📊

Confidence Intervals and Predictions

Jul 12, 2025

Overview

This lecture explains when a confidence interval allows for making predictions about an outcome, using a political ballot example where the passing threshold is 50%.

Making Predictions with Confidence Intervals

  • You can make a prediction when the entire confidence interval fits within your specified restriction.
  • For example, if predicting a ballot will pass requires more than 50% votes, the confidence interval for support must be entirely above 50%.

Political Ballot Example

  • A poll of 1,000 likely voters found 515 support a proposition.
  • Using 1-PropZInt, the resulting confidence interval for support is approximately 48% to 54%.
  • This interval includes possible support levels both below and above 50%.

When Prediction is Not Possible

  • If any part of the confidence interval falls outside the required restriction (e.g., less than 50%), you cannot make a definite prediction.
  • In the example, since the interval includes values below 50%, you can't be certain the proposition will pass.
  • Both scenarios exist: some possible support levels predict passing, others predict failing.

Visual Representation

  • A visual (blue/orange/red) illustrates when the interval is fully within or partially outside the restriction.
  • Prediction is only possible when the entire interval is within the restriction.

Key Terms & Definitions

  • Confidence Interval — a range of values that likely contains the true population parameter.
  • Restriction — the condition that must be met for a positive outcome (e.g., over 50% votes).
  • 1-PropZInt — calculator/statistical function to compute a confidence interval for one proportion.

Action Items / Next Steps

  • Review class notes and visual aids on confidence intervals and prediction criteria.
  • Practice with sample problems identifying when predictions can and cannot be made from confidence intervals.