Doing some Googling and found an old post, it looks like one I wrote, and in 2015 "Stick" copy and pasted into another forum.  It was part of a study I did on various joint strengths.  Repeating here because I cannot find the original

-------------------------------------

All the data I see on dowel joints test against "young" joints, not one that's been through several seasonal variances in humidity.
I thought R. Bruce Hoadley put this one to bed 40 years ago in Fine Woodworking...
As he says, "If good dowel joints aren’t the oldest joints ever made, loose ones must be."
Preview - The Dowel Joint - Fine Woodworking Article

tests run on "fresh" joints. And the tests are "racking" stress only.
And I am skeptical of anyone who says, "500 pounds of pressure" Force is measured in pounds, pressure in pounds per square inch, and torque in foot-pounds (or the metric counterparts).

"For every complex problem there is an answer that is clear, simple, and wrong." (H.L. Mencken)

Joint strength is a complex problem....

There are variables in the material

• Variations in wood species, moisture, density, propensity to split, wood movement indices, etc. etc.
• Lumber vs. engineered wood
• Variations in wood axes - Anisotropic – Orthotropic properties of radial, tangential, and axial planes (Hooke's Law)
• Fit / gaps (is a "weaker joint" that fits well better than a "stronger joint" that is ill-fitting?)
• Surrounding wood failure (even if the "joint" A is 10 times as strong as another type of joint B, if the wood fails first on both, it's a moot point)
• Depth of penetration of joint from one component to another

Another of the complexities are the types of stress. I identified 6 kinds o stress on a joint --

1. shear
2. compression
3. tension
4. racking
5. cleavage
6. peel.

Did you ever see a dovetail on a chair leg or a half-lap on a drawer?

There are also different ways the wood is joined

• side-grain to side-grain (at 0 or 90, or some other angle)
• end-grain to end-grain
• end-grain to side-grain
• end-grain to face-grain

Joints also have multiple attributes

• Speed to make
• Appearance
• Cost, especially if new tooling or special per-unit materials involved
• Strength

And there are multiple stress-failure considerations

• Soft fail (gradual loosening) vs. hard fail (fails with a crash)
• flexibility vs rigidity
• Repairable vs. non-repairable (the "shear pin" factor)
• moisture fluctuations over time, differential contraction's effects on the joint
• Dynamic (sudden) vs. static (continuous) stress

Most every test is done on a "fresh joint" not one that's gone through multiple wet-dry seasonal cycles. (one of the dowel jig vendors that claims they're the strongest joint specifies materials & their dimensions, type of test, and other considerations to tip the scale.
Unless you have end-grain to end-grain (a very uncommon joint in practice), the wood will be moving in different directions and tend to break the glue joint in at least one side of the joint
You have minimal glue surface (on a round dowel) and most of that will be end-grain to side grain
You have limited penetration distance if you are using dowel pins (usually max at 3/4 to 1" each side)

And the big question is, "Is it strong enough?" You have different strength requirements for a chair and a picture frame.

The only dowels I use are short ones to cover countersunk screws.

I have a bunch of dowel pins.  When I was regluing dining chairs regularly, I'd normally replace the old ones with new.  Now, I don't know what I'm going to do with all them.

I think the comment “is it strong enough” is the key. Is it as strong as mortise and tenon, probably not but I think for most applications it is.  To me chairs would be the application where it would be inadequate. I think loose tenon type jointery has largely replaced doweling ( biscuits, dominoes ). Of all of them biscuits are probably on the bottom rung for strength but in most tests they have exceeded the strength of the wood itself and work pretty well, but I mostly use them for orientation during glue ups rather than increasing the joint strength. The dominoes are by far the strongest joint but the cost factor of the machine is definitely a factor. I still use doweling for full size doors. It’s hard to beat a couple of 1” or 1 1/2” dowels. I have made a lot of interior/exterior doors in my day and no failures I know of. 
Paul

Edited September 8, 20223 yr by Masonsailor

Just repaired a table this morning.  It had dowel joints and it looked like there was no glue on one end of the dowels. (this is consistent with other findings)  They were "sort of" reinforced with pocket screws.  But consistent with Chinese furniture, they used drywall screws instead of pocket hole screws.  All but one of the joints had fallen apart because the drywall screws went in too deep and fell thru the joint.  Cr@p like this kept me in business.

10 hours ago, Masonsailor said:

but I mostly use them for orientation during glue ups rather than increasing the joint strength.

 

 

This would be my typical use.  Saves alot of time and fussing around while doing a glue up.

Successful joinery between furniture parts is usually due to a little bit of experience, thoughtful wisdom, and some luck.  I generally only use floating tenons (dowels or dominos) when the boards have been precut leaving no overlap for tenons or anything else.  When I taught furniture design I'd tell the students to make sure they allowed length on their parts for joinery to be made between the connecting parts. 

Dowels or dominos or biscuits between board being edge glued together are useful when the boards all have a little competing bow to help align, but otherwise pointless IMO.   

4D