An Analysis of Variance Test for Normality (Complete Samples)
Citations
7,134 citations
Cites methods from "An Analysis of Variance Test for No..."
...For all months, an asymmetric distribution was found which required data transformation, since pvalues were close to zero by normality test (SHAPIRO and WILK, 1965)....
[...]
...Normality hypothesis was tested according to the W test at 5% (SHAPIRO and WILK, 1965)....
[...]
3,774 citations
2,890 citations
Cites background or methods or result from "An Analysis of Variance Test for No..."
...The following values of men's weights in pounds, first given by Snedecor, were used by Shapiro and Wilk [17] as an illustration of a test for normality: 148, 154, 158, 160, 161, 162, 166, 170, 182, 195, 236....
[...]
...W is a statistic introduced by Shapiro and Wilk [17], and DA, W' are subsequent extensions introduced by d'Agostino [2, 3] (there called D), and Shapiro and Francia [16]....
[...]
...Shapiro and Wilk [17] also included power studies for D, W2 and A2 (there called D, CVM, WCVM) and found...
[...]
...For this last case, EDF statistics have suffered recently from comparison with the W-statistic of Shapiro and Wilk [17]; no doubt, this is because the power studies reported in that paper gave very low power to EDF statistics....
[...]
...Power studies in support of new statistics have often been skimpy; since, after all, we don't usually know the exact alternative to the null distribution, it is urged that a range of alternatives comparable to those used by Shapiro and Wilk [17] or in Table 6 of this article should always be investigated....
[...]
2,796 citations
Cites background or methods from "An Analysis of Variance Test for No..."
.../#4 - (2Vir)-1]NU/0-02998598 is outside (D*, Du*), where e? is the ith order statistic of u1,..., UN; (iv) Pearson et al. (1977) R test: reject Ho if either Vbl is outside (R1L, R1~) or b2 is outside (R2L, R2U); (v) Shapiro & Wilk (1965) W test: reject Ho if W = (E a1Ne?)...
[...]
...…and omnibus tests proposed by D'Agostino & Pearson (1973), Bowman & Shenton (1975) and Pearson, D'Agostino & Bowman (1977), the analysis of variance tests of Shapiro & Wilk (1965) and Shapiro & Francia (1972), and the coordinate-dependent and invariant procedures described by Cox & Small (1978)....
[...]
2,540 citations
Cites background from "An Analysis of Variance Test for No..."
...Skewness and kurtosis statistics of the dependent variable fall well within the boundaries for normality ( Shapiro and Wilk, 1965 ), allowing parametric tests of significance....
[...]
References
2,085 citations
2,013 citations
1,035 citations