connects all projects

MSPnet Academy Discussions


All Topics for this Forum

Topic: "MSPnet Academy:Using Social Network Analysis"

Topic Posts

MSPnet Academy Discussion
June 6 - June 20, 2013

Bill Zoellick, Education Research Director, SERC Institute

MSPnet Academy: Using Social Network Analysis to Study How Teacher Communities of Practice Affect Science Curricula Enactment

A primary mission of the MainePSP is improving science teaching practice in rural settings by building and training a community of educators. The core mechanism for improving teaching practice has been formation of three teacher communities that enact common curricula materials across diverse, rural settings in grades 6-9. Consequently, study of the relationship between community participation and changes in teaching practice is a primary research focus. Social network analysis (SNA) is central to the research design. In this session we introduce key issues and questions that have emerged in application of SNA to the study of teacher communities, and we initiate an ongoing online conversation with participants to improve sharing of SNA know-how and research findings across MSP projects.

Archives for this academy

This archived topic is open to the public.

restart conversation
This topic has 22 posts, showing: 1-20
1   2   Next

Welcome! ... And things to talk about ...

posted by: Bill Zoellick on 6/6/2013 8:57 pm

Thanks for your visiting this discussion ...

My presentation earlier today touched on a number of topics ...

* how the MainePSP has made developing and supporting a teacher community one of the central elements in its program to improve rural science education.

* use of "sociometric," whole-network analysis to test assertions about overall community structure and growth

* use of "egocentric," analysis of local network structure to support investigations about teachers' learning from each other

* use of Gephi as a tool to interactively explore network structure

* methodological concerns regarding network extent, data collection, and more ...

Each of these topics can be a jumping-off place for further discussion. I am also very interested in hearing more about how others are using social networks in their research. What are you using that is working? What kinds of difficulties are you encountering?

My hope is that we can use this two week discussion as a kind of "test bed" for what could be a more permanent online "birds of a feather" group. My sense is that social network analysis has a lot of untapped potential as an important tool in mixed-methods investigations seeking to gain insight into how teacher communities grow and change, Sharing know-how and experience can improve our collective ability to access that potential.

-- Bill

Social Network Analysis

posted by: Frederica Frost on 6/9/2013 8:23 am

Thank you for the post. I wanted to attend your presentation but was unable. Do you have a transcript, PPT, something like that to share?
Frederica

Copies of the Presentation

posted by: Bill Zoellick on 6/9/2013 4:40 pm

Frederica --

A recording of the presentation and a copy of my slides are available on the MSPnet site at:

http://hub.mspnet.org/index.cfm/MSPnet_Academy_Using_Social_Network_An alysis

At the moment, the links are also on the MSPnet home page at http://hub.mspnet.org

-- Bill

Using Social Network Analysis

posted by: Valida Walker on 6/10/2013 10:55 am

Hi Bill, I was especially interested in the use of GEPHI, but have not found a tutorial that made up for a deficit in understanding how to start.

Using Gephi

posted by: Bill Zoellick on 6/10/2013 2:05 pm

Valida --

It would be great if someone would write a book on using Gephi. It IS pretty new.

Have you found the various tutorials at http://gephi.org/users/ ? If not, that would be the place to start. The first tutorial uses a dataset about Les Miserables that comes packaged with the Gephi installation. Other tutorials include links to different datasets that you can experiment with.

I found that learning to use Gephi requires some time to just play with it a bit. But ... I found the same to be true of UCINET and NetDraw.

There is also a wiki for Gephi (http://wiki.gephi.org/) and a users forum where I have found answers to questions about file importing and so on (http://forum.gephi.org/ ).

I found the documentation on importing datasets that include attributes for the nodes and for the edges to be kind of weak. It turns out that Gephi DOES do a good job of this. It is something I could provide more detail on once you get past the basics.

-- Bill

"member checking"?

posted by: Brian Drayton on 6/13/2013 2:27 pm

Hi, Bill
I enjoyed the presentation! I don't remember you discussing this: As you have done these analyses of the teacher community, have you done any ground-truthing with the people whose relationships/interaction patterns are represented in the networks? I don't know, but it might be an interesting way to help the participants think critically and concretely about their collegial habits.
-- brian

Member Checking, Ground Truthing, and Feedback

posted by: Bill Zoellick on 6/13/2013 3:43 pm

Brian --

Interesting question -- Thanks for asking.

Actually, I think there are a couple of questions here. One is with regard to "ground truthing." After the first year's survey we interviewed a sample of teachers who responded to the SNA questionnaire and asked them, among other things, who they turned to for advice and conversation. Our goal was to check the validity of the survey instrument. Pitts and Spillane (2009) did this more thoroughly and completely, using a more rigorous, exhaustive interview for a similar survey instrument -- we wanted to do a quick check to make sure that our version of the survey was "on track" and picking up the important contacts for teachers.

