Bias

Learn nonresponse bias, undercoverage.

Lesson 2 of 32 parts Practice Flashcards

1/2

Nonresponse Bias

Objective: Explain nonresponse bias, apply the stated method, and check its conditions.

Nonresponse can bias an estimate when responders differ relevantly from invited nonresponders. Response rate is responses divided by invitations, but it does not measure the size or direction of bias by itself. Investigate who is missing and how the design addresses them. A high response rate does not guarantee absence of nonresponse bias.

Formulas

Response fraction = responses / invitations

Worked examples

Try itA synthetic survey invites 70 people and receives 42 responses, with a<10. Find its response fraction.Show answer

Step 1: Response fraction is responses divided by invitations. Step 2: Use 42/70. Step 3: The fraction is 3/5; a higher fraction alone does not prove absence of bias.

Swipe sideways to move between sections

Nonresponse Bias

Objective: Explain nonresponse bias, apply the stated method, and check its conditions.

Formulas

Response fraction = responses / invitations

Worked examples

Try itA synthetic survey invites 70 people and receives 42 responses, with a<10. Find its response fraction.Show answer

Step 1: Response fraction is responses divided by invitations. Step 2: Use 42/70. Step 3: The fraction is 3/5; a higher fraction alone does not prove absence of bias.

Undercoverage

Objective: Explain undercoverage, apply the stated method, and check its conditions.

Undercoverage occurs when the sampling frame omits part of the target population. Increasing sample size inside the same incomplete frame does not restore the missing units. A clear population definition and frame comparison help diagnose this design problem. Larger samples cannot by themselves repair an incomplete sampling frame.

Formulas

Missing count = target population count − frame coverage count

Worked examples

Try itA target population has 700 people, but the sampling frame lists 560. How many target people are missing from the frame?Show answer

Step 1: Undercoverage means some target units are absent from the sampling frame. Step 2: Subtract 560 from 700. Step 3: 140 people are missing; this can bias inference if their responses differ.

You finished the reading!

Lock it in with a few practice questions, then move on to the next lesson.

Practice Next lesson

Previous lessonSampling Methods Next lessonObservational Studies Versus Experiments