Hypothesis Testing

By | February 9, 2018

 

Hypothesis Testing

A second indicator that the IS-CRM subfield is maturing would be the emergence in the litera- ture of hypothesis testing to develop support for IS-CRM theory. Specifically we made this recommendation:

Second, once theories have been developed, there is a need for lab and field experiments to test hypotheses in order to find support for them and rule out other possible explanations. (Romano and Fjermestad 2001–2002, p. 85)

A quick scan of the current IS-CRM literature reveals that hypothesis testing is more often incorporated in studies now than during the period covered in our study. Madeja and Schoder (2003) also tested their research hypotheses on 224 companies that target consumers (B2C) and found that information- and functionality-richness (interactive character) and continuous updat- ing of the Web site are key drivers of success for general companies. Chen et al. (2004) also tested four major hypotheses based on their proposed theoretical model and found that the model sub-

stantially explained and predicted consumer acceptance of virtual stores and also explained many of the factors that lead to the user’s intention to use, and actual use of, a virtual store. Madeja and Schoder (2004) constructed a covariance structure (LISREL) model and tested three research hypotheses with a data set of 469 cases of general companies in the German-speaking market. They found that CRM is a critical success factor in electronic commerce, independent of how long the company has been on the Web, and that CRM is especially critical for B2C and small companies. These are three among many papers emerging in the literature that illustrate an in- creased use of hypothesis testing—providing evidence of the maturation of IS-CRM over time.

Instrument Validation

In our study only a very small percentage of the survey articles discussed instrument validation, and not all of these discussed reliability testing or validity in detail. We made the following rec- ommendation:

Third, there is a strong need for researchers to validate the instruments they employ and to explain these procedures in their articles in order to evoke confidence that the results are meaningful, interpretable, and reliable. Instrument development and validation must be carefully undertaken prior to use. Researchers in this new subfield need to explore referent disciplines, such as psychology, and use methods that validate instruments from a number of perspectives, including convergent validity, discriminate validity, construct validity, and reliability. (Romano and Fjermestad 2001–2002, p. 85)

 

In the current IS-CRM literature we were able to easily find articles that discuss instrument validation in greater detail than we found in our earlier study. Susarla et al. (2003) provide a detailed analysis of both discriminant and convergent validity as well as the reliability of the instrument they developed for their study. Croteau and Li (2003) also discuss both discriminant and convergent validity, and they used structural equation modeling (SEM) with partial least squares (PLS) to as- sess their model. Stefanou, Sarmaniotis, and Stafyla (2003) relied on previously developed instru- ments when possible and performed a detailed factor analysis and additional tests for reliability, sampling adequacy, and sphericity. Pennington, Wilcox, and Grover (2003) also discuss reliability, convergent validity, and unidimensionality of the instrument they used and adopt existing measures when possible. These papers illustrate how IS-CRM researchers have begun to seriously validate their instruments and discuss the results of these tests in their published studies, and they clearly demonstrate maturation in the IS-CRM subfield of MIS research.

Cumulative Tradition of Research

We also found that, although there were many papers on CRM, there was not a growing tradition of research or body of knowledge that was being cited. We made the following recommendation:

Fourth, there is a need for a cumulative tradition of research in which replication, extension of theories, models, and instruments, and development of standard constructs and metrics define the subfield and give each new study contextual meaning within a common body of knowledge. There is a need for depth as well as breadth of research. (Romano and Fjermestad 2001–2002, p. 85)

This is the most difficult aspect of IS-CRM research to assess without doing a full co- citation analysis (Culnan 1987), which is beyond the scope of this chapter. To truly establish a tradition of research would require a much longer time frame—a decade or more. We have, however, been able to see—at least in the area of trust in IS-CRM—that there appears to be a growing body of knowledge that is leading to a tradition of research. Two of the leading IS researchers in this area, D. Harrison McKnight and Norman Chervany, have published a num- ber of papers on the topic that are cited in the majority of papers on trust in e-commerce and trust in information systems (see McKnight 2000; McKnight and Chervany 2000; McKnight and Chervany 2001; McKnight and Chervany 2001; McKnight and Chervany 2001–2002; McKnight et al. 2002).

The AMCIS and HICSS papers were published in the CRM minitracks of these respective conferences, and the IJEC paper was published in the first special issue on CRM that the authors guest edited. These papers are frequently cited now in articles on trust both inside   and outside the IS discipline, showing the emergence of a tradition of IS-CRM research on trust. We hope that over time other areas of IS-CRM research will also begin to develop traditions  of research.

 

Publication of IS-CRM Research in Top MIS Journals

 

In our assessment of IS-CRM research (Romano and Fjermestad 2001–2002) we found that few articles had yet been published in the core MIS journals, such as JMIS, ISR, and MISQ, relative to other IS research topics. We made the following recommendation:

 

There is a need for researchers in the ECCRM subfield to submit their work to the core MIS journals (Journal of Management Information Systems, MIS Quarterly, Information Sys- tems Research, Information and Management) in order to increase the perception of its maturity within the MIS research community at large and make it a true subdiscipline of MIS. (Romano and Fjermestad 2001–2002, p. 85)

 

We explored the IS-CRM literature in IS journals by reviewing every paper published in the top-tier IS journals, based on past and recent published journal rankings (Gillenson and Stutz 1991; Holsapple et al. 1994; Walstrom et al. 1995; Hardgrave and Walstrom 1997; Walczak 1999; Whitman et al. 1999; Mylonopoulos and Theoharakis 2001; Walstrom and Hardgrave 2001; Peffers and Ya 2003; Lowry et al. 2004), also the second-tier journal Information and Management, and the top-ranked journal in e-commerce, International Journal of Electronic Commerce. The latter two had a relatively large number of IS-CRM articles in the initial study, so we thought they were likely to publish additional articles in this area. The criteria for inclu- sion in the count were that the main theme of the paper was IS-CRM, a secondary but substan- tial theme was IS-CRM, or the paper built on previous IS-CRM research. Table 1 shows the number of publications from 1984 to early 2002 and then from December 2001 to November 2004, illustrating that CRM publications in top-ranked IS and related journals continue to in- crease in number. The overlap in the dates has to do with the availability of papers at the time  of the first and the present review of the literature.

The data in Table 1.1 show that IS-CRM articles continue to be published in top-tier IS journals at an apparently increasing rate. In the less-than-two-year period of the second review of the literature every journal published at least half as many articles on IS-CRM as it had in the previous 16–year time span of the first review. Overall, since the first study,

Leave a Reply

Your email address will not be published. Required fields are marked *