Normality Test. When the underlying variable is normal, the traditional performance evaluation measure will not miss anything by only considering mean and variance. However, one issue in performance evaluation is that the returns of mutual funds are non-normal. Table 3 shows the Kolmogorov-Smirnov and Xxxxxx-Xxxx test results. For any group of mutual funds, the normality hypothesis is strongly rejected. Previous literature has also documented non-normalities in mutual fund returns. Kosowski, Timmermann, Xxxxxxx, and White (2006) suggest these non-normalities arise for three reasons. First, individual stocks within a typical mutual fund portfolio realize returns with non-negligible higher moments and managers often hold heavy positions in relatively few stocks or industries. Second, individual stocks exhibit varying levels of time- series autocorrelations in returns. Third, funds may implement dynamic strategies that involve changing their levels of risk-taking when the risk of the overall market portfolio changes. Kosowski, Timmermann, Xxxxxxx, and White (2006) argue that normality may be a poor approximation in practice, even for a fairly large mutual fund portfolio. The stochastic dominance test is based on the entire distribution. Unlike the Xxxxxx ratio, it does not require the return to be normally distributed.
Normality Test. The normality test was conducted on the data of pretest and posttest between control class and experimental class. If the p value is higher than 0.05 it means that the data is normal. Table 10: the result of normality test using one sample Kolmogorov-smirnov test One-Sample Kolmogorov-Smirnov Test pretest control class posttest control class pretest experimenta l class posttest eperimental class Normal Parametersa,b Mean 57,412 59,529 55,971 67,059 Std. Deviati on 9,4244 8,7776 11,0588 7,7477 Most Extreme Differences Absolut e ,188 ,122 ,094 ,124 Positive ,135 ,114 ,094 ,117 Negativ e -,188 -,122 -,088 -,124 Test Statistic ,188 ,122 ,094 ,124 Asymp. Sig. (2-tailed) ,111c ,200c,d ,200c,d ,200c,d a. Test distribution is Normal.
Normality Test. One-Sample Kolmogorov-Smirnov Test pretest control class posttest control class pretest experimenta l class posttest eperiment al class Normal Parametersa,b Mean 57,412 59,529 55,971 67,059 Std. Deviation 9,4244 8,7776 11,0588 7,7477 Most Extreme Differences Absolute ,188 ,122 ,094 ,124 Positive ,135 ,114 ,094 ,117 Negative -,188 -,122 -,088 -,124 Test Statistic ,188 ,122 ,094 ,124 Asymp. Sig. (2-tailed) ,111c ,200c,d ,200c,d ,200c,d a. Test distribution is Normal.
Normality Test. Normality test is used to test whether a variable is normal or not. Normal here means if the data have a normal distribution. The main reason of conducting normality testing in a research is that it is necessary for the researcher to know that the population or data involved in the research is in normal distribution. To test the normality of the data can use the One Sample Kolmogorov-Smirnov test with the provision that if Asymp. Sig > 0,05 the data were normally distributed (Asmarani, 2008). In this case the normality using SPSS (Statistical Product and Service Solutions) 23
