If you’re struggling, consider therapy with our sponsor BetterHelp. Click betterhelp.com/scottwalsh for a 10% discount on your first month of therapy with a licensed professional specific to your needs.

You forgot to test the most common joint: 1/8" pressboard tacked with one tiny staple, later repaired with clear tape.

I did some experimenting of my own years ago and found that a simple modification of technique made a massive difference when using white glue on joints where end grain is involved. I found that simply moistening the surfaces with a damp sponge a minute or so before applying the glue drew the glue just a little further into the wood fibres and when the joint was broken it would draw more of the fibres from both sides of the joint, as opposed to the clean break at the glue/wood interface. This was 50 years ago and I have always moistened the surfaces since.

Interesting results with a different testing setup than most. I'd like to see you expand on the theory of glue starved box joints. Compare box joints with varying amounts of clearance to see if a "sloppy" fit is actually stronger than a tight fit.

I have repaired lots of furniture over the years. Never saw a really old box joint that did not fail. Sort of. They look sloppy and loose and the glue turned into grit. I think the gritty flakes of glue tend to hold the joints together. Dovetails loosened up too but the mechanical advantage worked for them. A friend and I worked our way through college building kitchen cabinets and he continued making them until he retired. Made hundreds (Maybe a thousand) of drawers with butt joints and a couple of 1/4" crown staples. Over the fifty years he was in business he never got a call back for a drawer failure. Also important to consider the application and the loading. Drawers get pulled and pushed in one direction over and over. Some joints may suffer shock loading. Drop a box with glued miters and the shock may break the joint but it the miter were reinforced the shock might be dissipated by the reinforcement before the glue failed. Enough questions to keep Suman busy for years.

Fantastic video. It’s clear that you’ve stepped up your filming game and that you also put tons of time into this single video. It pays off. Excellent content and incredibly well presented. Hats off to you!

Excellent video and a great testing machine. Also, your findings were consistent with what I have confirmed in practice, in my own field which is beekeeping. Back 50'ish years ago when I started, bee boxes could be purchased in dovetail, box joint, or rabbet. I assumed dovetail would be the best but my boss who was an old guy and been a beekeeper all his life only used rabbet. A bee box has a hard life. It is exposed to damp then dry then damp, and is outside in the sun. They get loaded up with honey and in a commercial situation are then banged around on a truck and through the honey extraction process. They are expected to give many years service. Through many years of using all types it had become plain that rabbet joints were the hardest wearing of the three options. Another factor, beehives have to be made from untreated timber. The dovetail and box joint corners will rot much faster than a rabbet joint.

With modern drawer slides you can build any joint you want and it will be fine. I can close and open my kitchen drawers with one finger. There is almost no stress on the drawer box because of the operation of the undermount slide. This includes a tray that I built to hold my stand mixer which weighs 40lbs. Very cool test Scott. Some of the drawers in my kitchen are not even actual boxes. They have metal sides with melamine back and bottom screws into them attached to a wood drawer front. It is impressive engineering.

Oh man. So much work! Thank you for the dedication! Definitely one of the best joint test videos in existence

This is actually a very well devised testing setup, and that seems to be especially rare among joint strength tests. My only nitpick is that I wish there had been an equal number of samples for each joint so that half of them could be placed in reversed orientation in the test harness. Doing that would indicate how much the asymmetry of the joint design translates into asymmetry of joint strength and also produce more generalized overall averages. Nevertheless my hat's off to you for really adding something useful to a discussion that often thrives by instead exploiting hidden subjectivity and manufacturing pointless controversy for engagement or self indulgence.

Like you implied, there are a number of difference forces acting on a joint over its life. I’m in the same camp that a box joint seems to be the best due to modern glues. That said, this vid was incredibly informative. I would have loved to see pocket holes just to see what the haters would say.

I need to start thinking outside the box joints.

My son and I made very lightweight (2 lbs all in) "battlebots" for the Midwestern "Critter Crunch" competitions. I used "Reinforced Mitre #2) with aircraft plywood as the splines, and never had a frame get destroyed in all the competitions. Love that technique!

it makes me so happy to someone challenging their preconceptions

Very informative and entertaining. From my memory of my degree course, 50 years ago, failure on wood glue interfaces is complex and is not as simple as looking at the strength of the glue and the strength of the wood, if I recall the thickness of the glue layer is very important. I now wish I hadn't failed that part of the course! A lot of this science was developed in WW2 when they were building wooden airframes in furniture shops e.g. de Haviland Mosquito. From a statistical point of view comparing averages is not a good way to say whether one method is better than another, you need to look at the spread of the results for each test. If there is a big variation from minimum to maximum on a test, in simple terms it could overlap with the range of results of a comparative test and would mean that no statistically significant conclusion can be reached. You also need to know whether the results for a series of tests on a particular joint are what is called normally distributed i.e. equally, randomly spread about each side of the average.

