7. 1 The present study is intended to investigate about Library Effectiveness by
maintaining
Library
Automation
in
the
University
Libraries
of
Chhattisgarh. Using appropriate statistical tables, common reference of Fratio was applied to test the significance of differences for verification of various hypotheses. Scores for all the scales collected from all the samples were analysed separately. The outcome of statistical analysis is compiled in tables and figures followed by interpretation.
The major prerequisites for effective data analysis using ANOVA technique are the normally distributed and homogenous data. For judging normality of any given data, the best ways prescribed are to calculate the skew ness and kurtosis of the distribution pertaining to the data.
A distribution is said to be skewed when the mean and median fall at different points in the distribution, and the balance of the centre of gravity is shifted to one side or the other – i.e. to the left or the right. Ideally the skew ness value should therefore be exactly “0” (zero) for mean and median to fall at the same point. The skew ness of the dependent variable Library Effectiveness in this case is -0.233 with a standard error of 0.121. Therefore, the data was found to insignificantly skew.
The term kurtosis refers to the flatness of a frequency distribution compared with the normal distribution. While according to Garrett (1989) the standard Kurtosis value of normally distributed data should be exactly
DATA ANALYSIS & INTERPRETATION
7. 2 0.263, the kurtosis value of dependent variable Library Effectiveness in this study was found to be -0.003 with a standard error of 0.241. Therefore, it was inferred that the data being insignificantly Kurt was almost normally distributed. Further, using the Levene‟s test for measuring homogeneity of the data, significant results (p=.000) were obtained which denied the homogeneity of data. Therefore, following McNemar (1962), as the two assumption were not absolutely fulfilled, but ANOVA being a robust technique that overlooks near miss cases, for further analysis of results, the level of significance as an accept/reject criteria was kept stringent at 0.01 level i.e. (p
