Data and Evidence
8th Grade Science · 10 lessons
Start lesson 1
Lessons
- 1Claim, Evidence, Reasoning — and the RebuttalA complete scientific argument has four parts, not three: the claim, the evidence for it, the reasoning that links the two, and a rebuttal that names the strongest rival explanation and says why it fails. Leaving out the rebuttal is what separates an argument from an assertion, because an explanation that has never been set against a serious alternative has not actually been tested — it has only been stated attractively.Lesson
- 2Independent, Dependent and Controlled VariablesA fair test changes one independent variable, measures the dependent variable, and holds every other controlled variable fixed. If two things change at once, the result cannot be attributed to either, no matter how large the effect looks. Listing the controlled variables out loud before starting is the step that catches the ones which would otherwise drift by accident — room temperature, the age of the sample, the person taking the reading.Experiment
- 3Random Error Scatters, Systematic Error ShiftsRandom error scatters readings either side of the true value, so averaging more repeats narrows the spread and improves precision. Systematic error shifts every reading the same way, so the mean of a thousand readings from a balance reading 2 g high is still 2 g wrong — repeating changes nothing at all. The two faults are fixed by different actions: repeats for the first, calibration against a known standard for the second, which is why precision and accuracy are reported separately.Experiment
- 4Percentage Uncertainty Rewards Measuring Something BiggerPercentage uncertainty is the absolute uncertainty divided by the measurement, times 100. A ruler good to ±1 mm gives 20% uncertainty on a 5 mm object and only 0.2% on a 500 mm one, which is why the route to a precise result is so often to measure something larger and divide at the end — twenty sheets of paper at once rather than one, ten oscillations of a pendulum rather than a single swing.Lesson
- 5Anomalies Are Investigated, Not DeletedA reading far from its neighbours is a signal that something happened, not a licence to remove a point. The correct response is to repeat that measurement: if the value recurs it is data and the explanation lies in the system, and if it does not, you record that you excluded it and state why. Quietly dropping inconvenient points is exactly how a mistaken result comes to look beautifully consistent, and it is undetectable from the final graph alone.Experiment
- 6A Small Sample and a Biased Sample Fail DifferentlyA small sample is dominated by chance, and it announces that honestly through a wide uncertainty — collect more and the problem shrinks. A biased sample is wrong in a fixed direction, because of who volunteered, which site was easiest to reach, or which plants looked healthiest, and enlarging it only makes it more confidently wrong. Only the first fault is cured by collecting more data, which is why how a sample was chosen matters more than how big it is.Experiment
🖨️
View in ShopGet all 10 Data and Evidence lessons as PDF
Printable worksheets & answer keys — $4.99