Experiment

I'm good except you don't need to do the duration test portion, that data is available. Its like this for wood:

10 years 1.0

2 months 1.15

1 week: 1.25

10 minutes: 1.6

Impact 2.0

Meaning that wood can take 2x the impact load that it can a 10 year occupancy load. 

The 10 minute load is typically used for peak wind and earthquake loads for buildings but our rolling earthquake-mobiles get a lot more than 10 minutes of "earthquakes". Probably the 1 week duration would be more appropriate for us.  

So probably the best thing would be to apply the test loads for 10 minutes to each sample screw, average those loads, divide that by 1.6/1.25=1.28 and call that the max bump load rating. Since the bump loads are about 2gs that means that a factor of about 2.56 over the actual weight on the screw should probably be used. 

There is also the typical engineering factor of safety added on top of that. I just did a quick look for safety factor for wood fasteners and found 6x is typical for pullout loads and 1.6 to 2 for lateral loads because of variability, humidity, and aging, 

Using the 1.6 number I get an overall multiplier of 2.56 x 1.6 - 4.1, call it 4, or an allowable 1/4 of the average lateral test load at failure. 

Personally, I think my method of a combination of a coarsely threaded short screw with a patch of 3M double sided extra strong foam tape will win the contest, but I have been wrong so many times in my life ......  Suffice it to say, I don't patronize our fine casinos here in Dayton, NV.  

Love it!  I'm with OG.

For corroboration, We used a factor of safety of 4 for wood poles carrying transmission lines.  The poles were Southern yellow Pine, one of the strongest woods. 

Generally, bumps should be considered impact loads and doubled for equivalent weight. 

Please retest with your screw almost screwed all the way in and loaded parallel to the wood panel (90 deg/in shear).  I'll bet the result is about twice the 37 lbs you got. 

Then, your load rating should be that divided by 4 for safety and divided again by 2 for impact.  Let's guess and call it 10 lbs.  *I* would allow double that for 2 screws on the top row if they are at the same "elevation", but only allow half that for screws in an additional row (5 lbs/screw).  There is a shear strength and buckling load limit for the panel, so don't go hanging propane bottles on your walls.    I also wouldn't place the screws closer than about 4" apart, as the stresses from one overlap and add to the stress from the other.  Admittedly, 4" is a guestimate without further data. 

+1. We engineers are often accused of being a conservative bunch, but generally our customers aren't happy with us when our designs fail. You might go a long time without a failure with lower safety factors, but statistics being what they are it could happen tomorrow.