Why I Stopped Asking People to Proofread Everything They Submit
I spent about eight years screening candidates and vendor proposals before I realized the spelling filter was doing more harm than good. Not because spelling doesn't matter. It does. But it matters as one data point among dozens, and treating it as a primary filter meant I consistently overlooked people who could do the actual work while letting polished mediocrity through the front door. Signaling theory in economics explains what's happening here without getting academic about it. When someone produces well-spelled, grammatically clean text, they are sending a costly signal. Good spelling and grammar don't come free. They require years of exposure to standard conventions, usually through formal education or deliberate practice. That makes them a reasonable proxy for educational background and attention to detail in controlled environments. The problem starts when you treat that proxy as the thing itself.
The Great Signaling vs Signaling Showdown Why Spelling Matters In A Globalized World
In a globalized economy, the spelling signal gets even more complicated because standard English conventions aren't universal. British spelling, American spelling, Indian English, Singaporean English — these are all legitimate variants with their own internal consistency. When I started recruiting internationally around 2018, I caught myself rejecting candidates whose documents used "organisation" instead of "organization" or who placed commas in ways that felt wrong to my American sensibilities but were correct under the style guides they grew up with. That was a bad look in retrospect. Here's the practical reality most people miss: in many industries, particularly software development, data analysis, and technical writing, spelling accuracy correlates poorly with job performance. I once interviewed a senior engineer whose cover letter had three spelling errors. Her code sample was elegant, her system architecture explanation was clear, and she asked questions that revealed deep practical knowledge. She also got rejected by our automated screening tool before I ever saw her application. That tool was flagging her resume based on keyword matching and formatting expectations that had nothing to do with engineering competence. I had to manually override it twice before she got a proper interview. The reverse happens too, and it's worse. I've seen candidates with immaculate spelling and perfect grammar deliver work that was structurally unsound, logically incoherent, or completely missing the point of the assignment. Spelling is a surface-level signal. It's easy to achieve with autocorrect and hard to fake if you don't have the underlying skill. That's why it persists as a hiring filter despite being a flawed one.
There's another layer most people don't consider. AI writing assistants have fundamentally changed the cost structure of good spelling. Tools like Grammarly, Copilot, and various LLM-powered editors can now produce publication-quality prose from a rough draft in seconds. That means the signaling value of spelling is decaying rapidly. What was once a reliable indicator of education and effort is increasingly just an indicator of whether someone has access to the right tools. Not everyone does. That creates its own form of bias. I developed a workaround after the engineer incident. Instead of screening for spelling perfection, I started using work samples as the primary filter. A five-minute coding task, a one-page analysis of a real dataset, a brief written explanation of how they'd approach a problem in their field. These are much harder to fake with an AI editor and they correlate significantly better with actual performance. The tradeoff is time. Evaluating work samples properly takes about 20 to 30 minutes per candidate compared to the 2 minutes it takes to skim a resume for red flags. Over a busy hiring quarter, that adds up quickly. But the quality of hires improved noticeably after the switch. For individuals trying to navigate this landscape, the advice is straightforward but not always comfortable. If you're applying to jobs and your spelling isn't polished, fixing it matters — at least for the initial screening stage. Most Applicant Tracking Systems and overworked recruiters will not give you the benefit of the doubt on basic mechanics. Use whatever tools you need. Run it through a checker. Have a friend review it. Then move on. Don't spend three hours obsessing over whether "affect" or "effect" is correct when your actual experience is what determines whether you get called in.
Get the Full Details
On the other side of the table, if you're the one making hiring or vendor selection decisions, treat spelling as a minor signal, not a major one. Note it if something stands out badly, but don't let it be a decisive factor. The people who benefit most from strict spelling filters are usually the ones who already have advantages — access to good schools, time to practice, native-language fluency. The people who suffer most are the ones whose actual contributions would likely exceed their typo count. The globalized workplace has made this tighter. Communication across languages and cultures happens constantly now. A developer in Lagos writing to a product manager in Berlin will make spelling choices influenced by both Nigerian English conventions and German-influenced sentence structures. That's not a problem to fix. That's the reality of how professional communication actually works. The signal gets noisy in those contexts, and noisy signals make for unreliable filters. Bottom line: spelling matters, but not as much as the people who designed hiring processes seem to think it does. Pay attention to it. Don't let it run the show.