I'm so curious how not as tight dovetails and box joints would compare. Great tests!

Wow, what timing. I just finished making a spline jig for my table saw for mitred picture frames. And used it for the first time not 1 hour ago. Thank you so much for your brilliant joint test video. So glad I used the strongest type of splined mitre joint for my picture frame, because, yes, you need all that advantage in something so structurally important 😂

Thanks Scott. Really good video. Just what I needed. I am a hand tool woodworker and don't have access to things like a Domino or a 45 deg. Dowell jig. So I have been avoiding Long mitre to joints on cabinetry because I worry about their strength and therefore have resorted to the lengthy process of dovetailing. This gives me more confidence to try using nice long miter joints in the future. BTW: I would love to say this same joint strength test tried with 3/4" ply Cheers David

Just found this channel and I really like the content and probably the best delivery to date. I have a medium sized well appointed shop going into retirement and information like this is paramount for success 👍

loved your video. loved the presentation, the humor, the striving for accurate testing methods, and the lack of claim that you know and have proved it all!

Great stuff! I have two comments. Regarding glue-starved joints, even the winning joints could be glue-starved with overly enthusiastic clamping. Second, there are a pair of wonderful books that I learned about some 50 years ago that vastly informed my understanding of such things. They were recommended in The Whole Earth catalog, and a few years later I found paperback copies and grabbed them immediately. Both were by J E Gordon. They are: Structures: Or Why Things Don't Fall Down and The New Science of Strong Materials: Or Why You Don't Fall through the Floor Amazon has them, of course. I guarantee your intuition will be better informed by them.

Mr. Walsh, this is one of your best yet! The tests were great, but the way you told that potentially boring story was so entertaining! Congrats on your growing channel.

Great video Scott. I have a video suggestion for your test rig. It would be cool to see testing done to compare fluted dowels, spiral dowels, regular dowels, and maybe miller dowels too for the fun of it. I'm curious if the friction of a regular dowel causes glue starvation, or if the different types of dowels cause any significant difference in strength in the joint. I've never seen anyone test this before and many woodworkers use dowels in their joinery.

Thanks for doing this work. I’ll refer to it when I’m choosing which joints to use on future projects.

This video was very informative as well as entertaining. You hit a great mix of detail, analysis, presentation.

It’s amazing what old timers like us can still learn! Keep busting the myths! 🤘

One relevancy to glueup times is ... humidity vs. glue. The two most obvious wood glues to use are Titebond II and Titebond III. **These glues work completely differently.** And it's not just that TBIII is "water resistant", hence good for use on outdoor furniture. I live in Florida, typical humidity, even in winter, is probably not less than 50%, and runs up to 70%. Higher at other times of the year... Titebond II takes FOREEEEVER to cure in FL (probably true anywhere along the gulf region). Why? Because Titebond II cures by GIVING OFF MOISTURE. So clearly, if it's in high-humidity locations, it is generally not going to cure fast. I tried to do a joint with TBII and literally 48 hours later I pulled it apart by hand and a lot of the glue was still "thick fluid", not curing at all. In THESE locations, you're better off with Titebond III. Because TBIII uses a different chemistry to cure -- it ABSORBS moisture from the air. Right, the exact opposite from TBII. So it's GREAT in humid locations. Note: Downside, it "foams", so you typically have to do some cleanup after to smooth the area around your joint. But it is what it is. It will actually cure, which TBII might not. On the other hand, if you live in Arizona, you almost certainly want to use TBII and not TBIII.

Nothing more valuable than empirical evidence. Thanks for sharing, and yeah...I still love box joints and my carefully built jig will still get plenty of use, but I'll definitely give the splined mitre joint a chance on a future project.

Very thorough and convincing. I never even knew about the the 2nd reinforcement technique; so I'm making off like a bandit here. Thank you!

I use the splined miters like the last one you tested on all of my picture frames - looks great, good to see that it's actually good. For my boxes/humidors, etc. I actually use a mitered dovetail - do a miter joint, glue it up, then do a dovetail through the miter. Those look awesome as well, and have to assume that it'll make those joints even stronger than the splined miters - would be interesting to test that.

