Testing the Difference Between the Respondents Demographic Groups

948 Words2 Pages

Testing the difference between the respondents demographic groups As mentioned in chapter 4: table 4-3, there was a very different between the number of respondents within some demographic groups, e.g. the gender groups compromised 185 males and 25 females and similarly, the academic background groups compromised 3 high school, 185 graduate and 85 postgraduate holders. Therefore, generalising the results of testing the difference between the means of these groups may be invalid and meaningfulness. However, it could give general indicators about the opinion of each group regarding the performing the SECI and innovation processes within the Egyptian banks. In this term, there are two methods to test the difference means between groups based on the measurement scale of the tested variables and the normality distribution of data. The parametric test (e.g. t-test and one–way ANOVA) is suitable for the ratio or the interval scale and for the data which is normally distributed. In contract, the non-parametric tests (e.g. Mann-Whitny and Kruskal-Wallis tests) are suitable for the nominal and ordinal scales and for the data which is normally or non-normally distributed (Field, 2009; Kleinbaum, et al. 2008). As mentioned earlier, the normality distribution of the research data had been investigated by skewness and kurtosis tests (see Table 5-1) and the study used the factor scores (which are not ordinal or nominal scales) yielded by the factor analysis for the further statistical analysis. Therefore, the parametric tests could be proper techniques to test the difference between the respondents’ demographic groups regarding the SECI and innovation activities in the Egyptian banks. In this term, there are two parametric statistical techniq... ... middle of paper ... ...the males and females staff in terms of performing theses processes in the Egyptian banks. In contract, the levene’s test for the externalisation and innovation processes shows that the F ratios were significant (p < .01and p <.10, respectively). Therefore, the Equal variances not assumed row will be used for the t-test. As shown in the table, the t-test results were not significant for the both processes therefore, it could be included that although there was difference between the means agreement of males and females staff in terms of performing the externalisation and innovation processes (the males means was a bit less than the females means for the externalisation process, Mean difference =.-171 and the males means was a bit higher than the females means for performing the innovation process, Mean difference= .065), however this difference is not significant.

Open Document