I have been a power engineer for a long time. I have personally acquired the following pieces of test equipment:

1. Several scope voltage probes, (one HV voltage probe (V's > 500V))
2. A good oscilloscope current probe.
3. A "cheater plug" which isolates earth ground from the oscilloscope. Helps to preserve ground clips during AC/DC investigations.
4. A network analyzer, if you do not want to do the trial and error method. My book and Ray's are good for estimating the feedback needs, but they do not provide the ultimate answer. Ray's analyzer is great as well as Dean Venable's.

If anyone else wants to make a contribution, please do.

I don't like to recommend the cheater plug. I prefer and isolation transformer and variac so I can ground things.

Yes, I use them sometimes, but floating your test equipment can bite you.

You need a good function/signal generator for stability measurements.

Programmable AC source. In addition to not having to float your scope, the adjustable peak and average current limits reduce the potential for explosive surprises.

Programmable electronic load is also very handy to see dynamic response of the power supply

All good suggestions. I do have programmable load (a MOSFET and resistor also work), and several DMMs.

I brought this thread up because I had gotten into many discussion with young engineers that may not appreciate the depth to which a power engineer must peer into the operation of his/her power supply. I appreciate the response and I am glad all of you are sensitive to this.

 I have given up the cheater plug and gone to (and now love) using a HV differential probe such as a Tek P5205. When using regular scope probes, I've given up the usual ground lead and will mostly use the spring-type ground like the Tek 016-1077-00, which is also available from third parties.

I found a HP 3577A Network Analyzer on Ebay for $700 about 4 years ago... The problem was that it was not working and sold as is . I took a chance and bought it... Downloaded the service manual and tore down the unit to find the problem... I fixed it, calibrated it and cleaned it up... It now works perfectly .. I built an isolation box so I can hook into the power supply and measure Gain and Phase plots for stability right on my bench at home..
I eventually picked up a HP pen plotter with IEEE cable and now have the pen plotter graphs.. I hook it up to everything I can, even measure audio amplifiers, RIAA phono stage measurements. You can hook it up to reel to reel tape recorders to see how jittery the wow and flutter looks like..

In general double check the result of the measurements regardless of the tools you have.
Two brief examples in my experience. Although there are many more.

Sent a customer a power supply which was fine. They came back with it's unstable when we tested it. Checked again and it looked fine. The problem was with the model/brand of the scope they were using. Showed them in the lab that with a Lecroy, Tek and HP the step load was consistent. When the test was repeated using the scope they had it failed.

Doing PFC testing of different diodes I expected to see a trend in efficiency based on diode technology. When I compiled the data rather than an upward trend as expected I got a zigzag pattern. The problem was with the meters I was using not the parts. Noise sensitivity at the different ringing of the diodes skewed the measurements.

The lesson I learned in these cases was if you don't see what you expect don't assume it is a problem with the design. Check the test set up and instrumentation.