Adoption of information and communication technology

3 downloads 19 Views 880KB Size Report
Type III(a) is an IS product and business technological process innovation .... A youthful CIO could possess greater ICT knowledge and it may impact technology.

The current issue and full text archlve of t h ~ journal s www.emeraldinsight.com/0263-5577.htm

IS available at

Adoption of information and communication technology Impact of technology types, organization resources and management style

Information and communication technology 1257

Kyung Hoon Yang Department of I~zfo~*mation Systems, Wing Technology Center, University of Wisconsin-LaCrosse, La Crosse, Wisco~zsin,USA

Sang M. Lee of Management, College of Business Administration, -Department Unive~~sitv of Nebraska-Lincoln,Lincoln, Nebraska, USA, and

Sang-Gun Lee Department of Management, College of Business Administration, Ajou University, Suwon, Soutlz Korea Abstract Purpose - The aim of this paper is to find out why some organizations adopt ICT later than the others, and whether organizations have different adoption strategies based on the type of ICT. Design/methodology/approach- This paper is an empirical study of the ICT diffusion process between the early and late adopters of relational database and local area network. Findings - The results of the study indicate that there are significant differences between early and late adoption organizations with regard to management characteristics such as the age and interests of the CEO and CIO, and also the adoption process such as the eviluation period and initiation time. However, no significant differences were found in organization resource or corporate strategy factors. Furthermore, the results of the study indi~ltethat there are significant differences in organizational characteristics such as sales volume, organization slack and rewards between the two types of ICT. Research limitations/implications - The sample size is relatively small. Replication of this study with additional organizations in the sample will allow stronger validation of diffusion theory. Originalitylvalue - It is believed that the results of the study will provide useful guidelines in strategy development for managing ICT diffusion by organizations and IT vendors. Keywords Innovation, Communication technology Paper type Research paper

1. Introduction Recently, ICT has become a strategic asset which can help improve business processes and change the function of markets. Thus, it is necessary for organizations to continue their efforts in developing and implementing the up-to-date technology. Nevertheless, d even believe IT does not many organizations still hesitate to adopt new 1 c ~ - a n some matter as a strategic resource because of its commoditization (Cam, 2003). The purpose of this study is to answer the following two questions: (1) Why do some organizations adopt ICT later than others? (2) Do organizations have different adoption strategies according to the type of ICT?

lndustnal Management & Data Systems Vol 107 No 9, XQi pp 1257 1275 8 Emerald Group Publ~shmgLlrn~ted 0263 55ii DO] 10 UOs/026!3557071OKB956

IMDS 107,9

1258

To answer these research questions, we did the following first we classified several factors that affect ICT adoption into three categories: corporate strategies, organization resources and management style. Then we investigdted whether there were significant differences in these categories between early adoption organizations (EAOs) and late adoption organizations (LAOS). Second, we examined whether organizations, which have been using a certain type of ICT, influenced other organizations that were planning to adopt it. Third, we tried to find whether there are differences in the ICT adoption initiation time and evaluation periods between the two groups. Fourth, we tried to find the difference in ICT performance between EAOs and LAOs. We wanted to ascertain whether LAOs delayed the adoption of ICT on purpose or they delayed it due to organizational limitations. We believe the results of the study will provide useful guidelines in strategy development for managing ICT diffusion by organizations and IT vendors. 2. Theoretical background 2.1 Organizational factors for ICT adoption Since, Ryan and Gross (1943) discussed the diffusion of technology innovation, this has been a fascinating research topic and is still one of the hot issues for many researchers in several fields. In the MIS field, Rogers (1995) raised this issue and many researchers suggested several theories, frameworks, and research methods. As research advanced, the research questions have focused on the pattern and extent of ICT diffusion and the propensity of an adopter to adopt and assimilate ICT (Fichman, 2000). According to Gallivan (2001), the following theories were advanced and popularly used: the diffusion of innovations (Rogers, 1995), theory of reasoned action (Ajzen and Fishbein, 1980), the technology acceptance model (Davis, 1989), theory of planned behavior (Taylor and Todd, 1955), and social cognitive theory (Compeau and Higgins, 1995). Based on the above-discussed research, Swanson (1994) classified ICT into three types. Type I innovation is an IS process innovation. There are two sub-types of Type I: (1) Type I(a) is an IS administrative process innovation such as IS outsourcing. (2) Type I(b) is an IS technological process innovation such as computer-assisted systems engineering (CASE) tools, database management systems (DBMS), object-oriented processing systems (OOPS).

