I have DynoLicous (iPhone) and have done about 6 runs on it with the '90, 1.6, 5 speed, open diff, bone stock. It does have sticky, nicely worn Star Specs on the rear axles.
For averaging purposes, I deleted the highest and lowest result for dry and wet, which is why only two results remain, but none of the 6 were out of line.
With a very tired 1.6 (smokes, rattles, 270+K on the clock) I have 8.35 and 8.31 for 0-60 times, and 81 and 87 RWHP with a 23% estimate (mine) entered for drivetrain loss and a 2390 pound vehicle weight (again, my estimate).
Other settings I have are rollout at zero feet and trigger at 0.06 G's. It's my understanding that with these apps the setup is critical for accurate-ish results. Also, the phone must be rigidly mounted. I jam mine tight between the console and the passenger seat. It's worth it to keep in mind that due to the way the timer is triggered, your reaction time is pretty much perfect, something that would probably not be the case at a real track.
If I remember right, I dropped the clutch at 2500 each time. Just an arbitrary number I picked for consistency and without wanting to stress things too much on some tired driveline components.
I don't see any glaring reason to think the HP values are not close, especially considering the motor condition and the fact that I could be light on my vehicle weight guess, and of course the times are measured.
So, at $100/hr ... NOT ... I'm OK with the app results as an alternative to real dyno time. I think these apps do offer some value, provided that you accept that it's no substitute for real track or dyno numbers, but a useful tool none the less.
When I do my 1.8 swap I plan to throw a hundred bucks at the dyno and get a baseline right off the hop. I'll get it on the scale sometime too.
Then I'll do a few runs and adjust the drivetrain loss factor so that the app and the dyno agree. I'm hoping then that it will be a little more useful and can save some hours and bucks for the "real" dyno for tuning.