ANALYSIS: Trump at -20.6: What a New Second-Term Low Actually Measures

Introduction

On 7 August 2026, Silver Bulletin put Donald Trump’s net approval at -20.6 points, calling it a new second-term low. The figure is hard because it is an aggregate, not a ballot box. It collects surveys with different field dates, samples and methods; it cannot by itself forecast an election or explain every cause of a public mood. The date, the source and the stated limit set the terms of this article.

The hard fact comes first. The unsupported conclusion does not get to follow it.

This analysis follows the assigned record rather than turning a headline into a verdict. Each section names a documented element, its consequence, and the question left open. That method is not restraint for its own sake. It is the difference between reporting a public record and manufacturing one.

The -20.6 floor is a measure, not an election

Silver Bulletin’s dated low

In the dated record, Silver Bulletin reports that Silver Bulletin said Trump’s net approval reached -20.6 points on 7 August 2026, below its prior second-term readings. This supplies a dated fact with a defined source for the -20.6 floor is a measure, not an election. It is not decorative context; it is the piece of the record that makes the -20.6 floor is a measure, not an election answerable. A low is evidence of weakness. It is not a result.

The consequence is practical: a single aggregate cannot supply a constituency-by-constituency forecast. For the -20.6 floor is a measure, not an election, the article gives the stated evidence its full weight while refusing an invented extension. For the -20.6 floor is a measure, not an election, scope, attribution and timing decide what the evidence can honestly establish. No extra certainty is needed for the fact to matter.

A net rating has two sides

What can be verified here is The same update put the share of respondents who were strongly favourable at 21%. Read beside the first item on the -20.6 floor is a measure, not an election, it adds a second verifiable detail rather than a rhetorical echo. Its value within the -20.6 floor is a measure, not an election is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The fact changes the frame in one way only: the material on the -20.6 floor is a measure, not an election has a defined reach. In the -20.6 floor is a measure, not an election, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in the -20.6 floor is a measure, not an election. The distinction makes the conclusion durable.

The record has a direction. It does not surrender its limits.

January 2021 is a comparison, not a copy

The earlier -19 mark

At the centre of this section, Silver Bulletin reports that Silver Bulletin compared the new figure with -19 after the events of 6 January 2021 during Trump’s first term. This supplies a dated fact with a defined source for january 2021 is a comparison, not a copy. It is not decorative context; it is the piece of the record that makes january 2021 is a comparison, not a copy answerable. A number travels with its date. So does its context.

That distinction matters because historical resemblance is not proof of identical public motives. For january 2021 is a comparison, not a copy, the article gives the stated evidence its full weight while refusing an invented extension. For january 2021 is a comparison, not a copy, scope, attribution and timing decide what the evidence can honestly establish. The record is stronger when it keeps its edge.

Different moments carry different causes

On this part of the timeline, The comparison comes from one historical series; it does not establish that the political causes of the two downturns are identical. Read beside the first item on january 2021 is a comparison, not a copy, it adds a second verifiable detail rather than a rhetorical echo. Its value within january 2021 is a comparison, not a copy is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

It is enough to establish a pressure point, since the material on january 2021 is a comparison, not a copy has a defined reach. In january 2021 is a comparison, not a copy, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in january 2021 is a comparison, not a copy. The date belongs inside the analysis, not beneath it.

A date can anchor a judgment without completing it.

37.8% and 59.3% show the aggregate gap

A second published compilation

The file first establishes that the Wikipedia polling compilation reports that A Wikipedia polling compilation dated 7 August 2026 listed 37.8% approval and 59.3% disapproval. This supplies a dated fact with a defined source for 37.8% and 59.3% show the aggregate gap. It is not decorative context; it is the piece of the record that makes 37.8% and 59.3% show the aggregate gap answerable. The gap is real in the compilation. Its method still matters.

Its immediate meaning is limited: a compilation is not a universal census of the electorate. For 37.8% and 59.3% show the aggregate gap, the article gives the stated evidence its full weight while refusing an invented extension. For 37.8% and 59.3% show the aggregate gap, scope, attribution and timing decide what the evidence can honestly establish. That limit is part of the public fact.