Type 11 innovation is an IS product and business administrative process innovation such as the information center, executive information systems (EIS), payroll, personnel record systems, accounting information systems, teleconferencing, expert systems, e-mail, etc. Types 111 innovation involves inter-organizational IS innovation and has three sub-types. Type III(a) is an IS product and business technological process innovation such as material requirements planning (MRP), and computer integrated manufacturing, Type I II(b) is an IS product and business product innovation such as airline reservation systems. Type III(c) includes IS product and business integration innovation such as electronic data interchange (EDI). Swanson's framework was verified by Grover et aE. (1997). Table I collectively shows a summary of important studies applicable to this study.

w

2.2 Dflusion of innovation theory The diffusion of innovation theory has been applied to understanding the innovation behavior of adopters. In this paper, we are interested in knowing how the adopters affect the potential adopters. The pattern of the cumulative frequency of innovation adopters over time forms an S-shaped curve. This curve describes the behaviors of adopters, hnown as the d f i s i o n model. While there are many variant types of diffusion models, their basic forms can be classified into the external influence model, the internal influence model, and the mixed influence model (Mahajan et al, 1988; Venkatraman et al., 1994).

Information and communication technology 1259

-

Innovation type

ICT

Researchers

Type Ia.innovation

Outsourcing

Type Ib innovation

IS services Work station CASE Tool

Lob and Venkatraman (1992) and Grover et d.(1997) L i d and Zmud (1991) Moore and Ber~basat(1991) Orlikowski (1993) and Grover et al (1997) Nilakanta and Scamell (1990) and Grover et al (1997) Swanson (1994) Grover et al. (1997) Brancheau and Wetherbe (1990) Fichman and Kemerer (1993) Moch and Morse (1977) Zmud (1984) Fichman and Kemerer (1993) Grover et d (1997) Grover et al (1997) Grover et d (1997) Grover et al (1997) Swanson (1994) Johannes and Harianto (1992) George et al. (1992) Moore and Benbasat (1991) and Wildemuth (1992) Attewell (1992) and Garcia-Morales et d (2006) Burkhardt and Brass (1990) Grover and Gosler (1993) Swanson (1994) Grover et d.(1997) Chau and Tam (1997) and Huang and Lin (2006) Chen et aL (2002) Grover et d (1997) and Godlez-Alvarez and Nieto-Antolin (2005) Lai and Guynes (1994) Gurbaxani (1990)

DBMS and DDBMS

Type 11innovation

Data administration OOPS Spreadsheet 4GLs EDP Modern software practices Software engineering technology Expert systems E-mail EIS Teleconferencing Information center ATM GDSS End-user computing Business computing

Type W innovation

General purpose individual computing Telecommunication tech MRP CADICAM Open system

Type IIIc innovation

On-line trade ED1

Type IDa innovation

ISDN BITNET

\

Table I. ICT type and relevant studies

IhfDS 107,9

2.1.1 Internal influence model. The internal influence model, also known as the pure imitation difference model, posits that diffusion occurs only through individual contacts. This model is the most useful when early adopters' experience and information are needed for adoption. This model can be expressed mathematically as follows:

where N(t) is the cumulative number of adopters, q is an imitation coefficient or internal influence coefficient, m is the number of potential adopters, and interaction between the adoption units is expressed as N ( t ) [ m- N(t)].This model presumes that the diffusion rate is a function of social interaction or internal communication between early adopters and potential adopters in a social system. 2.1.2 External influence model. In this model, the diffusion process is driven only through an infomation or communication source external to -the social system (Venkatrarnan et al., 1994). This model assumes that the diffusion rate at time t depends only on the number of potential adopters in the social system at time t. The external model can be expressed mathematically as follows:

where N(t),m, p, and dN(t)/d(t)represent the cumulative number of adopters at time t, the number of potential adopters of innovation which is a non-negative constant, the innovation or external influence coefficient, and the diffusion rate at time t, respectively. 2.1.3 Mixed influence model. The mixed influence model is a combination of the internal and external influence models. This model is widely used as it more precisely explains reality than the other two models. The following is the mathematical formula of this model: -a(t) - p[m - NO)]

dt

+ q[m - N ( t ) ] N ( t )

(3)

where p and q are defined as the non-negative innovation coefficient and the imitation coefficient, respectively. For a successful innovation, the contition p S q should be satisfied.

