, pp.61-66 http://dx.doi.org/10.14257/astl.2014.48.12 An Empirical Study on the Effects of Software Characteristics on Corporate Moon-Jong Choi 1, Won-Seok Kang 1 and Geun-A Kim 2 1 DGIST, 333 Techno Jungang Daero, Hyeonpung-Myeon, Dalseong-Gun, Daegu, 711-873, South Korea {mj, wskang}@dgist.ac.kr 2 Kyungpook National University, 80 Daehakro, Bukgu, Daegu, 702-701, South Korea twinkle2204@hanmail.net Abstract. This research suggests innovativeness, standardization, and flexibility as three important software characteristics, and aims to investigate the relationship between these three characteristics while observing their influence on performance. The results reveal that software innovation and standardization positively influence flexibility, and that the three software characteristics variables, innovation, standardization, and flexibility, have a positive effect on performance. This study not only focuses on software characteristics, but also provides specific information that hands-on software development staff can use to strengthen the competitiveness of the industry. Keywords: SW innovativeness, SW standardization, SW flexibility, 1 Introduction With the emergence of mobile software along with various other fields, the domestic software industry faces new opportunities and threats due to the increased competitiveness. In this changing environment, increased competitiveness through technological innovation is very important. Thus, the small and medium scale domestic software companies need to grasp the software characteristics that can enhance a company s performance. Therefore, this research suggests innovativeness, standardization, and flexibility as the three important software characteristics, and aims to find the relationship between these three characteristics as well as investigate their influence on performance. Moreover, as there is a dearth of empirical studies on the current state of the software industry, this research will not only be a useful addition to the extant research, but will also inform hands-on staff of the factors necessary to achieve competitive advantage. Thus, this study will make a significant contribution to the creation of a successful software industry. ISSN: 2287-1233 ASTL Copyright 2014 SERSC
2 Hypothesis Development Figure 1 elucidates the research model proposed in the study. SW Innovativeness H4 H1 SW Flexibility H3 H2 SW Standardization H5 Fig. 1. The research model and hypotheses. Hypothesis 1. SW innovation has a positive influence on SW flexibility. Hypothesis 2. SW standardization has a positive influence on SW flexibility. Hypothesis 3. SW flexibility has a positive influence on performance. Hypothesis 4. SW innovation has a positive influence on performance. Hypothesis 5. SW standardization has a positive influence on performance. 3 Research Methods 3.1 Measures and Data Collection The study considers domestic software companies that have been developing new software as subjects and uses the survey method. The questionnaire used to measure the variables of the research model was based upon the 7-point Likert scale, ranging from (1) for strongly disagree to (7) for strongly agree. Measurement items developed for inspection of the suggested research model were initially tested for validity, followed by confirmation through theoretical concepts, and then revised to suit the research. In addition, survey questions were revised based on the responses obtained from the interviews of affiliates who implement software development in companies. Then, the final measurement items to be used in this research were developed. For the refinement and screening of these measurement items, a pilot test was conducted, focusing on companies that are currently developing software or are planning to develop software. The pre-investigation results showed that there were no factors that hindered the measurement items reliability and validity. A total of 1000 final questionnaires were distributed using multi-dimensional means such as email, mail, direct visit, and web survey for the duration of one month from April to May 2013, and a total of 220 surveys were collected. However, only 211 surveys were used in 62 Copyright 2014 SERSC
the analysis after discarding nine due to insincere responses. Table 1 shows the characteristics of the respondent companies and individuals. Table 1. Respondents. Demographic Characteristics Frequency (n=211) Percentage (100%) Classification of business (multiple responses) System software 82 21.7% Development software 39 10.3% Applications software 76 20.1% Services related to computers 48 12.7% Digital content development service 53 14.0% Embedded software development service 80 21.2% Position Associate level 4 1.9% Deputy section chief level 13 6.2% Section chief level 23 10.9% Deputy head of department level 26 12.3% Department head level 37 17.5% Executive level 58 27.5% CEO 50 23.7% Type of work Research service 83 39.3% Technological service 42 19.9% Administrative service 76 36.0% Other 10 4.7% Gender Male 200 94.8% Female 11 5.2% 4 Data Analysis and Results 4.1 Assessment of the Measurement Model In order to inspect the degree to which the data s characteristics agree with the evaluation model, before the examination of the reliability and validity of the evaluation model, AMOS 19.0 was used to inspect the suitability. The variables χ2, GFI, AGFI, RMSEA, NFI, CFI, and IFI were observed to analyze the overall suitability of the model. A total of 17 items were used to measure four latent variables in the examination of the initial measurement model. The results of the initial measurement model s suitability showed that the GFI and RMSEA index were below the promotion value suggested in existing research. Modification indices observation results showed that one item hindered the software innovativeness. The suitability test was conducted again after the removal of that item. Confirmatory factor analysis of Copyright 2014 SERSC 63
