Appendix · Methodology
How the report gets made.
The same pipeline, shown one step at a time. It narrows the public commentary on the 2024 loss down to 8,843 quotes sorted into 24 top-level themes. Every quote links back to the moment it was said.
- Sources
- 98
- Videos
- 203
- Hours
- 258.5
- Quotes
- 8,843
- Reach
- 118M
Cast the net.
We ran a set of search terms through YouTube: "why Democrats lost", "2024 election post-mortem", "Harris campaign", "coalition collapse" and "Trump 2024". The first pass is deliberately wide. Each step after it narrows the set.
candidate channels surveyed
Plot every voice on the compass.
We sample each candidate channel, score it on the Political Compass from its own content, and audit the score before including it. The sample is spread on purpose across the spectrum, audience sizes and formats, so small channels with a distinct view aren't drowned out by the largest ones.
sources after audit
spectrum tiers
CNN · audited
economic−2.0
social+3.0
tierCenter
Score from sampled segments. Audited 2026-04.
Lift claim-level quotes from raw transcripts.
We transcribe every video. Models pick out claim-level statements, each with a stance and a confidence rating, and pin each one to its timestamp, so every quote links to the exact moment it was said. Human reviewers check the borderline cases.
videos transcribed
hours of audio
quotes extracted
"That is why Democrats underperformed. They did not own the economic reality of paying at the gas, grocery, rent and mortgage counter… and you played down people's pain."
Let the themes emerge.
We group the 8,843 quotes from the bottom up, into the categories the commentary itself fell into. The top-level themes and sub-themes come out of the data. We write the taglines and edit the prevalence ranking.
top-level themes
sub-themes
- 01 Flawed Strategy & Tactical Incompetence 436
- 02 Neglected Coalition & Demographic Collapse 378
- 03 Ineffective Economic & Policy Messaging 309
- 04 Elitist Culture & 'Woke' Alienation 266
- 05 Flawed Candidacy & Leadership Vacuum 265
- 06 Internal Party Dysfunction & Organizational Decay 244
- +18 more top-level themes…
Layer the audience back in.
Each video's views, likes and comments are joined back to the quotes taken from it, so engagement can be shown by political leaning as well as by channel. Daily timelines come from publication dates.
views
likes
comments
- farLeft 2.28M
- left 45M
- center 43M
- right 1.07M
- farRight 26M
Dataset
2024 Election Analysis
| Range | 2024-11-05 → 2025-05-05 |
|---|---|
| Sources | 98 |
| Videos | 203 (258.5 hours) |
| Quotes | 8,843 |
| Themes | 24 top-level / 121 hierarchical |
| Corrections | 1 test row and 1 out-of-scope theme removed · 1 channel score corrected · 6 unscored channels placed by hand |
| SHA | 185f02c64ecf |
Frequently asked
Common questions about the method.
- What dates does the study cover?
- The study covers videos published between November 6, 2024 and July 22, 2025.
- How is engagement data reported?
- View, like and comment counts come from a YouTube snapshot taken on May 27, 2026. They have changed since.
- How are political leanings determined?
- Each source is scored on the Political Compass by sampling transcripts of its content, and audited before inclusion. We don't use self-reported labels. Where the scoring was blank or clearly wrong, we placed the channel by hand; the dataset table above lists those corrections.
- How is quote extraction done?
- Transcripts are processed with a mix of Gemini, Claude and OpenAI models, followed by human review. Each quote is stored with its timestamp and the claim it supports.
- What was the primary research question?
- "How do political commentators across the spectrum explain the 2024 election outcome: what they blame, who they think should change, and what they think comes next?"
- Is the full source and content list available?
- Yes. The Sources page lists all 98 channels, and every video is on the Videos page. Each pull quote's "Why we trust this" panel links to the video at the moment the quote was said.