3. Research model and hypotheses 3.1 Resea~chmodel This paper is an extension of previous studies. First, we adopted Swanson's (1994) framework. According to his classification, we chose relational database (RDB) to represent Type.1 ICT innovation and local area network (LAN) to represent Type I1 ICT innovation. Type III innovation is an inter-organizationalICT and is beyond the scope of our research. Second, we also analyzed organizational factors of the ICT diffusion process. We wdnted to h d out whether late adopters delayed the adoption of ICT strategically or they were forced to delay due to organizational limitations,

whether or not the evaluation period would be about the same between early and late adopters, whether there was a difference in the information source that the early and late adopters used, and if there was a difference in ICT performance between the early and the late adopters. To examine the above issues, this study undertook the following We selected the factors that the previous researchers found relevant. These factors were categorized into organization resources, management style and corporate strategies. Appropriate research instruments were chosen based on previous studies. We chose ICT using Swanson's (1994)framework of ICT adoption. We explored whether organizations are affected by imitation of ICT adoption and if so to what degree. Finally, we analyzed the performance of the sample organizations.

Infomation and cOmmunication technology

1261

The research procedure is shown in Table II. 3.2 Hypotlzeses

As one of the organizational characteristics, organization size may have an effect on ICT adoption (Zmud, 1984; Nilakanta and Scamell, 1990). Large organizations are likely to require a higher level .of information and technology. Accordingly, in general, EAOs would be larger than LAOs with regard to organization size: H1-1(a). In regard to RDB adoption, EAOs have a larger sales volume than LAOs. 1 - 1 ) In regard to LAN adoption, EAOs have a larger sales volume than LAOs. H1-2(a). In regard to RDB adoption, EAOs have a higher ratio of IT department employees/total employees than LAOs. ICT type Category

Factor

Instrument

Source

Type1

Human resources

IT employ/employ

Zmud (1984)

Organization resources

Financial resources Sales volume Reward IT resources Slack Reward Management style Decision structure Decision process Champion Corporate strategy Objectiveness Creativeness

Rogers (1995) Bourgeois (1981) Singh (1986) Hoffer and Alexander (1992) Zmud (1982)

Centralization Formalization Evaluation period Grover and Goslar (1993) Jung-Erveg et a! (2007) Initial time CIO characteristics Brancheau and Wetherbe (1990) CEO characteristics Lee ef d (2007) Petformane Venkatraman et a! (1994) Imitation effect Mahajan ef d (1988)

Note: 'Type is classified into Type I(b)or T w e I1 according to Swanson's classification

Table 11. Research framework

INIDS 107,9

1262

HI-2(b). In regard to LAN adoption, EAOs have a higher ratio of IT depa&ent employeesltotal employees than LAOs. The next set of hypotheses examines the differences in organizational resources of IT departments. In the first hypothesis of this set, we attempt to determine whether the reward system can be a catalyst for ICT adoption. Organizational slack may equate to higher ICT costs, training costs, and absorption force of risk, but it tends to promote ICT adoption. Organizational slack shows the degree to which uncommitted resources are available in the organization: HZ-l(a). In regard to RDB adoption, EAOs have a better reward system than LAOs. HZ-l(b). In regard to LAN adoption, EAOs have a better reward system than LAOs. HZ-Z(a). In regard to RDB adoption, EAOs have more organizational slack than LAOs. HZ-20). In regard to LAN adoption, EAOs have more organizational slack than LAOs. Lefebvre et d (1997) insist that CEO perceptions of the external environment are key issues with respect to technology policy formulation and enactment and their subsequent organizational impact in small manufacturing enterprises. ICT adoption is possible through CEO interest and support. If the CEO is young, well educated, and passionate for ICT adoption, it would be much easier for the organization to adopt ICT (Brancheau and Wetherbe, 1990). Thus, this study posits the following hypotheses: H3-l(a). In regard to RDB adoption, EAOs have younger CEOs than LAOs. H3-1(b). In regard to LAN adoption, EAOs have younger CEOs than LAOs. H3-Z(a). In regard to RDB adoption, CEOs' interest in ICT is stronger for EAOs than LAOs. H3-Z(b). In regard to LAN adoption, CEOs' interest in ICT is stronger for EAOs than LAOs. A youthful CIO could possess greater ICT knowledge and it may impact technology adoption. The last hypothesis in this set examines the relation between the CIO's attitude toward change and ICT adoption. If the CIO has a positive attitude toward change, helshe may motivate other members of the organization to adopt a new ICT. Consequently, a CIO's positive attitude toward change can advance the adoption time: H4-l(a). In regard to RDB adoption, EAOs have younger CIOs than LAOs. H4-1(b). In regard to LAN adoption, EAOs have younger CIOs than LAOs. H4-Z(a). In regard to RDB adoption, CIOs' attitude toward change is more positive for EAOs than LAOs. H4-2(b). In regard to LAN adoption, CIOs' attitude toward change is more positive for EAOs than LAOs.

