Juggling Multiple Networks in Multiple Streams
should be present on the site . End users would self-identify as actor-stakeholders in a specific policy domain , either by directly signing up to a specific ( selfdefined ) URL within IMPolS and then nominating network colleagues ( the tool will then send e-invites ), or by initiating an IMPolS instance during a network meeting ( for instance , an Annual General Meeting ) at which a first round of network data is entered .
At this stage , actor / stakeholders also choose a representative icon ( categories in Figure 2 shown in the red box at the bottom of the screen , but fully adaptable to other specific policy domains ), and may define and select categories of participants . Actors may continue to be added ; the expectation is that from the initiation stage onward actors will regularly access their domain and answer about a dozen questions relating to their position and connection in the network . These data will then be added to the database and first , the network visualization algorithm will recalculate the three network configurations , and , second , notify other members of the network that an actor has updated their position and connection ( thus prompting others to do the same ). Over time , with more data added , the network mapping visualizations ( and possibly actor behavior , about which more below ) will gain intricacy , and will allow for a dynamically animated , pulsing set of network configurations . Further sophistication could be added , either by the self-selected network members or by a network manager ( again , see below ), through the refinement of the timeline with critical events , such as described by Kingdon above ( e . g ., elections , climactic events , policy change or press release , etc .).
In our alpha testing of the IMPolS tool we have found a number of things : first , a visually attractive and transparent architecture of both the input screens ( user identification and network variable entry ) and the network screens would increase the likelihood of actors engaging with the tool . This is precisely what the developers of the Gephi software platform ( Gephi is an interactive visualization and exploration platform for all kinds of networks and complex systems , dynamic and hierarchical graphs — see gephi . org ) found : applying visualization principles from the gaming sector enhances the attractiveness of the application ( Bastian , Heymann , and Jacomy 2009 ). Second , and in full concordance with both the propositions by Kingdon and our initial research , virtually all actors in their “ facevalue analysis ” of the network outputs focused more on the problem stream than on either of the other streams . They found that problem stream graphic network visualizations provided them with arguments and impulses to ( re ) consider
a ) the nature of the problem they are engaging with ; b ) their framing of the problem , and how it might link with other actors if reframed ( Kingdon ’ s “ alternative specification ”); and c ) how to seek alliances with actors found to be similar ( sometimes called “ homophilic network relations ,” for example , Monge and Contractor 2003 , and Provan and Kenis 2008 ), either in their position in the problem network ( in terms of connectedness and centrality ), or in perceived similarities in mission or vision of the institutional characteristics of the actor .
206