Comps readings this week Diffusion of innovations questions figured prominently in the folder of comps questions - seemed like nearly everyone had a question relating this area to another area, so this finishes up the readings I had on diffusion of innovations. (this post was added to throughout the week and finished after the post on comps preparations. there will probably still be some posts on comps readings, but I'm supposed to be doing more integrating now - and not in the fun math way!)

Ilie, V., Van Slyke, C., Green, G., & Lou, H. (2005). Gender Differences in Perceptions and Use of Communication Technologies: A Diffusion of Innovation Approach. Information Resources Management Journal, 18(3), 13-31

This article looks at how gender impacts "intention to use" IM. The standard Rogers things: perceived ease of use, relative advantage, compatibility, observability (broken down into visibility and results demonstrability), plus perceived critical mass. Note that all of these have been studied by people other than Rogers, and they are all "perceived" - it's not Gartner's assessment of relative advantage, it's what the user thinks. They also review the extensive literature on gender differences in communication- both in person and online. The participants were business students who were of course heavy users of icts. They did a survey, and used scales from other studies for most items and made their own for perceived critical mass. I'll leave the stats to anyone who's interested to read. The men were into relative advantage,semonstrability , and perceived critical mass. Women were into ease of use and visibility. This matched with the hypotheses.

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425-478
MIS Quarterly articles are pretty meaty. This one seems even more so - jam packed with information. It's been cited about 520 times, 387 times from journals (from WoS).

This article compares 8 models of IT acceptance research and comes up with a unified model. The original 8 models were tested to see how much variance they explained, then the new model tested against the first set of data (improvement) and then the new model against new data (decent adjusted R^2).
The 8 models:

1. Theory of Reasoned Action (TRA)
   • attitude toward behavior
   • subjective norm


2. Technology Acceptance Model (TAM)
   • perceived usefulness
   • perceived ease of use
   • subjective norm (added as TAM2)
     


3. Motivational Model(MM)
   • extrinsic motivation
   • intrinsic motivation


4. Theory of Planned Behavior (TPB)
   • attitude toward behavior
   • subjective norm
   • perceived behavioral control


5. Combined TAM and TPB
   • attitude toward behavior
   • subjective norm
   • perceived behavioral control
   • perceived usefulness


6. Model of PC Utilization
   • job-fit
   • complexity
   • long-term consequences
   • affect towards use
   • social factors
   • facilitating conditions


7. Innovation Diffusion Theory (not really strictly Rogers, more like various IS takes on Rogers)
   • relative advantage
   • ease of use
   • image (if use enhances user's image)
   • visibility
   • compatibility
   • results demonstrability
   • voluntariness of use


8. Social Cognitive Theory
   • outcome expectations-performance
   • outcome expectations-personal
   • self-efficacy
   • affect (liking the behavior)
   • anxiety

For each of these things, there are "moderators" including experience, voluntariness, gender, age. There might also be things about the industry, job function of user, and things about the technology itself - but these aren't considered here.

Some of the limitations of previous studies were that they were mostly done with students, they were done retrospectively, they were done with completely voluntary innovations (but managers need to know how to get employees going), and the technologies were fairly simple.

The authors find 4 organizations in 4 different industries with samples drawn from different functional roles, and give a survey immediately after training on a new technology, a month later, and 3 months later. They also gather "duration of use" for 6 months after the training to look at actual usage. The questions are taken from scales from the studies supporting the 8models and were pre-tested and tweaked... then follows a lot of statistics and testing for validity, reliability... etc. The 4 orgs were grouped into 1a + 1b (voluntary) and 2a + 2b (mandatory). Essentially the older models each accounted for at most between 40-50% of the variance in intention and usage.

The new model (for fear of getting in trouble, I'm re-drawing)
Each of these pieces is pulled together from pieces of the 8 models - so they use the questions from the other models, and group them together this way. They did the standard tests to see if they hung together and tested some hypotheses about the moderators. They then got data from 2 more orgs and ran the new model against that. It essentially worked pretty well. The new model accounted for like 70 percent of the variance of usage intention. There are definitely some limitations - partially due to sample size and the shear number of variables.

Labels: comps

  ¶ 10:40 PM| |cites (technorati) |