. J

In addition, we examine the relationship between the degree of formalization and the time of ICT adoption, and between the degree of centralization and the time of ICT adoption. Centralization means the degree to which power and control in an organizational system are concentrated in the hands of relatively few individuals. Formalization refers to the degree to which an organization emphasizes its rules and procedures concerning the role of its members. Generally, fornlalization acts to inhibit consideration of innovations by organization members: H5-l(a). In regard to RDB adoption, the degree of formalization is greater for EAOs than LAOs. [email protected]). In regard to LAN adoption, the degree of formalization is greater for EAOs than LAOs. H5-Z(a). In regard to RDB adoption, the degree of centralization is greater for EAOs than LAOs. [email protected]). In regard to LAN adoption, the degree of centralization is greater for EAOs than LAOs. Grover and Goslar (1993) investigated whether or not environmental uncertainty, organizational factors, IS maturity and other IS factors affect ICT adoption in each phase of the process. Hence,.we attempt to investigate differences in ICT adoption between EAOs and LAOs. Earlier adopters usually try to utilize the effectiveness of ICT. Thus, the ICT evaluation period of EAOs can be shortened by intensive investment of necessary resources. Again, initial time means the time that the organization began to consider the new ICT adoption: H6-I (a). In regard to RDB adoption, the evaluation period for a new ICT is shorter for EAOs than LAOs. [email protected]). In regard to LAN adoption, the evaluation period for a new ICT is shorter for EAOs than LAOs. H6-2 (a). In regard to RDB adoption, the initial time for a new ICT is earlier for EAOs than LAOs. [email protected]). In regard to LAN adoption, the initial time for a new ICT is earlier for EAOs than LAOs. Under conditions of poor business performance, organizations often seek to streamline their operations. According to Loh and Venkatraman (1992), organizations are increasingly making ICT directly accountable for contribution to overall corporate profitability. Thus, ICT adoption might clearly help ascertain the contribution of ICT. We design the next hypothesis to detect the differences in ICT performance between EAOs and LAOs: H7-l(a). In regard to RDB adoption, an' organization's adoption behavior is affected by its previous ICT adoption experience. H7-10). In regard to LAN adoption, an organization's adoption behavior is affected by its previous ICT adoption experience.

Infomation and cOmmUnicatiOn technology 1263

~Iv~DS 107,9

H7-2(a). In regard to RDB adoption, organizational performance of EAOs is better than that of LAOs. [email protected]). In regard to LAN adoption, organizational performance of EAOs is better than that of LAOs.

1264 4. Research method 4.1 Data cohction The unit of analysis in this study is the organization. Data were collected in Korea. Data were collected by two ways: a mail survey and interviews. The survey instrument was first pilot tested by ten ICT experts such as professors, CEOs and CIOs. The modified questionnaires were mailed to 450 companies randomly selected from publicly listed ICT corporations. We sent questionnaires to CIOs of the organizations. To ensure a high-response rate and credible responses, we explained the purpose of the survey ind asked for the respondents' help by telephone calls. Out of the 450 questionnaires mailed, 80 were returned, an 18 percent response rate. A total of 64 responses were useable, with 16 excluded because of accuracy issues, such as personnel changes involving CEOs or CIOs during the ICT adoption period. In the sample of 64 firms, the number of manufacturing firms was 33, and others were 31. The other types of industries included banks and financial service firnls, distributors, and retailers, etc. Therefore, we classified the sample as manufacturing and non-manufacturing corporations. The sales volume ranged from $10.3 million to $2.8 billion, with an average of $185 million. To obtain data related to the corporations' financial statements, we used the yearly business report (1998-2000)provided by the Korea Securities Dealers Association. Data were collected including total sales, sales costs, general management costs and the number of employees.In this study, ICT was categorized according to Swanson's (1994) classification: Type Iand Type 11. Table III shows which RDB and LAN technologies were used by the organizations participating in this study.