"Crossed the border with a trunk full of joints". The scripting on this is incredible

Comprehensive and fun! Nice work. Someday, if you get around to it, I would like to see the locking rabbet tested.

You put so much time and thought into this. Thank you

When I make drawers for shop furniture, I don't even glue them. I use butt joints and screw them together. If they fail (and none of them have yet in over 20 years), I replace the plywood panel and go on with life. Goal is cheap, quick and effective. For fine furniture a more visually pleasing joint is used. But even on some of my closet organizer drawers I use the same technique and am happy. Those drawers are rarely seen and they work great, cost little and last forever. The whole joint craziness is largely due to UKposts phenomenon and talented woodworkers on YT displaying their skills. That is great and I have hand cut dovetails just to do it and to enjoy the process, but as a joint it is obsolete in a modern drawer. I still love them and I love the box joints and all other classic joinery and woodworking in general. What I don't like is to be made to think that a certain joint is absolutely the best. It all depends and it all evolves with materials, glues, etc. Thanks for the fun video.

Now I NEED to see these tests with joints no self-respecting woodworker (schlups like me) would use: pocket holes, butt joints with drywall screws (in sheer), and butt joints with 16 gauge brad nails (in sheer). I LOVE these kinds of tests!

Long time woodworker and home builder here. LOVE that you do this stuff and your presentation is fluid, clean and entertaining. A few comments on the joints; I have 240 drawers in my shop. Most of them made with butt joints and #8 wood screws (they're shop drawers..) and the early ones made with 1/8" dowels. I have not done any testing like you have, but with that many drawers, some holding a couple of hundred pounds of tools or hardware, I feel I'm in a good position to make some claims about these joints. 1/8" dowels have not failed yet and one of these drawers is only 2.5 inches deep and holds 120# of galvanized bolts. Of the screw joints only one has failed, also with more than 100# load, and it failed because of glue starvation (my opinion) and being slammed too often by a gorilla (AKA new employee). If you ever revisit this, I'd be interested in the 1/8" dowel joint and how strong it is. Seems it might be stranger due to less material being compromised...?

Wow! Excellent video. Some of the results make sense to me. I have always thought that the miter joint was stronger than popular belief. But not as much as you have demonstrated. Normally a few other videographers put out rebuttal videos. That will just add more data to be reviewed. I think your tests are accurate and make perfect sense to me.

Hey Scott, nice work! In your final test, where the splines slip out of the joint without breaking - it's my belief this is caused by glue starvation. Hot hide glue is the solution that works for me. I make hardwood art frames. Over 2000 frames (8000 joints) in the last few years. 3/8 x 1.25 mouldings, where each miter only has 0.675" of surface area for the glue. I use two 1/8" thick hardwood splines across each joint, placed 3/16" in from each edge. The miters use Titebond I, which seems fine. But when I used Titebond on the splines, they failed by slipping out, not by breaking, just as you found. The majority of the frames have been made using 315 gram strength hot hide glue for the splines. While I can break a Titebond frame with my bare hands, these are nearly impossible to break either by pulling, levering, or twisting. When they do fail (when I use tools to break them), the splines are cracked or pulled apart with the grain. Usually there's other damage to the mouldings / miters as well. The difference from my perspective is night and day. Much, much stronger. Without even knowing if one glue is inherently stronger than the other, I believe hot hide works so well because the warm water in the glue swells the wood components, making the joint immediately tighter. Then the glue dries, and the surfaces can't come apart. It's telling that I need to make the splines about .003" thinner than the slot. With Titebond, I was able to make the splines the same as the slot. If I left slop to avoid pushing out the glue, I'd have an unsightly glue line outlining each spline. With hide, if I make the spline as thick as the slot, I can't push it in fast enough to bottom out before it begins grabbing. The glue doesn't set up quite that fast. That's the wood swelliing. There's no visible glue line because ... there just isn't. Hide glue dries clear, and accepts finishes. If you ever re-do this test, it'd be great to see an objective measurement of what I see anecdotally day after day.

Nice summary of joint options. I think that they don't matter if you're making drawers, which don't get stressed much, but it seems likely that other types of boxes -- for example, the box used to make a large 8'x10' book shelf -- might be stressed more than your typical drawer-box would. So I think the vid, and your tests, does provide real value.

Crude as the testing is, the info remains very useful. Thank you. I will stick to using the humble rabbet for most of my joinery needs.

I just found this channel the other day, thought it was interesting so I subscribed. Glad I did. I love stuff like this.

