Author: eduKate Asia
-
PSLE Science Reality Lab Vol No.458 | “98% Accurate on the Training Data” — Has the Model Proven It Can Predict New Cases?
A Primary 5/6 Reality Lab guide to model accuracy claims: learn why a high training score does not prove performance on new cases, and how to check independent test evidence, leakage and scope.
-
PSLE Science Reality Lab Vol No.457 | “12-Month Running Mean” — Did the Raw Measurements Really Change That Smoothly?
A Primary 5/6 Reality Lab guide to running means and moving averages: learn why a smooth scientific line is a calculated summary, not the raw measurement path, and how to test claims against the original data.
-
PSLE Science Reality Lab Vol No.456 | “HDOP = 0.8” — Does That Mean the GPS Error Is 0.8 Metres?
A Primary 5/6 Reality Lab guide to reading HDOP on GPS/GNSS displays: separate satellite-geometry dilution from metres of position error, then test the wider evidence before trusting an accuracy claim.
-
PSLE Science Reality Lab Vol No.455 | “This Fluorescence Image Is Bright Green” — Were the Cells Naturally Green?
A Primary 5/6 Reality Lab guide to fluorescence-image pseudocolour: separate measured signal from display colour, check channel labels and scaling, and avoid treating a scientific image palette as the specimen’s natural colour.
-
PSLE Science Reality Lab Vol No.454 | “Orthorectified Image” — Is Every Pixel Now in Exactly the Right Place?
A Primary 5/6 Reality Lab guide to orthorectified imagery: learn what terrain and sensor-geometry corrections improve, what residual registration error can remain, and how to evaluate map-image position claims.
-
PSLE Science Reality Lab Vol No.453 | “NoData = −9999” — Did the Sensor Measure Minus 9999?
A Primary 5/6 Reality Lab guide to NoData and fill values: learn why −9999 can mean missing or invalid data rather than a physical measurement, and how metadata, valid ranges and masks protect scientific reasoning.