The arithmetic is plain

The documented detail is Subtracting the two produces a gap close to the Silver Bulletin direction, while still reflecting the surveys that compilation selected. Read beside the first item on 37.8% and 59.3% show the aggregate gap, it adds a second verifiable detail rather than a rhetorical echo. Its value within 37.8% and 59.3% show the aggregate gap is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The record has force precisely because the material on 37.8% and 59.3% show the aggregate gap has a defined reach. In 37.8% and 59.3% show the aggregate gap, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in 37.8% and 59.3% show the aggregate gap. The known fact carries enough weight on its own.

The strongest conclusion is the one the evidence can carry.

CNN’s tracker lands at 36% to 63%

A separate aggregation

For this precise point, CNN Poll Tracking reports that CNN’s polling tracker displayed 36% approval against 63% disapproval in the material assigned to this article. This supplies a dated fact with a defined source for cnn’s tracker lands at 36% to 63%. It is not decorative context; it is the piece of the record that makes cnn’s tracker lands at 36% to 63% answerable. Approval measures a climate. Elections allocate seats.

The public implication is clear: national approval and local electoral outcomes are different questions. For cnn’s tracker lands at 36% to 63%, the article gives the stated evidence its full weight while refusing an invented extension. For cnn’s tracker lands at 36% to 63%, scope, attribution and timing decide what the evidence can honestly establish. The evidence deserves the discipline of its own scope.

One tracker is not every tracker

The available evidence places That reading makes disapproval the larger response in that tracker, but it does not reveal how every congressional district would behave. Read beside the first item on cnn’s tracker lands at 36% to 63%, it adds a second verifiable detail rather than a rhetorical echo. Its value within cnn’s tracker lands at 36% to 63% is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

That is not a minor qualification: the material on cnn’s tracker lands at 36% to 63% has a defined reach. In cnn’s tracker lands at 36% to 63%, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in cnn’s tracker lands at 36% to 63%. The boundary is not evasive; it is accurate.

Facts do not need an invented echo to hit hard.

The Economist’s 35% to 60% widens the picture

Another published average

The reporting identifies The Economist aggregation reports that The Economist-linked aggregation in the dossier put approval at 35% and disapproval at 60%. This supplies a dated fact with a defined source for the economist’s 35% to 60% widens the picture. It is not decorative context; it is the piece of the record that makes the economist’s 35% to 60% widens the picture answerable. Different tools can point the same way without matching exactly.

The analytical burden is therefore simple: the dossier does not provide a single method that makes every tracker interchangeable. For the economist’s 35% to 60% widens the picture, the article gives the stated evidence its full weight while refusing an invented extension. For the economist’s 35% to 60% widens the picture, scope, attribution and timing decide what the evidence can honestly establish. Nothing is gained by making the file say more.

Method creates variation

The reported sequence shows Its difference from CNN and other trackers is not a flaw to hide; field windows, weighting and which surveys enter an average can change the final snapshot. Read beside the first item on the economist’s 35% to 60% widens the picture, it adds a second verifiable detail rather than a rhetorical echo. Its value within the economist’s 35% to 60% widens the picture is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The distinction protects the account: the material on the economist’s 35% to 60% widens the picture has a defined reach. In the economist’s 35% to 60% widens the picture, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in the economist’s 35% to 60% widens the picture. The article does not need a larger claim to be clear.

A source is part of the story, not a decoration beneath it.

RealClearPolitics reaches 39.3% to 58.4%

The higher approval reading

What can be verified here is RealClearPolitics reports that RealClearPolitics showed 39.3% approval and 58.4% disapproval in the cited period. This supplies a dated fact with a defined source for realclearpolitics reaches 39.3% to 58.4%. It is not decorative context; it is the piece of the record that makes realclearpolitics reaches 39.3% to 58.4% answerable. A convenient number is still only one number.

