# 1-Way ANOVA: Interval Estimation - Quick Tab

Noncentrality Interval Estimation and the Evaluation of Statistical Models

Select the Quick tab of the 1-Way ANOVA: Interval Estimation dialog box to access options to implement the procedure (Fleishman, 1980; Steiger & Fouladi, 1997) for computing a confidence interval for RMSSE, the root mean square standardized effect size . Note that STATISTICA can also compute exact post hoc confidence intervals on power, and the sample size necessary to achieve a power goal, based on results from a study that has already been performed (see discussion below).

Observed F. In the Observed F box, enter the observed value of the F-statistic.

No. of Groups. In the No. of Groups box, enter the number of independent samples on which the analysis is based.

N per Group. In the N per Group box, enter the sample size per group.

Alpha. In the Alpha field, enter the type I error rate used as a target in calculating post hoc confidence intervals on power and required sample size.

Conf. Level. In the Conf. Level box, enter the confidence level for the confidence interval calculation. Note that confidence limits are given for the following quantities:

Delta. These confidence limits are for the noncentrality parameter of the noncentral F-distribution.

RMSSE. These confidence limits are for the RMSSE.

Power. These confidence limits are post hoc statistical bounds on power.  Because the F-statistic provides information that allows one to set confidence limits on delta, and because power and required sample size for any particular ANOVA design is a function of delta, it is therefore possible to set confidence limits on power, given the observed data.  This technique is a relatively recent idea, and, as such, is not discussed in the classic textbooks on ANOVA. It is also not discussed in Cohen's (1983) power analysis text. For an example of this calculation and its interpretation, see Steiger and Fouladi (1997, page 252).  For a general theoretical development of the method, see Taylor & Muller (1995).

Required N. Given an observed F-statistic, the program calculates a confidence interval on the sample size necessary to achieve a particular level of power, and reports the confidence limits here.

Power Goal. In the Power Goal box, enter the power goal used as a target in calculating post hoc confidence intervals on sample size.