Identifying coalitions / One-mode congruence networks

[Hendrik]

Hello everyone,

I am currently working with DNA + Visone and would like to find out how coalitions develop over time.

To visualise these coalitions, I am creating congruence networks with the variable Organisations.

I can now see the coalitions, but I lack the information on which statements the coalitions are based. A ‘manual’ link by looking up all statements in a two-mode network or by scrolling through the event Excel is not possible with the amount of data I am working with.

Am I making a mistake?

Thank you very much in advance!

Best
Hendrik

[Kimmo Elo]

Hi Hendrik,

being quite familiar with 'dna' (you refer to Discourse Network
Analyser, right) I try to answer your question. I think you have a bit
misunderstood how the export of a congruence network works. It creates a
an adjacency matrix (org x org), in which "the cells represent on how
many concepts any two actors (i e., the row organization and the column
organization) had the same issue stance" (see
https://usermanual.wiki/Pdf/dnamanual.2049511603.pdf, p. 56 for
details). The point is that such a matrix simply counts agreements. You
can easily test it by exporting the congruence network into a csv file.

What you could do is to export a TWO-mode network into a csv file. This
network should have CONCEPT (=statement) as "Varible 1" and ORGANISATION
as "Variable 2". To simplify things you could set "Qualifier
aggregation" to "ignore".

If you now open the exported csv file, you will see a table, where
statements are rows and organisations are columns. Already here you
could rather easily find shared statements by filtering appropriate
columns (e.g. organisations) for "1".

HTH,

Kimmo

> --
> You received this message because you are subscribed to the Google
> Groups "visone-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to visone-users...@googlegroups.com
> <mailto:visone-users...@googlegroups.com>.
> To view this discussion visit
> https://groups.google.com/d/msgid/visone-users/d1998c42-0cd3-46e5-9420-51375df19195n%40googlegroups.com <https://groups.google.com/d/msgid/visone-users/d1998c42-0cd3-46e5-9420-51375df19195n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Kimmo Elo
Bergeninkatu 2 F 59
20320 Turku
GSM: +358 44 342 2122
Sposti: kimmo...@gmail.com
===============================

[Hendrik]

Hi Kimmo,

thank you for your response and help. 

I would like to refer to those networks I found in Ohlendorf et al. (2023) (https://doi.org/10.1016/j.eist.2023.100692)

Here the authors somehow bring together the congruence network of the actors with storylines. 

Am I missing a step?

Best
Hendrik

[Kimmo Elo]

Hi Hendrik,

I took a look on the article and now understand better what you are
looking for. Actually, the paper is rather week in describing how the
network data was created. On p. 7 they write:

"We create an actor congruence network to visualize individual actors
using similar storylines. Links between actors are normalized to account
for unequal numbers of coded storylines between actors, by dividing the
edge weight with the average number of storylines (Leifeld, 2017). Fig.
3 shows the resulting discourse network in detail. Fig. 4 shows the same
discourse network, but with all edges highlighted that include at least
one storyline in support of the respective conflict position. All
figures are created with Stata and edited with Inkscape."

They do NOT say having used DNA to create their congruence networks.
Actually, they mention "Stata" so I suspect that they have exported a
two-mode network (actor to concept) as CSV and then created the
congruence network in Stata by linking actors by the storyline variable
and then summarising the co-occurrence. I could do the same with R, so
this is not a big deal :-) But I should have been described better in
the article.

What also strikes me is the reference to Inkscape. So it seems they have
produced something with Stata and the edited this graph (?) in Inkscape.
To be honest, I find this a bit odd... But this is also beyond visone,
so we can leave it here.

Best,

Kimmo

'Hendrik' via visone-users kirjoitti 13.11.2024 klo 22.12:
> Hi Kimmo,
>
> thank you for your response and help.
>
> I would like to refer to those networks I found in Ohlendorf et al.
> (2023) (https://doi.org/10.1016/j.eist.2023.100692)
>

> 1.png2.png

> Here the authors somehow bring together the congruence network of the
> actors with storylines.
>
> Am I missing a step?
>
> Best
> Hendrik
> On Wednesday 13 November 2024 at 08:58:31 UTC+1 Kimmo Elo wrote:
>
> Hi Hendrik,
>
> being quite familiar with 'dna' (you refer to Discourse Network
> Analyser, right) I try to answer your question. I think you have a bit
> misunderstood how the export of a congruence network works. It
> creates a
> an adjacency matrix (org x org), in which "the cells represent on how
> many concepts any two actors (i e., the row organization and the column
> organization) had the same issue stance" (see
> https://usermanual.wiki/Pdf/dnamanual.2049511603.pdf

> <https://usermanual.wiki/Pdf/dnamanual.2049511603.pdf>, p. 56 for

[Hendrik]

Thank you for your help Kimmo!