latent variables resulted in: IFI=0.948, GFI=0.901, AGFI=0.832, CFI=0.944, χ2/df=1.881, and RMSEA=0.054; the recommended level of the overall index of the measurement model was satisfactory. After the suitability test, the final data (n=211) gathered before the structural model test was used to test the measurement tool s reliability and validity. For the reliability, the most commonly used Cronbach s Alpha coefficient (threshold 0.7 and above) was employed [3]. Validity can be divided into convergent validity test and discriminant validity test. For testing convergent validity, factor loading, composite reliability, and Average Variance Extracted (AVE) results of Confirmatory Factor Analysis (CFA) were used that employed AMOS 19.0. Generally, if the factor load was greater than ± 0.4, it was considered significant [1], and the composition reliability index had to be greater than 0.7 and each latent variable s average value had to be greater than 0.5 to say that convergent validity exists [2]. Lastly, the discriminant validity test uses the Average Variance Extracted (AVE) suggested by Fornell and Larcker [2] and the Pearson correlation analysis method. To test whether the discriminant validity exists, each latent variable s AVE square root value has to exceed the Correlation coefficient between the latent variable at hand and other latent variables. The test of the measurement model provides evidence for the internal consistency and validity statistically. Table 2 and Table 3 show the results of the reliability and validity tests of the measurement model. Table 2. Reliability and convergent validity. Construct Item t-value Loading SI1-0.921 SW innovativeness SI2 15.795 0.813 SI4 16.486 0.832 SS1-0.893 SS2 16.575 0.845 SW standardization SS3 15.984 0.828 SS5 16.872 0.852 SF1-0.818 SW flexibility SF3 14.295 0.861 SF4 14.309 0.862 P1-0.771 P2 13.759 0.896 P3 11.892 0.786 P4 12.666 0.829 Cronbach s Alpha CR AVE 0.886 0.892 0.734 0.936 0.916 0.731 0.883 0.884 0.717 0.888 0.892 0.676 Table 3. Results for discriminant validity. Latent Variable 1 2 3 4 1. SW innovativeness 0.857 2. SW standardization 0.542 0.885 3. SW flexibility 0.626 0.621 0.847 4. 0.509 0.454 0.516 0.822 64 Copyright 2014 SERSC
Note: Items in bold type along the diagonal represent the square root of the AVE. For discriminant validity, diagonal values should exceed off-diagonal correlations. 4.2 Assessment of the Structural Model After the validity test of the measurement model, the Structural Equation Modeling (SEM) approach was used to find the relationship between the variables suggested in this research model using the collected data. The analysis of structural models not only helps to observe the suitability of structural models and relationships between the research model variables, but also the coefficient determination of the R-squared (R 2 ) of the endogenous variables. First, the results of the structural model suitability test used the index used in the measurement model suitability test, and the results were: IFI = 0.936, GFI = 0.903, AGFI = 0.823, CFI = 0.928, the relative chi-square (X2/df)=2.029, and RMSEA=0.05. Thus, there was no notable difficulty in testing the research hypothesis. Second, structure equation analysis was used to obtain the path coefficient (β). This shows the relationship between two variables [4]. Third, the structural model analysis was used to determine the coefficient of determination R 2 of the endogenous variable. The coefficient of determination R 2 refers to the ratio explained by the regression line and exogenous variable of the changes in the research model. The results of the structural equation analysis are shown in Figure 2. SW Innovativeness 0.569*** (3.734) 0.491*** (6.939) SW Flexibility 0.436** (2.436) SW Standardization 0.388*** (7.016) R 2 =0.567 R 2 =0.391 0.193* (1.672) * : p<0.1, ** : p<0.05, *** : p<0.01 Fig. 2. The results of the structural equation model (SEM). 5 Conclusions By studying the factors that influence the performance of software companies, this research aims to provide various insights not only to other researchers in the field, but also to officials in charge of implementing software industry policies and technological development within software companies. Most research on software has been focused on direct technical fields within software; however, this research Copyright 2014 SERSC 65
focuses on software characteristics of innovation, standardization, and flexibility, and analyses their effects and influence on company performance. Despite the significance of the characteristics for software development as reflected in multidimensional analysis, research on software companies has been lacking. This research will be able to provide a foundation for researchers aiming to further the research in this field in the future. In addition, the actual increase in innovativeness, standardization, and flexibility that influences software characteristics, leading to company s success, is due to the distinctiveness of the software industry. As the life cycle of techniques in the software industry is short, an endless technical innovation is necessary. Further, rather than tangible assets, such as manufacturing equipment on sites, intangible assets, such as systematic process management techniques and standardized management manuals for new and innovative product development, are used to improve software companies competiveness Acknowledgments. This work was supported by the DGIST R&D Program of the Ministry of Education, Science and Technology of Korea. (14-BD-01) References 1. Barclay, D., Thompson, R., Higgins, C.: The Partial Least Squares Approach to Causal Modeling: Personal Computer Adoption and Use as an Illustration. Tech. Stud. 2(2), 285-324 (1995) 2. Fornell, C., Larcker, D.F.: Evaluating Structural Equation Models with Unobservable Variables and Measurement Error. J. Market. Res. 18(1). 39-50 (1981) 3. Nunnally, J.C.: Psychometric Theory, New York: McGraw Hill (1978) 4. Wixom, B.H., Watson, H.J.: An Empirical Investigation of the Factors Affecting Data Warehousing Success. MIS Q. 25(1). 17-41 (2001) 66 Copyright 2014 SERSC