Skip to main content

PROPOSAL EVALUATION AND SCORE REPORT

How Typewiser evaluates your proposal​

Typewiser reviews your draft in two ways. It assesses the quality of a single section while you are working on it, and it scores the complete proposal against the evaluation criteria of the call you are applying to. Both reviews return comments, scores and recommendations.

Neither one changes your text. You decide what to act on.

Use the section 📋evaluate function as you draft, one section at a time. Use the Score Report once you have a first full draft and want to know how a reviewer would read it.

Why Typewiser looks at how you write​

An evaluator reads your proposal for two things at once. They check whether the project meets the criteria of the call. They also form an impression of how well the case is made. Writing quality is rarely scored on its own, but a passage that is vague, unsupported or hard to follow makes the underlying idea harder to credit.

Typewiser assesses that second layer using pragmatics, the branch of linguistics that studies how meaning depends on context. Pragmatics starts from a simple observation. When two people communicate in good faith, the reader expects the writer to follow a few unspoken rules. The philosopher Paul Grice set out four of them, known as the conversational maxims: say what is true and can be backed up, say enough but no more than is needed, stay on the point, and be clear.

Those four expectations are close to what a proposal evaluator applies when reading your text. Typewiser turns them into four criteria and reports against them.

MaximTypewiser criterionThe question it answers
QUALITYTruthfulnessIs the claim supported by appropriate references?
RELATIONRelevanceIs the content aligned with the expectations?
QUANTITYInformativenessDoes the text give enough to understand it fully, without padding?
MANNERClarityIs it clear for the intended reader?
tip

You do not need the theory to use the feature. Read the four criteria as the four questions in the table above. They are the questions a reviewer asks in the margin of your proposal.

Evaluate a section​

The section review checks one section against the four linguistic criteria and leaves its findings as comments beside your text. It is quick, it is local, and you can run it as often as you like while you work on a section.

Run a section review​

  1. Open the document in the editor and go to the section you want to check.
  2. Make sure the section has content.
  3. Find the Typewiser comment attached to the section heading.
  4. Select 📋Evaluate section.
  5. Wait while the review runs. A set of agents reads the section against Truthfulness, Relevance, Informativeness and Clarity.
  6. Read the comments that appear.

Each comment names its criterion, highlights the passage in question, explains what the reviewer sees, and suggests what to do about it. Where a claim needs evidence, the comment can also list supporting sources with their links.

Recommendations typically ask you to do one of the following.

  • Improve the clarity of a passage, so an evaluator can follow it without effort.
  • Move or restructure material, so a message sits where a reader expects to find it.
  • Add a supporting source for a claim that is currently asserted rather than evidenced.
  • Develop a point that is stated but not explained, or cut one that is repeated.

You stay in control​

A comment is a reading of your text, not an instruction. You can act on it or dismiss it. If you know that the evidence sits in another document of the package, or that a passage is deliberately brief because the call limits the page count, leave it as it is. The author decides.

note

The review runs on agents that are separate from the ones that generate text. This is deliberate. A reviewer that shares the writer's assumptions finds fewer problems. As a result, the review may flag passages that Typewiser generated in the first place. That is expected behaviour, not a fault.

Request a Score Report​

The Score Report is a full, simulated evaluation of your draft. It combines two assessments. The first applies the official evaluation criteria of your specific call. The second assesses the overall quality of the writing, using the same pragmatic principles as the section review. The result reads like a strict reviewer's report written against the requirements of your call, rather than against a generic standard.

The second assessment is not part of any official score. It matters all the same, because every reviewer forms a judgement about the quality of the writing while working through the official criteria, and a poorly made case weakens the impression the project leaves.

Request the report​

  1. Complete a first draft of the proposal. The report evaluates the entire document, including how the sections fit together, so run it when there is something complete to evaluate.
  2. In the Tool bar, open the Tools menu.
  3. Select Score Report.
  4. Leave the report to compute. It assesses the whole proposal, so this takes a few minutes. You can carry on working in the meantime.
  5. You receive an email when the results are ready.
tip

Run the report on a complete draft, not on a partial one. Sections that are empty or half written are read as missing evidence, and the score reflects that.

What is inside the report​

The report is organised as a sequence of passes over your proposal.

  1. Criteria mapping. The official scoring criteria of the call, for example Excellence, Impact, and Quality and efficiency of the implementation, are broken into granular sub-requirements. Each one carries its weight, its threshold, and the evidence expected of it, including where that evidence should appear. Formal eligibility rules are mapped as well, such as consortium composition, ethics and Do No Significant Harm.
  2. Per-criterion review. For each criterion, the report locates the evidence in your proposal, rates how fully each sub-requirement is covered (full, partial, weak or absent), flags content that is missing or deficient with a severity rating (critical, major or moderate), checks whether the evidence is well positioned, and rates the writing on Truthfulness, Relevance, Informativeness and Clarity. Each criterion closes with a score and a list of what is blocking a higher one.
  3. Scoring. You get an overall weighted score and a separate score for each criterion against its own pass threshold. This shows exactly where the proposal falls short, instead of hiding a weak criterion inside an average.
  4. Inconsistency check. A separate pass looks for internal contradictions: effort or budget tables that do not match the narrative, timelines that do not hold together, claims that conflict between sections. Each finding is logged with its locations, its impact and its severity.
  5. Prioritised fixes. Findings are ranked by how much they cost you, then turned into a tiered correction plan: what to fix, where to fix it, what evidence to add, and the effect each change is expected to have on the score.
  6. Writing quality. The draft is read for how well it is written, not only for what it claims. Four axes are scored from 0 to 5: Truthfulness, Relevance, Informativeness and Clarity.
  7. Final verdict. A summary gives the overall and per-criterion scores, the strongest and weakest areas, the single highest-value fix, and whether the proposal would be likely to pass.
warning

The score is a simulation, not a prediction. Typewiser reproduces the shape of an official scoring system, but it cannot tell you whether your proposal will be funded. Read the report as a critical review whose purpose is to improve the draft.

The assessment is deliberately demanding. It is written to find problems rather than to reassure you, so a low score early on is useful information and not a reason to stop. Work through the prioritised fixes, revise, and run the report again when the draft has moved on.