ICT categories RDB (Na=37)

'

Table III. IT adoption frequency

Frequency (percent)

Type ORACLE DB2 MSISQL SYBASE ADAJ3AS INFORMIX PROGRESS INGRES Etc. ~otal~

,

13 (25.0) 9 (17.3) 7 (13.5) 4 (7.7) 4 (7.7) 2 (3.8) 2 (3.8) 1(1.9) 10 (192) 52

Notes: 'number of &;bmultiple response

ICT categories

LAN (Na= 41)

Type

ETHERNET IBM LAN LOCAI.NET 101NET WANGNET NECNET DECNET OMNINET KREONET Etc. ~otal~

Frequency (percent) 34 (56.7) 10 (16.7) 4 (6.7) 4 (6.7) 2 (3.3) 2 (3.3) 1 (1.7) 1 (1.7) 1 (1.7) 1 (1.7) 60

4

L

4.2 Analysis methods 4.2.1 Reliability and validity. This study used the respondents' subjectiveperceptions as measures representing the hypothesized concepts. Accordingly, to venfy the concepts in terms of reliability and validity, Cronbach's a test was used. The questionnaire used is credible since the reliability coefficients of the variables measuring CEO characteristics, organizational structure characteristics, IT department, and ICT performance ranged from 0.702 to 0.897. All a coefficients were higher than the suggested value of 0.7 (Hair et al, 1998). Factor analysis was also conducted to verrfy the conceptual validity of CEO attitudes and interest, formalization, centralization, reward system, and ICT performance. All factor loadings were greater than 0.4. Factor analysis regarding the CIO's attitude toward change resulted in three factors, as in Ettlie and O'Keefe's (1982) analysis. Correlation analysis was used to evaluate internal consistency and reliability. Correlation analysis verified the conceptual validity of CEO attitudes and interest, formaiization, centralization, reward system, and ICT performance. Correlation was found among the CEO's age, education and interest, and also among the CIO's age, education and interest, sales volume, and total employees numbers. Correlation was also found between reward and the ratio of IT budget/sales. The correlation among independent variables are calculated and summarized in Table IV. 4.2.2 Data analysis. Survival analysis was used to compare EAOs and LAOs. Traditionally, in this field, the most common strategy is to ignore the non-adopters under the assumption that they will not a£fect the research result. However, Wildemuth (1992) analyzed the non-adopters in her research, after this tool was recognized as a suitable analysis method. Singer and Willett (1991) is a good source of information for survival analysis. For RDB, the number of the uncensored subjects was 52 and the number of the censored 12. For LAN, the number of the uncensored subjects was 60 and the number of the censored 4. The uncensored data were classified almost evenly and named EAOs and LAOs. To compare the EAOs and LAOs, we made three steps. First, we assumed as if the data were censored at the point of EAOs, and the EA group and non-EA group were compared. Second, we deleted the data of EAOs, and LAOs and non-LAOSwere compared. Third, EAOs and LAOS were compared. The same procedures were repeated for LAN (Figure 1). 4.3 Hypotheses testing 4.3.1 Discussion of hypotheses regarding organizational resources. All responding organizations were divided into EAOs and LAOS on the basis of the first time they installed RDB or LAN. EAOs include organizations in the first 50 percent of the adopter frequency distribution, while LAOS include the rest. The organizations that adopted RDB between January 1993 and January 1998 were classified as EAOs (within the upper 48.6 percent) for RDB analysis, while organizations that adopted LAN between January 1993 and July 1998 were classified as EAOs (within the upper 48.8 percent). 'Tables V and VI show the results of the Wald-test on the difference of organization factors, according to the adoption order of RDB and LAN. Tables V and VI show that there is a significant difference between EAOs and LAOs, such as the JCT evaluation period, the age of CEO, degree of CEO's interest, the

