Confidence gaps in STEM subjects rarely come from lack of ability. They come from specific, fixable gaps in how students were taught to approach the subject.
"I'm just not a science person" is one of the most common things a STEM tutor hears, almost always from a student who is entirely capable of the subject. The actual cause is usually narrower and more fixable than the student assumes: a specific type of question, a particular topic taught badly at some earlier point, or simply never having been shown how to approach unfamiliar problems methodically.
Generic confidence-building, encouragement, reassurance, "you can do this", has limited effect if the underlying skill gap isn't addressed directly. Confidence in STEM subjects reliably follows from students experiencing genuine, repeated success on problems they previously found difficult, not from being told they're capable in the abstract.
Many students who struggle with STEM subjects aren't struggling with content recall. They're struggling with problems presented in an unfamiliar structure or context, even when the underlying concept is one they actually know. This is a specific, teachable skill (pattern recognition across problem types), not a fixed trait, and it responds well to deliberate practice with varied question formats.
Each STEM transition raises the expected level of independent problem-solving substantially, often faster than confidence has caught up. A student who was comfortable at GCSE can genuinely struggle at A-Level not because the content is unmanageable, but because the expected way of engaging with problems has changed and nobody has explicitly named that shift.
Deliberately choosing problems just above a student's current comfort level, working through the specific reasoning that connects a familiar concept to an unfamiliar question format, and tracking visible progress over weeks, not vague encouragement, is what reliably rebuilds STEM confidence, because it gives students real evidence rather than reassurance.