This prevents a shortcut, because choosing the most favourable tracker would not settle the wider evidence. For realclearpolitics reaches 39.3% to 58.4%, the article gives the stated evidence its full weight while refusing an invented extension. For realclearpolitics reaches 39.3% to 58.4%, scope, attribution and timing decide what the evidence can honestly establish. The difference between claim and proof remains material.

Selection changes the frame

For readers following the chronology, That is a less severe approval figure than some other trackers, but it remains a finding of net negative sentiment in its own selected series. Read beside the first item on realclearpolitics reaches 39.3% to 58.4%, it adds a second verifiable detail rather than a rhetorical echo. Its value within realclearpolitics reaches 39.3% to 58.4% is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The point is not to weaken the evidence; it is to keep it exact: the material on realclearpolitics reaches 39.3% to 58.4% has a defined reach. In realclearpolitics reaches 39.3% to 58.4%, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in realclearpolitics reaches 39.3% to 58.4%. The available record still demands attention.

The difference between a claim and proof is where accountability begins.

Marquette’s 1,076-adult survey moves differently

The late-July fieldwork

On this part of the timeline, Marquette Law School via CNN Poll Tracking reports that Marquette Law School surveyed 1,076 adults from 22 to 29 July, with a stated margin of error of ±3.2 points. This supplies a dated fact with a defined source for marquette’s 1,076-adult survey moves differently. It is not decorative context; it is the piece of the record that makes marquette’s 1,076-adult survey moves differently answerable. One survey can rise while an average falls.

The wording sets a boundary: a two-point movement must be read alongside the stated margin of error. For marquette’s 1,076-adult survey moves differently, the article gives the stated evidence its full weight while refusing an invented extension. For marquette’s 1,076-adult survey moves differently, scope, attribution and timing decide what the evidence can honestly establish. Precision does not reduce the consequence.

Sampling uncertainty stays visible

The narrow fact is Published on 6 August, it recorded 40% approval, up from 38% a month earlier in the CNN tracking cited by the dossier. Read beside the first item on marquette’s 1,076-adult survey moves differently, it adds a second verifiable detail rather than a rhetorical echo. Its value within marquette’s 1,076-adult survey moves differently is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The consequence is practical: the material on marquette’s 1,076-adult survey moves differently has a defined reach. In marquette’s 1,076-adult survey moves differently, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in marquette’s 1,076-adult survey moves differently. No extra certainty is needed for the fact to matter.

The file is not weakened by its limits. It is protected by them.

Reuters/Ipsos brings a larger sample and a lower reading

The 4,505-adult poll

The documented detail is Reuters/Ipsos via CNN Poll Tracking reports that Reuters/Ipsos surveyed 4,505 adults from 29 July to 3 August and reported a margin of error of ±1.5 points. This supplies a dated fact with a defined source for reuters/ipsos brings a larger sample and a lower reading. It is not decorative context; it is the piece of the record that makes reuters/ipsos brings a larger sample and a lower reading answerable. The sample got larger. The answer got colder.

The fact changes the frame in one way only: a narrower stated margin does not make questionnaires, dates and weighting identical. For reuters/ipsos brings a larger sample and a lower reading, the article gives the stated evidence its full weight while refusing an invented extension. For reuters/ipsos brings a larger sample and a lower reading, scope, attribution and timing decide what the evidence can honestly establish. The distinction makes the conclusion durable.

Precision does not erase design choices

The underlying record states Its result was 35% approval and 63% disapproval, a more adverse snapshot than Marquette’s. Read beside the first item on reuters/ipsos brings a larger sample and a lower reading, it adds a second verifiable detail rather than a rhetorical echo. Its value within reuters/ipsos brings a larger sample and a lower reading is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

That distinction matters because the material on reuters/ipsos brings a larger sample and a lower reading has a defined reach. In reuters/ipsos brings a larger sample and a lower reading, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in reuters/ipsos brings a larger sample and a lower reading. The record is stronger when it keeps its edge.

Precision is not hesitation when public consequences follow.