Information and cOmmunicatiOn technology 1265

--

1. CEO-age 2. CEO-int 3. ClO-age 4. ClO-int 5. Sales 6. I'l'iAil E ~ n p 7. Reward 8. Slack 9. Fonualization 10. Cenlralization 11. I.AN 11111. 12. IAN Init 13. RDB L ~ x 14. KDB h i t

1 - ,244 *

-Q014 -0.%** 0.004

- 0.098 - 0.054

0.071 - 0.075 - 0.2(9 - 0.131 0.23 0.038 0.021

1 0.251 * ,349 * 0.2i9 0.048 0.109 * * -0.01 0.251 * * 0.127 0.K5 -0.159 --0.113 -0 229

:

1 0.217 0.216 * 0.274 * 0.1% - 0.217 0.165 0.018 0.12 -0.169 0.333* - 0.183

1 0.11 0.157 0.270' - 0.215 0.362 * * 0.267 * - 0.02 0026 0.023 - 0.023

1 0.0% 0.313 * * - 0.03 0.271 * 0.188 0.08 -0.085 0.154 -0.108

1 0.058 0.149 0.232 * 0.172 - 0.0.76 -0.016 -0.006 0.13.3

1 - 0.ORS

0.381 " 0.181 0.176 -0.151 -0.104 -0.281'

1 -0.077 0.057 0.017 0.043 0.012 0.059

1 0.041 -0.05 0.052 - 0.062

-0.043

1 0.182 0.074 - 0.168 0.111 -

1 -0.348 * * 0.258 -0.127

Notes: *Correlation is significant at the 0.05 level (hvo-tailed); **correlation is significant at the 0.01 level (two-tailed)

1 - 0.277

0.321 *

-

Information and communica~on technology

(a) RDB-EAOs

(c)

LAN-EAOs

NAOs

t/2

t

t

(b) RDB-LAOS

(d) LAN-LAOS

1. Assumed censored time point for EAOs . 2. Censored time point for AOs

Figure 1. Time frame of survival analysis

3. NEAOs = LAOs + NAOs

Variable CEO age CEOinte. CIO age CIOinte. Sales1 1~/all~ Reward Slack3 Formal Central Duration Initial

Average EA LA 54.70 3.64 38.69 3.62 177.70 0.028 2.68 24.43 3.15 3.54 16.87 104.40

55.98 3.23 40.16 3.56 170.37 0.018 2.38 20.51 3.03 3.53 39.93 127.83

Min

SD EA 9.05 0.59 4.37 0.37 196.07 0.067 0.78 28.34 0.82 0.62 17.54 28.93

LA

EA

7.43 42 0.78 4.89 5.85 26 0.52 4.50 369.41 17.20 0.023 0.345 0.83 4.40 28.19 1.65 5.00 0.70 0.68 4.60 29.25 56.00 38.95 146.10

Max LA

EA

LA

46 1.22

54 3.64 40 3.62 2664.1 0.028 2.68 96.29 3.15 3.54 16.87 146.10

74 5.00 53 4.90 781.16 0.166 4.80 94.95 4.33 4.60 111.00 182.63

29 2.50 9.69 0.000 1.00 2.43 1.50 1.20 0.00 0.00

Wald

Sig.

0.016 0.900 . 5.980 0.014** 7.135 0.008*** 4.730 0.030** 3.647 0.046** 8.103 0.004*** 1.303 0.054: 2.710 0.090 0.6210.031** 0.0390.844 7.107 0.008** * 8.871 0.003* **

P,o,tes: RDB (EAO: 48.6 percent (n = 18); LAO: 51.4 percent (n = 19)); *p < 0.1; **p < 0.05;

p < 0.01; 'unit: million dollar; 'ratio of number of IT department employees and total employees; 3average percentage of (sales and management cost)/(salesvolume) between 1998 and 2000

degree of formalization, reward system, the CIO's attitude toward change, and ICT performance. More details about the results of hypotheses tests are as follows: HI-1(a) is accepted and HI -1(3) is rejected with the significance level of 0.046 and 0.144, respectively.H1 -2(a) is accepted with the significancelevel of 0.004, and HI-20) is rejected with the level of 0.73. As for the financial and human resources, the mean values for the EAO group were a slightly higher than that for LAOs. That means that

