It is widely believed (Beliaet al., 2005; Schenker and Gentleman, 2001) that two 95
percent CIs having zero overlap – just touching end to end – are equivalent to statistical
significance withp¼0:05. In fact, for 95 percent CIs on two independent means, overlap
by about one quarter of the total length of one interval corresponds to apvalue of
approximately 0.05 (Cumming and Finch, 2005; Wolfe and Haney, 2002).
As displayed in the right side of Figure 1, two CIs touching end-to-end correspond to
an approximatepvalue of 0.01. Using the chart located in the right hand frame, one can
easily obtain a visual to determine the statistical difference between two point estimates.
For example, aligning the top of the chart, which equals 1.0, with the mean in group one
and tracing the chart downward until aligning with the second group mean provides a
useful estimation of statistical significance utilizing confidence intervals.