psychology - What research is there suggesting modal dialogs are disruptive?

It seems "Everyone" knows they're highly disruptive for many situations but often times clients don't. Particularly in desktop applications this form of dialog is often misused, often disrupting a workflow.

I intellectually know all of the reasons to avoid them and some rare cases they may be warranted, and we have a great deal of anecdotal evidence and professional opinions in our modal-dialog questions.

I know there is published research on modal dialogs, but to my dismay few people reference this research when supporting or opposing the use of modals dialogs.

I know for example that distracting users while in a state of flow (such as writing an essay in Microsoft Word) can be easily destroyed by interruptions such as modal dialogs, having disastrous results on productivity.

In my particular case the issue is attention (working on the mistaken opinion that modal dialogs attract attention and comprehension) and productivity as the dialog disrupts workflow for a "confirmation".

What are the most convincing resources to give to a client who believes modal dialogs are not problematic?

I am particularly looking for scholarly articles, but high level explanations (which cite real data) may be applicable as well.

Wikipedia has a good list of "common knowledge" about and criticisms of modal windows, but even they mostly cite blog posts and a couple books.

Answer

Generally speaking, disruptions and distractions negatively affect human performance, a common finding in cognitive psychology. Many studies have shown that distraction greatly increases task time on a wide variety of tasks.

There also exist many Quantitative studies showing task performance is negatively affected by distractions (note these figures are not specifically for modal dialog distractions):

We can all relate to the above distractions, but what is their quantitative effect on our performance?
- Forgetting the as-left conditions: 45%
- Forgetting to return to the original task: 25%

- Original task out of control during distraction: 17%
- Not knowing changes after returning to original task: 13%

How does this relate to modal dialogs? Users find dialog windows distracting. The linked University of Minnesota study found users rated dialog windows to be considerably more distracting than two other tested window types.

Dialog windows in particular show an extreme effect on task performance (emphasis mine):

Results show that when peripheral tasks interrupt the execution of primary tasks, users require from 3% to 27% more time to complete the tasks, commit twice the number of errors across tasks, experience from 31% to 106% more annoyance, and experience twice the increase in anxiety than when those same peripheral tasks are presented at the boundary between primary tasks.

Subjects unanimously rated the dialog window as the most distracting (intrusive) awareness technique.

Other research has specifically found that interruptions from modal dialogs greatly increase error rates in a simple form workflow exercise.

On average, errors were made 7.14% (SD = 8.06%), 17.26% (SD = 7.28%) and 25.89% (SD = 9.88%) of the time in the zero-, one- and two-interruption conditions respectively

In addition, research on video games have shown that Calm Messaging improves task performance (pdf download link). Emphasis mine:

Games have shown that reducing demands on the user’s attention can aid performance; through the use of sound, speech, transient text, and animation, games communicate in a calm manner that promotes a fluid, uninterrupted workflow.

Bottom line: Modal dialogs can double error rates, increase time to task completion, and are near-universally despised by users. Alternate means of notification are often available and should be utilized wherever possible and appropriate.