Table V. Difference between early adopters and late adopters in RDB

IMDS 107,9

1268

Table VI. Differencebetween early adopters and late adopters in LAN

Variable

Average EA LA

SD EA

Max

Min

LA

EA

CEO age 53.25 57.37 9.81 4.97 40 CEO inte. 3.45 3.30 0.71 0.69 1.89 CIO age 37.40 41.68 5.09 4.53 30 CIO ~nte. 3.60 3.63 0.40 0.47 3.00 Sales1 212.47 162.43 223.84 369.74 24.73 1 ~ 1 ~ 1 1 ~0.038 0.013 0.078 0.007 0.004 Reward 2.53 2.45 0.84 0.78 1.00 Slack3 27.92 14.81 31.77 20.12 1.65 Formal 3.19 3.00 0.63 0.63 2.17 Central 3.46 3.63 0.76 0.54 1.20 Duration 27.00 32.32 29.39 21.28 0.00 Initial 67.6 74.83 29.87 28.3 0.000

LA

EA

LA

Wald Sig.

65 4.780 0.029*** 42 76 2.11 4.89 4.56 0.460 0.02** 29 46 53 3.735 0.053* 2.60 4.40 4.90 0.606 0.036' 9.7 2664.10 781.16 2.137 0.144 0.004 0.345 0.038 0.119 0.730 1.00 4.40 3.6001.9760.160 2.43 96.29 93.19 1.248 0.264 2.00 4.50 4.17 0.7290.393 2.60 4.60 4.40 0.943 0.332 3.00 97.00 73.00 2.275 0.091* 0.000 104.6 109.6 2.783 0.095*

E,fes: LAN (EAO: 48.7 percent (n = 20), LEAO: 51.3 percent (n = 21)); *p < 0.1;* *p < 0.05;

p < 0.001; 'unit: million dollar; 'ratio of number of IT department employees and total employees; 3averagepercentage of (sales and management cost)/(salesvolume) between 1998 and 2000

the influence of organization size on ICT adoption remains unclear (Lefebvre et al, 1997). For HZ, the significance levels were calculated as 0.054 and 0.16, respectively, from RDB and LAN analyses.Thus, H2-1(a) is accepted and HZ-1(6) rejected It is evident that a good reward system in the organization provides a potent influence to organizational adoption of technological innovations, especially for the case of RDB. HZ-Z(a) is accepted and HZ-2(6) rejected. Contrary to general thought, organizational slack is not always a factor that guarantees early ICT adoption. The test results of the difference between EAOs and LAOs show significance levels of 0.09 O B ) and 0.264 (LAN). 4.3.2 Discussion of hypothses regarding management styk. H3-1 (a) is rejected and H3-l(b) accepted. In RDB and LAN analysis, the age of CEO is different at the significancelevels of 0.9 and 0.029, respectively.For the adoption of RDBs, the average CEO age difference between EAOs and LAOs was 1.28 years. For the adoption of LANs, the average age of CEOs for the EAO group was 4.12 years younger than that for the LAO group. These results can be interpreted to mean that younger CEOs tend to have wider experience and broader technological knowledge, and they try to actively adopt ICT for LAN, but the same result cannot be found for RDB. H3-2(a) and 2 0 ) are both accepted with significant levels of 0.014 and 0.02. The degree of the CEOs' interest in ICT for the EAO group was greater than that of LAOs for both RDB and LAN analyses. That means the degree of the CEO's interest can affect ICT adoption in the organization. As hypothesized, the CEO's perception of external change is an important factor with respect to ICT strategy formulation and action programs (Lefebvre et al, 1997).The implication of this result is clear, as misread or misunderstood environments will result in ineffective ICT strategies.Thls in turn will probably translate into sub-optimal resource allocation, which could be detrimental to ICT performance. H4-1(a) and 1 (6) are both accepted. The significancelevels of the age of CIOs between EAOs and LAOs were 0.008and 0.053, respectively.It appears that the age of the CIO has a significant iduence on ICT adoption. H4-2(a) and H4-Z(b) are also both accepted.

