ADDIE (Analysis, Design, Development, Implementation, and Evaluation) and HPT (Human Performance Technology) have broad similarities, but there are enough differences to merit a deeper look. The biggest difference, for me, is that ADDIE focuses explicitly on building training programs, while HPT aims to develop an appropriate solution, which might be "training, performance support tools, a new or re-engineered process, the redesign of a workspace, or a change in compensation or benefits" (ISPI, p. 3). HPT advocates a systems view of performance improvement, in which training is one potential tool.
Both models include evaluation, but ADDIE (as classically defined) places it at the end of the process whereas HPT welds it into the middle portions of the system as well. HPT also explicitly asks about retention of information by including a "Confirmative" evaluation of learners ("Continuing Competence"), organizational goals ("Continuing Effectiveness"), and value ("Return on Investment (ROI)").
To some degree, it seems like a red herring to compare ADDIE to anything on a fine-grained level since, as Morrison et al. (2007, p. 13) point out, ADDIE does not exist as a formal model and it has never been written down in an authoritative way. The beauty of ADDIE--its simplicity--is also its downfall here, since it tends to morph into whatever some author wants to call it. ADDIE takes a lot of flak because authors want straw men to argue against. At its heart, ADDIE offers a reminder to consider certain kinds of tasks when designing instruction, and its value as a "mental checklist" is high.
I think HPT is valuable for the same reason: it offers a systematic, visible way of thinking about performance deficiency and improvement. As Reigeluth says, it reminds us to ask important questions about the solutions we propose, and gives us tools for knowing whether the process is finished.
For my example of ADDIE versus HPT, I will take a real-world example that actually crossed my desk a few months ago. In my work at the hotline, I spoke with a woman who was very upset about her husband's habit of leaving the light on in the bathroom after he left the room. His habit was very firmly ingrained, and the woman found herself being significantly annoyed every time she saw the bathroom light left on. She asked me how she could teach him to turn the light off.
ADDIE might have me analyze the situation, determine that the husband did know how to use the light switch, and eventually design some sort of training that explained the deleterious effects of light-switch-left-on syndrome on their marriage and their shared bank account, and went on to teach light extinction skills. Had I done this, I suspect my evaluation would have shown a failure to change anything.
HPT addresses this as a performance problem. In analyzing the learners, I determined that the husband, though very good-hearted, was often absent-minded. The problem wasn't that he chose not to turn off the light--it was that he didn't notice the light at all. Is training the right solution? No! I suggested a workplace modification: that she install a motion-sensing light switch in the bathroom. Evaluation will show that the original performance problem--leaving the light on--was eradicated immediately and did not recur.
A few other thoughts that interested me:
Gram (2009, September 9) feels "Adherents and crankites alike view ADDIE as an 'instructional design' methodology when in fact it should be viewed more as a project management process for learning projects. Viewing Instructional Design as synonymous with ADDIE does both a disservice. There is loads of ID going on inside ADDIE but it is primarily in the Design phase of the process, and it can be much more creative than the original model prescribes." This meshes nicely with my own observations: most experts learn a model and then improvise using its tools, rather than slavishly following the model.
Bhattacharya (2006) argues that structural models like ADDIE and HPT are losing their relevance because they demand adherence to a relatively strict procedural model in an era where rapid development is the norm. He also argues that most actual developers follow something he calls a "seismograph" pathway through models like ADDIE, rather than the linear "waterfall" other authors sometimes describe. Although some of his points involve arguments with straw men, the central tenet is valuable: instructional design ought to stay in touch with its learners and employers throughout the process, not merely during Analysis and Evaluation.
Jarche (2006, November 7) points out that models like ADDIE and HPT are, by nature, better suited to _training_ tasks (those with specific end-point goals) than to _educational_ tasks (those that aim to increase general knowledge, skills, etc.) because these models demand specific measurements. He goes on to argue that models like ADDIE and HPT are still valid today, but that they need careful application. He writes, "too often we see training as a solution looking for a problem." At its heart, this favors HPT somewhat, since HPT explicitly includes non-training solutions. But the over-arching admonition is to analyze the situation, determine what kind of solution is needed, and _then_ choose a model for getting there.
Bhattacharya, A. (2006, November 8). Big question: ISD / ADDIE / HPT: still relevant? Retrieved May 12, 2010, from http://incsub.org/soulsoup/?p=739
Gram, T. (2009, September 9). ADDIE is dead! Long live ADDIE! Retrieved May 12, 2010, from http://gramconsulting.com/2009/09/addie-is-dead-long-live-addie/
International Society for Performance Improvement. (2009). What is human performance technology? Retrieved May 12, 2010, from http://sylvan.live.ecollege.com/ec/courses/46324/CRS-CW-4119171/What_is_Human_Performance_Technology_%28HPT%29.pdf
Jarche, H. (2006, November 7). Whither ISD, ADDIE & HPT? Retrieved May 12, 2010, from http://www.jarche.com/2006/11/wither-isd-addie-hpt/
Morrison, G. R., Ross, S. M., & Kemp, J. E. (2007). Designing effective instruction (5th ed.). San Francisco: John Wiley and Sons, Inc.