Interpreting output for a t test for two related samples You are a marketing expert fora company that produces a new kind of energy bar. You want to provide evidence that consuming the new bar 1 hour before engaging in strenuous exercise of a long duration- such as running a marathon-will have an effect on performance. You decide to conduct a test at an annual half-marathon race. Because the course and conditions for this year and last year are similar, you decide to provide the bar to 40 randomly selected participants who also participated last year and compare their times between this year and last. The average time of your sample this year when they consumed the bar (TWITHBAR) is 2 hours and 24 minutes (2:24, or 144 minutes).
The average time of your sample last year when they did not consume the bar TWITHOUTBAR) is 2 hours and 28 minutes (2:28, or 148 minutes). You use a statistical computing package to conduct a t test for related samples. The following tables list the output.
Paired-Samples Statistics Std Deviation 27.98 22.45 Std Error Pair TWITHOUTBAR 148 40 4.42 1 TWITHBAR 144 40 3.55 Paired-Samples Correlations Correlation Sig Pair 1 TWITHOUTBAR &TWITHBAR 40 0.56 Paired-Samples Test Paired Differences 95% Confidence Interval of the Std Std Error Mean 3.82 Sig Mean Lower Uppert 3.72 11.721.05390.301 Upper d(2-tailed) Pair 1 TWITHOUTBAR TWITHBAR 4.00 24.15