What we found was that the teachers named more colleagues in their online responses than they identified in their telephone responses. Our takeaway was that in constructing their online responses the teachers thought more carefully about ALL the people they talked to ... which is a good thing.

A SECOND question in your posting is with regard to sharing findings with teachers. I do provide an annual analysis of the shape and changes in the teacher community to teachers at our project "summit" meeting and retreat. I talk about things such as the degree to which the community reflects connections within districts, the degree to which connections are growing among teachers working together on implementing a particular part of the program, and so on.

My perception is that the teachers find this mildly interesting and understand that it is important to the project as a whole (which they tell me is why they bother to fill out the survey each year) -- but my perception is also that they don't find it to be directly relevant to their own work. They are, not surprisingly, more focused on how to best use the new materials and methods with which the project has presented them. There is certainly understanding that the opportunity to work with other teachers is a valuable resource the assists in effective use of materials and improvement of methods -- they TELL us they value that. But it is reasonable to assume that the view of the community that teachers find to be most important and relevant is LOCAL (their own personal network) -- rather than systemic.

Could we present teachers with pictures of their own personal networks, or of the second order connections they have? My concern there would be that we would be violating our commitment to them to keep their responses private and to only present data in ways that are de-identified.

Again ... thanks ... this was a useful question for opening up these issues.

-- Bill

Pitts, V. M., & Spillane, J. P. (2009). Using social network methods to study school leadership. International Journal of Research & Method in Education, 32(2), 185-207.

Leveraging the data beyond your study?

posted by: Heidi Larson on 6/14/2013 10:19 am

Who will own this data at the end? I'm wondering, if the participants, principals, superintendents will not see results except in summary fashion from your reports, if there is a way that the data can be used in ways other than to benefit your study. For instance, it seemed that some of the teachers were more central than others -- teachers tended to turn to them more often than to others. Is there a way to make use of this information without sacrificing the confidentiality of the study? They could become teacher leaders, or even if they are the only science teacher in their school, they could, for example, consult with the Maine DOE, or be invited to give webinar presentations on whatever it is that makes them such magnets, or asked to contribute to a journal that can be kept for future use...

Basically, just looking for your thoughts on other ways to use your data to benefit education for Maine kids.

Thanks.

--Heidi

Using SNA for evaluation as well as research

posted by: Bill Zoellick on 6/15/2013 5:48 pm

Heidi --

As I understand your question, you are asking about use of SNA in formative program evaluation -- in addition to its use as research tool in the service of construction or testing of theory.

One of the powerful things about SNA is that it results in pictures -- graphical representations of structures and relationships that would be difficult to describe in words. It follows that SNA could be really useful for helping a client see important things about a system or community. Maryann Durland has a chapter titled "A Formative Evaluation of the Integration of Two Departments" that provides a good example of this kind of use ... it is in Durland, M. M., & Fredericks, K. A. (Eds.). (2006). Social Network Analysis in Program Evaluation. San Francisco, CA: Jossey-Bass.

However ... the issue of confidentiality is REALLY important. Respondents are telling you who they work with ... and, by extension, who they don't work with. It is often useful to ask about the importance of the relationships they identify ... which caries the risk that someone else might be hurt -- or angry -- to find that their relationship with a respondent is less important than others. An administrator's learning that he or she is not an important resource for a teacher might change the administrator's relationship with the teacher in unproductive ways.

In short, since SNA data is about relationships between people, making those data available to the people who are involved in the relationships carries risks of damage to the relationships unless anonymity can be preserved. And, providing information about a person's responses is unethical unless everyone knows in advance that you intend to make their responses public. (And, if they know that, you are unlikely to get candid responses.)

Further, even if you don't "name names" -- the chances are good the anonymity will be comprised unless the diagram you present includes responses from a large number people.

So ... SNA seems like a useful formative evaluation tool. But protecting anonymity is essential.

-- Bill

An example of a simpler application of SNA

posted by: David May on 6/14/2013 9:21 am

Hello all,

Our first MSP, called VIP K-16, included several different kinds of programs for teachers and faculty to get involved with each other. We conducted surveys that simply asked them to name each person with whom they had meaningful collaborations, on such things as redesigning curriculum or their teaching methods, both before our MSP's programs commenced and after 3-4 years. The resulting networking diagrams were quite striking. You can see them, along with a quick summary, here:
http://hub.mspnet.org/index.cfm/lnc08_may/page/index

More detail about the study can be found here:
http://hub.mspnet.org/index.cfm/13774

And although the data are not online anywhere, I can tell you that a year after our MSP ended, we surveyed them again to find that a large majority of the professional collaborations they had developed were still ongoing.

(The VIP K-16 partnership included the University System of Maryland, Montgomery County Public Schools, and Montgomery College. Our networking study was conducted by USM and our external evaluators at Westat.)

