A researcher wants to collect data to create a confidence interval. Which change can the researcher make to decrease the width of the interval?

Increasing the level of confidence widens the confidence interval. For example, moving from a 95% to a 99% confidence level requires a broader interval to ensure that the true parameter is captured within that range. This is contrary to the goal of decreasing the interval's width.

Collecting another sample does not guarantee a decrease in the width of the confidence interval unless the new sample size is combined with the existing data. If the additional sample is not larger than the original, it will not reduce the width. Thus, without increasing the overall sample size, this choice is ineffective.

Using a different estimator may not necessarily affect the width of the confidence interval. Different estimators can have varying properties, but simply changing the estimator doesn't ensure a narrower interval. The width primarily depends on the sample size and the level of confidence, not the estimator itself.