Move your eye from side to side and watch what the pointer does against the scale.
01Look for the gap first: is the pointer touching the scale, or standing clear of it?It tells you whether parallax is a risk at all. A scale printed on the same surface as the thing you’re reading has none.
02Move your head until you are looking along the pointer, square to the face.Square on is the one viewing position where the eye, the pointer and the correct mark are the same line. Every other position reads a different mark.
03If the instrument has a mirror strip, move until the pointer hides its own reflection.The image moves the opposite way to the pointer as you lean, so they coincide at exactly one place, and that place is square on. It’s a test you can see rather than a position you have to guess.
04Read one end of a length at a time, moving your eye to be square on to each.Reading the far end of a rod from where you read the near end puts your eye at a slant to it. Two ends means two viewing positions, not one.
05Where you truly cannot get square, read from both sides and average.The errors are equal and opposite, so the mean of the two removes most of it. This is a last resort: it takes two readings to recover what one properly-taken reading would have given.
06Keep your eye a reasonable distance back, not right up against the scale.The closer your eye, the more sharply the sightline swings for a given sideways movement. Further back makes the geometry more forgiving of a small lean.
Laboratory
Your eye is about 8 cm to the left of the pointer, which stands 1.2 cm clear of the scale. You’re seeing the pointer against a mark roughly 4 mm to the right of where it really is. Move until you’re looking straight along the pointer, square to the scale.