Doug Bowman and Ryan McMahan of Virgina Tech wrote an interesting article titled "How Much Immersion Is Enough" which appears as a cover feature of July's IEEE Computer magazine.
They discuss several aspects of Immersion (defined as "the objective level of sensory fidelity a VR system provides") and presence ("a user’s subjective psychological response to a VR system") and talk about the main factors that drive immersion, including field of view, field of regard, resolution, stereoscopy, head-tracking, frame rate and more.
The article describes several studies that investigate how the degree of immersion impacts the performance of a person trying to complete complex tasks (such as planning the path of an oil well or visualizing a tunnel through rock structures) and describes the significant and measurable benefits attributed to higher degree of immersion.
It is nice to see measurable benefits tied to the "wow" feeling that people experience when they try on immersive head-mounted displays that offer both panoramic field of view and high resolution. My company has built several high-immersion models by optically stitching together small micro-displays. One unique side-effect of this approach is that during demonstrations, we are able to turn off individual displays and thus take a user through a full range of immersion options (e.g. from 150 degree field of view to 120 to 100 to 80) within seconds. This has been an excellent way to experience the benefits of true immersion.
Once you've had a chance to try a "business class" seat in an airplane, it's not easy to go back to economy. Similarly, once you try on an HMD or a CAVE with wide FOV, it will be difficult to go back to narrow displays.