Normal probability plot of DEA efficiency scores with t

Analyzing the efficiency of small and medium-sized enterprises of a national technology innovation research and development program. Park S - Springerplus (2014) Bottom Line: However, the R&D collaboration type, "SME-University-Laboratory" Joint-Venture was superior to the others, achieving the largest median and the smallest interquartile range of DEA efficiency scores.Second, the differences in the efficiency were statistically significant between government R&D subsidy sizes, and the phenomenon of diseconomies of scale was identified on the whole.As the government R&D subsidy size increases, the central measures of DEA efficiency scores were reduced, but the dispersion measures rather tended to get larger. View Article: PubMed Central - PubMed Affiliation: Department of Business Administration, Baekseok University, Cheonan, ROK 330-704 Korea. ABSTRACT This study analyzes the efficiency of small and medium-sized enterprises (SMEs) of a national technology innovation research and development (R&D) program. In particular, an empirical analysis is presented that aims to answer the following question: "Is there a difference in the efficiency between R&D collaboration types and between government R&D subsidy sizes?" Methodologically, the efficiency of a government-sponsored R&D project (i.e., GSP) is measured by Data Envelopment Analysis (DEA), and a nonparametric analysis of variance method, the Kruskal-Wallis (KW) test is adopted to see if the efficiency differences between R&D collaboration types and between government R&D subsidy sizes are statistically significant. This study's major findings are as follows. First, contrary to our hypothesis, when we controlled the influence of government R&D subsidy size, there was no statistically significant difference in the efficiency between R&D collaboration types. However, the R&D collaboration type, "SME-University-Laboratory" Joint-Venture was superior to the others, achieving the largest median and the smallest interquartile range of DEA efficiency scores. Second, the differences in the efficiency were statistically significant between government R&D subsidy sizes, and the phenomenon of diseconomies of scale was identified on the whole. As the government R&D subsidy size increases, the central measures of DEA efficiency scores were reduced, but the dispersion measures rather tended to get larger. No MeSH data available.	© Copyright Policy - open-access Related In: Results - Collection License Show All Figures getmorefigures.php?uid=PMC4128955&req=5 Fig2: Normal probability plot of DEA efficiency scores with the 95% CI (n2 = 69). Mentions: Meanwhile, based on the Normal probability plot of DEA efficiency scores with the sample of n2 = 69 in Figure 2, it is not appropriate to assume the Normality. Additionally, two test statistics such as Anderson-Darling test statistic AD = 7.633* (P-value < 0.005) and Kolmogorov-Smirnov test statistic KS = 0.302* (P-value < 0.010) strongly support that their hypotheses, assuming the Normality should be rejected. Hereafter, , , ** indicate statistical significance at the significance level of α = 10%, 5%, 1%, respectively.Figure 2

Analyzing the efficiency of small and medium-sized enterprises of a national technology innovation research and development program.

Park S - Springerplus (2014)

Bottom Line: However, the R&D collaboration type, "SME-University-Laboratory" Joint-Venture was superior to the others, achieving the largest median and the smallest interquartile range of DEA efficiency scores.Second, the differences in the efficiency were statistically significant between government R&D subsidy sizes, and the phenomenon of diseconomies of scale was identified on the whole.As the government R&D subsidy size increases, the central measures of DEA efficiency scores were reduced, but the dispersion measures rather tended to get larger.

View Article: PubMed Central - PubMed

Affiliation: Department of Business Administration, Baekseok University, Cheonan, ROK 330-704 Korea.

ABSTRACT

This study analyzes the efficiency of small and medium-sized enterprises (SMEs) of a national technology innovation research and development (R&D) program. In particular, an empirical analysis is presented that aims to answer the following question: "Is there a difference in the efficiency between R&D collaboration types and between government R&D subsidy sizes?" Methodologically, the efficiency of a government-sponsored R&D project (i.e., GSP) is measured by Data Envelopment Analysis (DEA), and a nonparametric analysis of variance method, the Kruskal-Wallis (KW) test is adopted to see if the efficiency differences between R&D collaboration types and between government R&D subsidy sizes are statistically significant. This study's major findings are as follows. First, contrary to our hypothesis, when we controlled the influence of government R&D subsidy size, there was no statistically significant difference in the efficiency between R&D collaboration types. However, the R&D collaboration type, "SME-University-Laboratory" Joint-Venture was superior to the others, achieving the largest median and the smallest interquartile range of DEA efficiency scores. Second, the differences in the efficiency were statistically significant between government R&D subsidy sizes, and the phenomenon of diseconomies of scale was identified on the whole. As the government R&D subsidy size increases, the central measures of DEA efficiency scores were reduced, but the dispersion measures rather tended to get larger.

No MeSH data available.

Normal probability plot of DEA efficiency scores with the 95% CI (n2 = 69).

Related In: Results - Collection

License

Show All Figures

getmorefigures.php?uid=PMC4128955&req=5

Fig2: Normal probability plot of DEA efficiency scores with the 95% CI (n2 = 69).

Mentions: Meanwhile, based on the Normal probability plot of DEA efficiency scores with the sample of n2 = 69 in Figure 2, it is not appropriate to assume the Normality. Additionally, two test statistics such as Anderson-Darling test statistic AD = 7.633*** (P-value < 0.005) and Kolmogorov-Smirnov test statistic KS = 0.302*** (P-value < 0.010) strongly support that their hypotheses, assuming the Normality should be rejected. Hereafter, *, **, *** indicate statistical significance at the significance level of α = 10%, 5%, 1%, respectively.Figure 2