T-Test for Proportions? Making Do When Your Software Can't Do Confidence Intervals for Proportions
Presentation Type
Poster/Portfolio
Presenter Major(s)
Biostatistics, Statistics
Mentor Information
Sango Otieno, otienos@gvsu.edu; Gerald Shoultz, shoultzg@gvsu.edu
Department
Statistics
Location
Kirkhof Center KC23
Start Date
13-4-2011 12:00 PM
End Date
13-4-2011 1:00 PM
Keywords
Mathematical Science
Abstract
In Introductory Statistics courses, students are taught how to calculate confidence intervals for the population proportion and the difference between two population proportions. However, statistical software packages often lack syntax for computing such intervals. We assume that if proportion data is recoded in a binary fashion, the resulting t-confidence intervals for one and two sample problems are equivalent to the corresponding z-intervals. We determine mathematically and by simulations, at what sample size, the t and z intervals can be considered equivalent for various confidence levels.
T-Test for Proportions? Making Do When Your Software Can't Do Confidence Intervals for Proportions
Kirkhof Center KC23
In Introductory Statistics courses, students are taught how to calculate confidence intervals for the population proportion and the difference between two population proportions. However, statistical software packages often lack syntax for computing such intervals. We assume that if proportion data is recoded in a binary fashion, the resulting t-confidence intervals for one and two sample problems are equivalent to the corresponding z-intervals. We determine mathematically and by simulations, at what sample size, the t and z intervals can be considered equivalent for various confidence levels.