I hope this information is useful!
-David May, USM

SNAs

posted by: Kavita Mittapalli on 6/15/2013 7:41 am

Hi, David
Thanks for posting the paper link and study description. I checked it out and have sent a note to Dr Parsad at Westat asking him the software he used to make the SNs. Hope to hear from him soon. I have used Gephi is the recent past but found it a bit cumbersome -maybe I need some additional practice with some play data. Thanks again.

Kavita Mittapalli, Ph.D.
MN Associates, Inc.
www.mnassociatesinc.com

SNA software

posted by: David May on 6/18/2013 2:14 pm

Hi Kavita,

Hopefully Basmat will get back to you, but if not, I can tell you that she used UCInet to create the diagrams. I've used the trial version a couple of times, and it's finicky but gets the job done!

Good luck,
David

UCINET

posted by: Bill Zoellick on 6/19/2013 9:05 am

Kavita --

Just a quick addition to David's useful reply about UCINET ... adding a couple of points and a link to a resource ...

UCINET is a very powerful analysis package that includes a number of kinds of analysis that are not available in Gephi. So ... this is not an either/or situation ... I find myself using both tools. The big difference is that UCINET is not interactive -- you run an analysis and you get a result. The pictures are made with a companion package called Netdraw.

None of these tools is particularly "user-friendly" -- getting familiar with them and getting productive results takes some patience and time.

The is a really excellent online textbook by Robert A. Hanneman and Mark Riddle that both provides a very readable introduction to the basic concepts of social network analysis and ... along the way ... a nice introduction to UCINET and Netdraw. You can access it at: http://faculty.ucr.edu/~hanneman/nettext/

-- Bill

Thanks ... And a question

posted by: Bill Zoellick on 6/15/2013 5:53 pm

David --

Thanks for sharing this. It actually looks like a pretty ambitious study, rather than a "simple" one.

Slides 31 and 32 in your presentation look pretty interesting ... what is going on here -- with and without the central figure?

-- Bill

Your question about our central figure

posted by: David May on 6/18/2013 2:20 pm

Bill,

About our slides 31-32, with and without the "central figure:" We identified the one person with the largest number of connections - a project manager in the school system - and removed his connections from the data. The result is that there was still a dense network without him. Our interpretation is that through his work, or at least through the design of the program, many people were brought together into collaborative relationships that didn't exist before, and that they didn't depend on the presence of this central figure. A positive result, we think!

--David

Thanks ...

posted by: Bill Zoellick on 6/19/2013 8:48 am

Interesting ... thanks ...

Connection between the two cohorts?

posted by: Heidi Larson on 6/14/2013 10:21 am

I could never quite grasp the connection between the two cohorts. Did cohort two join in workshops with cohort one? Or were they kept separate?

Connecting Cohorts

posted by: Bill Zoellick on 6/15/2013 6:06 pm

Heidi --

During our "summer academy" for teachers in the program last summer we did provide separate tracks for the first and second cohorts. We did this because the new, second cohort needed an introduction to the curriculum materials that they would be using for the first time, whereas the first cohort, who had been using materials for a year, needed time together to talk about what worked and what didn't.

However, once the school year started, teachers from both cohorts attended Saturday and/or evening sessions together to talk more about implementation issues and to explore the intersection of science content and science practice. So, then they were mixed.

Now that we have done that for a year, we are reflecting on it. Feedback from teachers indicates that there might be important differences in what teachers from the two cohorts want to focus on.

SNA data that we are collecting now -- about relationships over the last year -- might also tell us something about the kinds of connections that the second cohort formed. (We have evidence that teachers from the first cohort continues to connect with each other.)

-- Bill

How did teachers communicate?

posted by: Heidi Larson on 6/14/2013 10:25 am

I can't remember if you talked about this in your presentation. Your surveys asked who teachers connected with and how often, but did they also ask how teachers communicated? I ask because part of my work is aimed at helping school administrators learn the benefits of social networking via mediums such as Twitter, both for their own personal development and for communicating with their school community. So, I was wondering how these teachers, especially in cohort one, were communicating and building their networks.

Thanks.

Teacher Communication

posted by: Bill Zoellick on 6/15/2013 6:22 pm

Heidi --

Yes, we did collect data about that. We provided them with a list of communication modes and, for each relationship, to check ALL of the modes that applied.

For teachers in our initial cohort, not surprisingly, 69% of the relationships included meeting with each other at the various meetings that are part of the MainePSP project. After that, the next most frequent kind of contact (38% of relationship) is among people who work together in the same school. THEN (and we find this to be exciting) -- the next most important means of connection (32% of relationships) is on our private website set up to support teacher collaboration. School district meetings and email connectons are next, at 28% and 26%, respectively.

For teachers who were not implementing the program, the primary modes of connection are email (42%), within building contacts (40%), school district meetings (36%), MainePSP sponsored meetings (30%), and outside school get togethers (24%). They don't have access to the private website.

-- Bill