Upper House Election 2025Election.comVote Matching
https://gyazo.com/aa48471a2bd225711b50a68d5ab9e6f7
https://votematches.go2senkyo.com/sangiin_2025/
[About the political parties and organizations listed
The following is a list of political parties and organizations that have responded to the Election.com policy questionnaire at this time.
Japan Reform Party is not listed because it has not responded at this time. We will post their responses as soon as we receive them.
This ballot matching is for political organizations that have filed lists for proportional representation in addition to national political parties that meet the party requirements. Political organizations that have candidates in only some electoral districts are not listed.
● 27th House of Councillors Election Calculation of Vote Matching
The "distance" between the party's answer and the user's choice is used to calculate the degree of matching. For example, if Party A answers "Agree" and Party B answers "Slightly Agree" to a question for which the user chose "Agree," Party A will receive more points than Party B because of the greater distance between the two parties' answers.
*(7/13) We received a response from the Japan Seishinkai and have added it to the ballot matching.
*(7/6) We received a response from Team Mirai and have added it to the ballot matching.
2025-07-14
https://www.facebook.com/kuchii.jun/posts/pfbid02n4znvqwg4Tr8g1FgDQiJqpE34ZMjFKrZpBAbXcnPkSwdsR92j1coe3LareSk1moXl
Very interesting analysis (I wanted to do it myself, but I don't have time...)nishio.icon
Perhaps there has been a bias that "observed opinions are biased toward the radical side" due to the fact that "the more radical the opinion, the more likely it is to be communicated"
The political parties made policy decisions in accordance with the "radical distribution of opinion" based on that biased observation, and as a result, each party's opinion became more radical.
The distribution of opinions among "those who use vote matching" rather than "those who transmit opinions" was actually more "no strong opinions"
As a result, the probability of clustering users into parties that give moderate answers to radical so-called "ideological" questions of division may have increased
This time, it's possible that it's vote matching optimization in a search engine optimization sense, although it's not being triggered for this itself (first published: Diary 2025-07-14 , moved).
I think we will have to have a discussion soon about what kind of vote-matching mechanism is preferable for the public. I think this disclosure of information by election.com is a good starting point for the discussion. (First published in Diary 2025-07-16 , copied)
2025-07-20
20 5-step questions
Criticality Survey of Enforcement
Of the following 20 questions you answered, please select three that are of particular importance to you. (Required)
If you choose all central, and the first three for the mandatory importance survey.
https://gyazo.com/fbb439357de15a2d97337bddf90b4531https://gyazo.com/f1a4b53fb5521f44a8064d1b2cad38ad
https://gyazo.com/ec85b5ad86ce4150b322b660a7586c2d
https://gyazo.com/86c2df70449c038db8d079eb4f7cdada
After researching the phenomenon of Team Mirai being unexpectedly suggested by Election.com's vote matching, the observation was made that when neutral is selected for all answers, the match with Team Mirai is 79%.
While public opinion as visualized on social networking sites is divided, the silent majority is more moderate, and Team Mirai's moderate responses to questions that are fueled by a policy of not fueling division seems to have matched those silent majority clusters as a result.
Related discussions: Division and the Silent Majority.
2025-07-24
The opinion of a political party or user can be regarded as a point in a -2~+2 hypercube in 20 dimensional space
Probably the Euclidean distance.
The three focus questions would expand that axis.
Each party divides the space
By considering "everyone's aggregate result" as the amount of points in the space, we can roughly obtain the distribution of users' opinions.
It's messy to assume a uniform distribution within each space, but it should work.
I'm not sure if it's possible to consider it as a superposition of normal distributions and use the EM algorithm to estimate it.
Once the distribution has been estimated, we can calculate "how to change the answer to any question to increase support."
Vote Matching Optimization
---
This page is auto-translated from /nishio/参院選2025選挙ドットコム投票マッチング using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.