Spotting bias and measurement error
Tell a slanted study (bias) from a noisy measurement (error), and random error from systematic error.
- Sample vs. population: Bias often comes from who ends up in the sample, so you must know a sample stands in for a larger group.
- Controlled variables: Recognizing what should be held constant helps you see when a study tilts the result.
Prerequisites are inferred: pending teacher review.
Re-learn the skill with worked practice and clear examples.
Bias is a tilt built into the study (often in who or what gets measured). Measurement error is how far a reading lands from the true value: random error scatters both ways, systematic error always pushes the same way.
A bathroom scale always reads exactly 2 kg higher than the true weight, every single time. What kind of problem is this?
Reviewed- A.Random error
- B.Systematic error
- C.A representative sample
- D.No error, because it is consistent
Show the worked solution ▾
Answer: B. Systematic error
- Step 1: Check the direction: Every reading is high by the same amount: it never wobbles low.
- Step 2: Match to the definition: An error that pushes every reading the same way is systematic error.
Why it's right: Because every reading is off in the same direction by the same amount, this is systematic error.
- A: Random error would scatter readings both high and low, but these are all high.
- C: A sample is about who is studied, not a scale reading.
- D: Being consistent does not make it correct; it is consistently wrong, which is systematic error.
Aligned to BI 2.1: random vs. systematic error · reading level ~grade 9
A company that sells a vitamin runs its own study and only counts customers who say they feel better, leaving out those who quit early. What is the main issue?
Reviewed- A.Random measurement error
- B.Bias, because the study is tilted toward a positive result
- C.A perfectly fair study
- D.Systematic error in a thermometer
Show the worked solution ▾
Answer: B. Bias, because the study is tilted toward a positive result
- Step 1: Look for a tilt: Dropping the people who quit and keeping only the satisfied ones slants the whole study before any reading is taken.
- Step 2: Name it: A study leaning toward a chosen result is biased.
Why it's right: Leaving out the people who quit tilts the study toward a positive result, which is bias.
- A: This is about who is counted, not the wobble in a single reading.
- C: Dropping unfavorable data is the opposite of fair.
- D: No thermometer or measuring tool is involved here.
Aligned to BI 2.1: bias · reading level ~grade 9
- A lab calibrates a scale against a known weight to catch systematic error before any experiment begins.
Fill these in as you work through the lesson.
- Bias (a study leans one way before measuring):
- Measurement error (how far a reading is off from the true value):
- Random error (unpredictable wobble that has no fixed direction):
- Systematic error (a built-in flaw that biases every reading consistently):
If something tilts who or what is studied, that is . If readings scatter high and low by chance, that is error; if every reading is off in the same direction, that is error.
- A survey is only handed out at a gym. Why might the results be biased?
- A scale always reads 2 kg too high. Is that random or systematic error, and why?
- Repeated stopwatch times that wobble a little above and below the true time show which kind of error?
A team weighs the same object 5 times on a scale and gets 50.1, 49.9, 50.2, 49.8, 50.0 kg, then learns the true mass is 52.0 kg. Describe the random error and the systematic error you see.