d

RDB and LAN analyses indicate significance levels of 0.03 and 0.036, respectively. This means that the CIO's attitude toward change is related to ICT adoption. A CIO who has a positive attitude toward change strives to acquire information about a new ICT as soon as possible, and helshe has a strong tendency to apply this knowledge to hlsher organization. Consequently, the organizational innovativeness is strengthened. H5-1(a) is accepted and H5-1(b) rejected, with the significance levels of 0.031 and 0.393. That indicates that a high degree of formalization plays a part in the early adoption of ICT in Type I ICT. H5-Z(a) and 2(b) are both rejected with significance levels of 0.844 and 0.332, respectively. Accordingly, the degree of centralization for EAOs is not different from LAOs. H6-1(a) and 1 (b) are both accepted with the significance levels of 0.008 and 0.091, respectively. The time lag of ICT adoption from planning to implementation for EAOs is shorter than that for LAOs. In the case of RDBs, the time lag is approximately 23 months. In the case of LANs, the evaluation period time lag for ICT adoption between EAOs and LAOS is roughly 5.32 months. The result also shows that the evaluation period for RDBs is shorter than that for LANs for EAOs and the opposite results for LAOs. H6-2(a) and H6-2(b) are both accepted with the significance levels of 0.003 (RDB) and 0.095 (LAN). Since, the initial time is the starting point that the organization began to consider the adoption of the new ICT, it means the EAOs are more interested in adopting new ICT than LAOs. 4.3.3 Discussion of hypotheses regarding coqorate strategy. H7-I (a) and [email protected]) were tested by using the time series data as shown in Table W. The white-noise model, representing only internal influences, is set as the null hypothesis. The external influence model and the mixed influence model are set as alternative hypotheses, using Mahajan et al's (1988) analysis process. The parameters of the white-noise model, external influence model and mixed influence model were tested by ordinary least square as shown in Table VIII, using the time series data for each ICT. Table IX presents the results of each ICT's test for H7-1. As shown in Table IX, the t-value of PI coefficient of the external influence model was significant at the 0.05 level and had the same direction as the estimated parameters in Table VIII. Both the adjusted R 2 and mean absolute deviation were superior to those of the white-noisemodel in terms of goodness of fit. Yet, the signs and values of the estimated parameters were contradictory to the signs and the values of the expected parameters in the mixed influence model. In addition, as noted in Table X, the null hypothesis was rejected at the p < 0.05 significance level (based on the F-test of the external influence model and white-noise model). In conclusion, it was established that organizations that have already adopted RDB do not affect decision making by organizations that are considering adopting RDB. Hypothesis testing regarding LAN (Column I1 in Table W) is as follows: as shown in Table IX, the sign and value of the estimated parameters for the external influence model match the sign and value of expected parameters in Table VIII, but that is not the case with the mixed influence model. The t value of the external influence model was 0.5640 and it was significant at the 0.019 level. Also, the adjusted R 2 of the external influence model was higher than that of the white-noisemodel, yet the value of the mean absolute deviation was similar. This result means that the external influence model is better than the white-noise mod61 (internal influence model) when explaining the LAN adoption behavior.

Infomation and cOmmunicatiiOn technology

1269

IMDS

-

Number of IT adoption organization

107,9

RDBrn

Date of adoption

1270

Table W. ICT adoption time-series data

40

93 112 93 314 94 112 94 314 95 112 95 314 96 112 96 314 97 112 97 314 98 112 98 314 99 112 99 314 00 112 00 314

RDB O

1 2 0 2 2 2 1 1 6 0 4 6 2 2 5 1

LANO

N(t) 1 3 3 5 7 9 10 11 17 17 21 27 29 31 36 37

1

o 0 0 1 0 1 0 5 2 9 6 4 2 9 1

N(t)

x(t)

1 1 1 1 2 2 3 3 8 10 19 25 29 31

2 2 0 2 3 2 2 2 1 11 2 13 6 4 14 2

40

41

+L ~ O N(t) 2 4 4 6 9 11 13 14 25 27 40

52 58 62 76 78

-

Notes: 112 - first and second quarters; 314 - third and forth quarters; x(t) - the number of ICT adopted organizations at time t; N(t) - the cumulative number of ICT adopted organizations at time t J

Parameter

Hypothesis

Model type

Equation

&

h

Null hypothesis

White - noise model Ex%ernalinfluence model Mixed influence model

x(t) = x(t - 1) + ~ ( t )

1

-

+ x(t) = Plx(t - 1)+ EN *(t- 1) + ~ ( t )

1

Suggest Documents