That was really interesting so thanks for sharing your results. I must admit I didn’t think the miter would do so well. Maybe the lock miter joint cut with a shaper/router would be stronger although the setup is such a pain. Cheers

Would have loved to see how the mitre lock joint holds up given the newfound knowledge of mitre joint strength and glue ups!

Snoop Dogg: I was expecting a different type of joint testing...😂

Thanks for making this. I think this also lines up with the results Bourbon Moth had as well in his tests, which is good to see two different tests corroborate similar outcomes.

Conclusion: make the joints how you want them to look 👍 Great video

Interesting test and valuable despite the very limited amount of samples. I'll definitely be using more rebates and miters! As a point of interest; discarding your top and bottom samples as "outliers" isn't a good idea. Statistically speaking, outliers can be calculated from quartiles (Q1, Q2/Median, Q3) and the interquartile distance (IQ). Top outliers would be >Q3+1.5IQ and bottom outliers

That was a very informative video, Scott. Thanks, man ! Butt joints it is. Think of all the time I'll save not having to watch dovetail, box joint and miter videos that I can now spend in the workshop. Or spend watching your other videos while the glue dries.

Just came across your channel. great content! The splash of humor is the added touch that makes your stuff really great! Thanks

Could have just asked me, Granpa told me that back in 76 i think. He was a cabinet maker/carpenter and at the end of his working life would trim out track homes. He mitered all the sute built cabinets and drawer box fronts using carpenters glue and nails. Ive been in many of those homes and his work was still going strong 45 years later.

I think, as you mentioned, that the main call for dovetails in drawers is not brute strength. it's long term resistance to wiggle and thousands of tugs.

What an excellent set of tests. I never thought much of a butt joint until I saw your tests. Never would have guessed they are so good. Thanks.

This is some fine content. Very nice work! It was so good that I liked and subscribed too. I really appreciate the thoughtful and thorough test scenario you created. Thanks!

Very interesting results, Scott. Potentially a dull subject, but as I sat here at 3:20 am, feeding my 5 week old foster baby, it was more than interesting enough to keep me awake at this ungodly hour. Loved the content, and it shall be very useful in my shop remodelling.👍

This test was awesome. Many thanks for sharing

Thanks great video. I always use mitred joint with dowels at an angle if I can.

100% of the time, I'm skipping the ad. But that cute doggo convinced me to sit through it. Nicely done.

You dont need better help if you have a doggo like that❤

Excellent video, thanks for posting. My typical joints (just dowels) are strong enough, so no need to change anything. I do like the look of box joints and splined mitered joints, so I still use those occasionally too. Dovetails are all but extinct in my shop.

Cool video! Just due to time I will probably stick with miters from here on out, though for the extra time I really love the *look* of the second reinforced miter. If I'm bored that might be my new kitchen cabinets in a few months

Ok - statistically there is no good reason to throw away the top and bottom values. They are important contributors to the estimation of your standard deviation. If one is miles away from the others then you may consider removing it but you do have to be really careful about biasing your results.

I would think that if you pulled on a drawer more than 754 lbs you should stop going to the gym and start working more on your brain and start thinking about removing something from your drawer.

You really put in the work for comparative and measured testing. This is very valuable. Thanks

Awesome video bro, I really enjoy these! Can't wait to see more!

He talks a lot without saying much and takes forever to get to the point.

Extremely interesting. Really. Thank you

Nice tests, even better narration, very entertaining. Thank you fellow canadian!

Fascinating test outcomes. Thank you!

spared me a lot of time and money, will just use box and simple miter joints for everything - thanks!

Haha I’ve been making most of everything with miters and I had some older woodworkers busting my chops about it. Thank you for the testing video and I’ll send it to him 😂

Amazing video, thanx for the tests

Fantastic test! Very eye opening. I would be interested to see how a 1/4-1/4-1/4 joint holds up, as well as a stepped dowel (Miller dowel). But this test has convinced me to not spend the $ (and time) on a dovetail jig for making my shop furniture.

My biggest take away was that no matter which joint I use, it will never be under this much stress. LoL! Great video and I thoroughly enjoyed it. Great data. Thank you for sharing. A plywood box I made when I was a teen, 40+ years ago, with butt joints, glue, and brad nails is still holding together after all these years. Mostly. Wood movement, the plywood warped and bowed on one side, caused the most likely glue starved joint to fail. However the nails kept it together. I merely squeezed more glue into that joint and clamped the crap out of it for 4 days. It will probably bow again in the future, but I'll probably be dead and IF any of my kids keep it, they will be the one gluing it back together again. LoL

Singer sewing machine bases around the 1950 had a simple joint with one small 4mm wood peg on each corner + a corner block glued inside for the machine to sit on. I have just had one in the oven to kill woodworm and make the glue fail, it is now ready for renovation. These cheaper wood bases usually had canvas / cotton material wrapped around that hid the joints. They cope with quite a lot of stress from a cast iron sewing machine even when the glue starts to give up with age. Many thanks for showing the strength of a Rabbet. I think the peg in the sewing machine base was for alignment while the glue set.