The 32% to 44% spread forbids false precision

A range across cited pollsters

The available evidence places the assigned pollster range reports that The fact block records approval readings from 32% to 44% among cited pollsters in late July and early August. This supplies a dated fact with a defined source for the 32% to 44% spread forbids false precision. It is not decorative context; it is the piece of the record that makes the 32% to 44% spread forbids false precision answerable. The spread does not cancel the trend. It disciplines it.

It is enough to establish a pressure point, since not every pollster in the compilation published an individual margin of error. For the 32% to 44% spread forbids false precision, the article gives the stated evidence its full weight while refusing an invented extension. For the 32% to 44% spread forbids false precision, scope, attribution and timing decide what the evidence can honestly establish. The date belongs inside the analysis, not beneath it.

The interval is part of the story

In the dated record, That range is wide enough to rule out language that treats one percentage as an uncontested national measurement. Read beside the first item on the 32% to 44% spread forbids false precision, it adds a second verifiable detail rather than a rhetorical echo. Its value within the 32% to 44% spread forbids false precision is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

Its immediate meaning is limited: the material on the 32% to 44% spread forbids false precision has a defined reach. In the 32% to 44% spread forbids false precision, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in the 32% to 44% spread forbids false precision. That limit is part of the public fact.

One measured fact can outweigh a page of unproven certainty.

Why the net number feels harsher

Approval versus disapproval

The reported sequence shows Silver Bulletin reports that A net approval figure compares favourable and unfavourable responses rather than reporting either one alone. This supplies a dated fact with a defined source for why the net number feels harsher. It is not decorative context; it is the piece of the record that makes why the net number feels harsher answerable. A balance is not a biography.

The record has force precisely because the evidence does not identify a single cause for every respondent’s view. For why the net number feels harsher, the article gives the stated evidence its full weight while refusing an invented extension. For why the net number feels harsher, scope, attribution and timing decide what the evidence can honestly establish. The known fact carries enough weight on its own.

Net ratings compress two movements

At the centre of this section, Silver Bulletin’s -20.6 therefore communicates a balance between two measured opinions, not a diagnosis of why people gave those answers. Read beside the first item on why the net number feels harsher, it adds a second verifiable detail rather than a rhetorical echo. Its value within why the net number feels harsher is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The public implication is clear: the material on why the net number feels harsher has a defined reach. In why the net number feels harsher, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in why the net number feels harsher. The evidence deserves the discipline of its own scope.

A public record is not a blank cheque for interpretation.

What the pollster methods do not share

Field dates and samples

For readers following the chronology, Marquette and Reuters/Ipsos reports that Marquette’s fieldwork ended on 29 July, while Reuters/Ipsos ran through 3 August; the two surveys also used samples of 1,076 and 4,505. This supplies a dated fact with a defined source for what the pollster methods do not share. It is not decorative context; it is the piece of the record that makes what the pollster methods do not share answerable. The dates are not decoration. They define the measurement.

That is not a minor qualification: the article cannot reconstruct methods not supplied in the fact block. For what the pollster methods do not share, the article gives the stated evidence its full weight while refusing an invented extension. For what the pollster methods do not share, scope, attribution and timing decide what the evidence can honestly establish. The boundary is not evasive; it is accurate.

Averages smooth differences

The file first establishes that Those differences explain why a responsible reading keeps methods beside results instead of treating all percentages as identical observations. Read beside the first item on what the pollster methods do not share, it adds a second verifiable detail rather than a rhetorical echo. Its value within what the pollster methods do not share is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The analytical burden is therefore simple: the material on what the pollster methods do not share has a defined reach. In what the pollster methods do not share, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in what the pollster methods do not share. Nothing is gained by making the file say more.

The unanswered question belongs in the article too.

What the record can say about Trump’s standing

Several measures, one direction

The narrow fact is the four named aggregators reports that Silver Bulletin, CNN, The Economist and RealClearPolitics all appear in the assigned record, with distinct approval and disapproval pairs. This supplies a dated fact with a defined source for what the record can say about trump’s standing. It is not decorative context; it is the piece of the record that makes what the record can say about trump’s standing answerable. The evidence converges. It does not become uniform.

