What I've learned is to be VERY systematic and methodical in diagnosing a problem. Most errors that I see in technicians are based on their "thinking" that they know the casue of the problem and heading right to that spot to begin their repairs. Of course, for the pro, time is money, and if he can beat the book a couple of times he might actually make a buck that day.
I work somewhat differently and I've been at if for 35 + years.
I always start with the basics -- every time -- even if it seems obvious that they are all covered. I also operate on a couple of principles that I've learned over my years of doing this sort of work.
1. Start with the battery, wires, fuses, etc. (charging volts is a simple meter test -- wiggle the cables to see if anything blinks)
2. Check fuel supply and PSI (takes only a couple of seconds to snap a guage on a fuel port fitting)
3. Check all fluids
4. Do a check for vacuum leaks - both visual, feel, sound, and guage (I use starting fluid - SPARSELY sprayed around vacuum fittings and areas of known leaks -- like the intake on Explorers)
4. Scan codes
5. Based on codes (or lack of codes) begin a diagnosis based on symptoms.
A. It is probably electrical rather than fuel related
B. It is probably vacuum rather than electrical related
C. After eliminating vacuum and electrical issues, check the fuel system
D. If none of the above turns out anything, start checking for hard-part failure -- rounded cams, bent pushrods, burnt valves, bad rings, etc. This may require disasembly and/or compression - leak-down testing.
6. Begin testing individual components. For this, use a Vantage, a Vantage Pro, or a GOOD VOM that you know how to use. It will need to be digital, need to have square wave capacity, and need to measure milli-amps.
7. Begin replacing bad components and re-test.
The above is typically my routine for diagnosis.
Also know again that NEW parts are not always GOOD parts. Especially with the cheap parts houses in business these days. I've seen so many "new" but bad fuel pumps (oh, they pump, but not anywhere near what is required -- test for this using a Modis or Vantage (Pro). You can actually count the individual stators on pumps with a good scope, and you will soon see if they are bad or not (or made with fewer than the factory spec stators!).
Dirty alternators (as in the power they produce not being even and "clean" - not that they have dirt on the cases) are one of the worst offenders for ECM problems. Your car computer is just like your home computer -- it needs a clean, stable power supply.
Testing individual components consists of "pinning out" (sticking little wires, pins, or probes of some type) into the proper wires or wire connectors that feed that particular sensor, motor, or circut, then reading a test voltage, square wave pattern, or amps draw to see if the part is within specs.
On an O2 sensor, one can check to see if they are reading properly by using either a good scanner (not a code reader) that will actually show real-time data from the computer to see if the O2 sensors are changing as the car warms up, and then if they assume a steady state once it does. Erratic readings indicate that the O2 may indeed be bad. O2 sensors can also be tested by the meter method above.
Tip for changing them... Heat them up (a small propane torch or a micro torch works great) -- and remove them while they are hot. NEVER use penetrating oil like WD 40. That will ruin a good O2 sensor and also will WELD a new one in the bore once it gets hot. Use anti-seize designed for that application only. Get the correct socket for removal - makes all the difference in the world. Don't scrimp here, unless you actually LIKE drilling and re-threading the bung for a new O2 sensor.
Hope all this stuff helps someone out there.
Now if I could only find enough spare money to get my hands on a Snap On Modis... Ultimate diagnosis tool in the world today. I can actually diagnose compression ratios with the scope in the right settings.