I really liked the editing in this. Nicely done.

Great video! Thanks!

doing the work. I will remember this, thank you

Very interesting tests. One thing no one pointed to I think is that "reinforced miter 2" is also more repetible, they fail at a more regular force and look therefore more predictable.

Oh, congrats on the new cute puppy. Just adorable.

Great video, thanks you for making and posting it.

Thanks so much for this impressive video! Used last method many years as the most simple to do, but did not realize the it the most durable in addition! I used plywood for inserts, by the way 😊

I think I have seen others do similar tests on UKposts, though off the top of my head I can't think of one. I'd be curious to know if you modeled your test on someone else's and if you found your results consistent with what others have found or not. Glad you posted the spreadsheet in the video. I paused it there to look at all the data points and examine the deviation. It's a lot of work to make 5 samples of each joint for testing, so kudos to you.

Very entertaining. Thank you.

Thank you for the very helpful tests.

Amazing amount of work went in to this! Even more strength to be had when you glue your plywood drawer bottom in.. plywood glue, solid panel don't glue

Would have been great to see a locking rabbet in this comparison as the 1/2 1/2 1/2 or 1/4 1/4 1/4 method is a pretty common and very effecient way to make drawers as you can do all of the jointery from a single table saw setup.

Great video! Thanks for all the effort 🤘🏻🤘🏻

That was really interesting to watch. Thank you. I would love to see a box joint with a looser fit to see the effect of scraped glue.

Nicely done! I realize there are many variables to consider in the evaluation of a joint but one that appears in the video is the amount of glue used was not necessarily consistent or allowed for visual evaluation of a minor amount of glue squeeze out.

This was such an informative vlog. So glad I watched it. Many thanks for posting.

Great job!!

As a picture framer of 40 plus years....45 degree mitre joints while being fast, are also extremely strong. We used to pre drill and put in a finishing nail on opposite sides. Some of the high end frame mfrs sold metal clips pounded in from the bottom. None are necessary. Suggest you use a light coating of glue...wait a couple minutes then glue as normal. Square clamping is essential. After that they are stable. Movement of a joint like drawers, requires the right pull push joint. If you look at cabinet doors that are 45 mitres...they last forever without any other reinforcement. Try to refurbish a 150 year old piece and study the joinery. You'll learn.

Outstanding work

These are some interesting results I never would’ve predicted!

Dang, this video definitely went against expectations. I need to watch all the other good wood failure tests of using forces in different directions to help get a sense of what's actually sensible. Now of only someone could make a test about the right tolerances with digital calipers or even a micrometer for testing purposes and in a similar vein just purely about what clamping force (pressure, not just force but per area) is too little or too much. Also some kind of accelerated longevity test for joint types, just taking your wood outside to gather moisture and inside to dry up again, maybe in a monthly basis. I know I'm not in a place right now to be able to do that and film it as well, but please, someone, take one for the team and advance woodworking for us all :D

THAT was a VERY well done, and SUPER interesting take on wood joints. Great Stuff.!.!.!

Ive got a whole covered patio with glass and doors to make and i plan to use mostly rabbits.....wow looks like its doing well too!

Very interesting, educating and convincing! Thank You so much for such extensive experiment so concisely presented! I always intuitively felt that miter with some reinforcement might be the strongest joint, now we have the proof!

Love your comment about the time it takes to set up the stupid dovetail jig. I've got one, works awesome, but it takes me forever to get it actually set up correctly; most of the time it's just as easy to make the couple of joints I need to make by hand.

Good job on finding the last miter joint with horizontal reinforcements! Perhaps another joint for you to test would be what I call a double rabbit joint. Not sure if it exists already, but both corner boards are cut like the original rabbit joint, but at half as much distance on each. Next joint to try is a double rabbit joint with horizontal reinforcements. The third joint to try is super complex, a double rabbit with dove tail horizontal reinforcements (to exploit that mechanical advantage you were talking about). If you have the time to manufacture and test these experimental joints, we may find an even stronger joint than the miter with horizontal reinforcements and I'd love to see the video and am sure others would too.