The distinction protects the account: agreement on direction is not a licence to erase methodological differences. For what the record can say about trump’s standing, the article gives the stated evidence its full weight while refusing an invented extension. For what the record can say about trump’s standing, scope, attribution and timing decide what the evidence can honestly establish. The article does not need a larger claim to be clear.

No single instrument owns the truth

For this precise point, Their readings vary, but each cited aggregate places the reported disapproval level above reported approval in its own tally. Read beside the first item on what the record can say about trump’s standing, it adds a second verifiable detail rather than a rhetorical echo. Its value within what the record can say about trump’s standing is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

This prevents a shortcut, because the material on what the record can say about trump’s standing has a defined reach. In what the record can say about trump’s standing, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in what the record can say about trump’s standing. The difference between claim and proof remains material.

The evidence is enough to demand attention, not fabrication.

The question beyond the floor remains open

A current mood is not a forecast

The underlying record states the dated polling record reports that The 7 August 2026 low is a dated statistical event, while the Marquette reading shows why later polling must be watched rather than assumed away. This supplies a dated fact with a defined source for the question beyond the floor remains open. It is not decorative context; it is the piece of the record that makes the question beyond the floor remains open answerable. The floor is documented. The verdict is not.

The point is not to weaken the evidence; it is to keep it exact: future electoral behaviour is not established by the current approval files. For the question beyond the floor remains open, the article gives the stated evidence its full weight while refusing an invented extension. For the question beyond the floor remains open, scope, attribution and timing decide what the evidence can honestly establish. The available record still demands attention.

The next data point can change the line

The reporting identifies Nothing in the assigned material converts these surveys into a prediction of turnout, seats or a final vote. Read beside the first item on the question beyond the floor remains open, it adds a second verifiable detail rather than a rhetorical echo. Its value within the question beyond the floor remains open is concrete: it identifies what the public record contains and keeps the reader from mistaking emphasis for proof.

The wording sets a boundary: the material on the question beyond the floor remains open has a defined reach. In the question beyond the floor remains open, a disciplined reading separates the reported fact from a reasonable inference and from an unknown requiring further evidence. This article does not hide that third category in the question beyond the floor remains open. Precision does not reduce the consequence.

The next decision must answer the fact already on the page.

Conclusion

The assigned record supports a hard conclusion: On 7 August 2026, Silver Bulletin put Donald Trump’s net approval at -20.6 points, calling it a new second-term low. The evidence establishes that public fact and the concrete consequences described above. It does not authorise a result, motive or future outcome that the sources have not documented.

Proof is not softer than rhetoric. It is what makes a hard judgment survive.

That is the standard for this analysis: keep the documented fact intact, keep the unresolved question visible, and judge the next decision by what it does rather than by a story forced onto the record. The record is demanding enough. It must now be answered.

Signature

Signed Maxime Marquette, columnist

Columnist's Transparency box

Editorial positioning

This article is pro-democracy, pro-rule of law and, where relevant, pro-Ukraine and pro-Western. That editorial position never upgrades a reported statement into an established fact or treats a warning as proof of a completed event.

Methodology and sources

The text uses only the assigned fact block and the linked sources listed below. It keeps dates, attribution, methodological limits and unconfirmed elements visible instead of filling gaps with supposition.

Nature of the analysis

This is a column of analysis, not a claim of firsthand reporting. It distinguishes documented facts, attributed statements and reasonable interpretation; where a fact remains unconfirmed, future or unresolved, that status is stated plainly.

Sources

Primary sources

The assigned fact block supplies no directly linked primary source for this article.

Secondary sources

• Silver Bulletin — Trump approval update — 7 August 2026
• Polling compilation — second Trump presidency — 7 August 2026
• CNN Poll Tracking — approval tracker — 6 August 2026
• The New York Times — approval-poll table — August 2026
• USA Today — poll context — 6 August 2026
• CNN Poll Tracking — Marquette and Reuters/Ipsos data — 6 August 2026