This is an attempt on a preliminary analysis supporting the presentation by Boda Zsolt at the POPREBEL policy event.
“Preliminary” is an important disclaimer, because we have not done a data audit yet. I have, however, cleaned up the #ethno-icqe2022 part of the corpus, and added it to the main #ethno-poprebel corpus. The latter consists of 5,904 contributions in 499 topics, for a total of 809,000 words. The CCN contains 1,691 codes and 46,965 (stacked) edges. The total number of co-occurrences is just over 100,000.
The research question is "how do affects surface in discussing politics in this corpus? @Nica and I decided to do a sort of ego network analysis, much like we did in the ICQE22 paper. She chose five codes; their ego networks should provide the beginning of an answer. The four are:
I have chosen to reduce by Simmelian backbone, for two reasons. The first: Simmelian backbones emphasize closely-knit communities of codes. By looking at other codes in the same communities, we can catch an idea of the kind of “company” our starting codes keep. In theory, we could do this with other reduction techniques too, but those would focus on high-depth and high-breadth edgeds rather than on the (relatively unreduced) communities to which these codes belong. I write “relatively unreduced” because Simmelian backbone extraction reduces the edges bridging across communities, leaving closely-knit communities more or less intact.
The second reason: these codes do not belong to very deep or very broad edges.
Below is a visualization of the CCN’s Simmelian backbone with r > 24 (850 codes, 14,708 edges). This is the highest value of r for which all four codes still appear. The maximum redundancy value of incident edges to each code is a rough indicator of how closely-knit (in the Simmelian sense) is the community to which that code belongs. These values are:
The easiest one is
anxiety, which is part of a tightly-knit community even at these high levels of edge redundancy. Besides
anxiety itself and two meta-codes (
Z Emotions - Category), it contains 74 other codes.
list of the codes in the same community as `anxiety` with *r* > 24 (`anxiety` itself is omitted)
military occupation labour working remotely city big cities variety pressure to succeed sense of purpose daily routine sense of personal space despondency placemaking equity inadequate income irrationality historical traumas police brutality anxiety increasing use gentrification public space sense of loss FOMO bilingual chance encounter urban geography belonging intersectionality freelance work improving places following passion cultural diversity democracy motivation village small changes unemployment sense of agency EU PPR urban infrastructure change "just city" feminist geography delayed projects social life social engagement performing rituals women mindfulness anti-ownership encountering difference ignoring rules cosmopolitan racism self-worth recognising shared humanity *sense of safety small town lack of involvement 'happiness' bodily reaction taking time off
Since I have later used the threshold value of 20 for r, here is a list that does the same, biut with edges for which r > 20. There are about 30 extra codes, but the flavour does not seem to change to my untrained eye.
list of the codes in the same community as `anxiety` with *r* > 20 (`anxiety` itself is omitted)
political inaction public space irrationality racism Germans *frustration sense of agency women big cities self-worth police brutality vaccinations sense of threat EU FOMO working remotely inadequate income social engagement unemployment *economic crisis despondency *sense of safety cosmopolitan taboo support groups homosexuality daily routine mindfulness change police state small town village city social life labour freelance work taking new approach 'good examples to follow' visibility suppression ignoring problems taking individual action 'alone against the system' self-righteous political problems 'escape from problems' environmental problems initiatives "hard to plan ahead" instrumentalisation of social problems left-wing politics wybory życiowe educational inequality ignoring rules eliciting emotions democracy sense of community self-actualisation recognising shared humanity security sense of loss chance encounter encountering difference belonging *building communities delayed projects small changes sense of decrease deliberate desensitization pessimism cultural diversity following passion variety 'happiness' urban infrastructure change gentrification sense of personal space bodily reaction increasing use performing rituals sense of purpose pressure to succeed motivation intersectionality equity "just city" lack of involvement anti-ownership improving places placemaking taking time off bilingual urban geography feminist geography
trust has only one edge connecting it to the reduced CCN when r > 24 (it goes to
access to information). However, if I reduce slightly less (r > 20), I get it as part of another very clear community:
list of the codes in the same community as `trust` with *r* > 20
access to information feeling restricted worry *democratic state (critique) changing rules *good health lack of clear guidance lack of agreement justifying decisions
I did the same thing with
empathy, except that I chose a lower threshold value (r > 20) to include some extra codes. At this level of reduction,
empathy has 20 neighbors. I should emphasize that these neighbors are not exacly a community in the sense of the very clear cluster that
anxiety is part of.
list of the codes connected to `empathy` by edges with *r* > 20
world is your oyster lockdown (lack of) affordable housing tolerance privilege Catholic Church France coping with death salary satisfaction internet users escaping "bubbles" "bubbles" emotionally difficult traditional values homophobia judgmental culture mental issues alleviating polarisation seeking civilised debate sexism
anger in the exact same way as
list of the codes connected to `anger` by edges with *r* > 20
wyjaśnienie naukowe entrepreneurs consumerism education system WHO vaccinations lockdown government inefficiency mental issues pedophilia *depression disappointment private healthcare religious upbringing climate change cynicism demonization financial crisis spirituality NGOs pandemic privilege universities parenting unpaid labour
@nica, any thoughts?