No, they don't, and I don't care who told you otherwise. However, I'll concede that they will happily help you lead yourself down the garden path, if you aren't wise to their tricksy little ways. It's easy to get a little fudge factor here, and a little fudge factor there, until at the end you've got enough fudge to open a candy shop, and a swatch that bears no useful resemblance at all to what's actually going to happen with your knitted item.
Here's the game plan: we'll talk about the common ways in which fudge creeps into your swatching, and then we'll go through a really rigorous approach to swatching that will remove as much of the uncertainty as possible, and then we'll talk about when you really do need to go through the whole rigorous process and when it's safe to do something less (and how much less).
Here are some common places where distortion can creep in:
- The swatch is too small
- Edge stitches are distorting the measurement
- You use needles and/or a method different from that you'll use for the project
- You push or pull on the swatch in the measurement process
- The fabric grows or shrinks when washed
- The fabric grows lengthwise from its own weight when worn
You may notice that the first three of those relate to actually making the swatch, and the last three more to what happens with it after that; things can go astray in either the swatching itself or in the measurement after. I'm sure someone can think of some other way swatching can go wrong, but those are the usual suspects, and if you eliminate those, your chances are pretty darn good of winding up with a finished item that's the size you had intended it to be. So let's get on with how to make a dependable swatch and how to measure it in a reliable way.
Make your swatch big enough. It's not so much that a little-bitty swatch causes distortion as that a bigger one averages it out. Look at it this way: if you make a 1/4-stitch error in measuring across 4" worth of stitches, then in a 40" sweater, that's going to add up to 2.5 stitches of difference, and your chances are fairly good that it won't matter enough to notice (that's about half an inch in a worsted-weight sweater); if, on the other hand, you make a 1/4-stitch error across only 1" worth of stitches, then in a 40" sweater you've got 10 stitches of difference, and that could matter a lot. The bigger your measurable area, the more accurate your results are likely to be.
Give it edges, and get away from them. Speaking of measurable area, it's not a great idea to cast on exactly the number of stitches that are supposed to add up to 4". Go look at the edge stitches on some of your items -- you'll notice that often they're a little off in size from the rest of the piece. If you include those in your measurement, you're likely to be off a bit. Also, remember that stockinette curls. What do you think your chances are of measuring fabric without pulling or pushing on it if you're trying to flatten it down so it doesn't roll up? To address both of these issues, cast on enough stitches to make up the number you want to measure across, plus 3-4 for each side to give you a garter or seed-stitch edging, plus another 2-4 for each side to get away from the edging; work 3-4 rows of garter before starting your stitch pattern, and work a few more rows than you intend to measure across, before ending with 3-4 rows of garter again. And bind off the swatch; measuring with it still on the needles can propagate error quite a way down into the fabric. This gives you a reliable measurement area, and makes the swatch behave itself while you measure. Yes, it takes longer to make a swatch like this. An adult sweater in worsted weight averages around 40,000 to 50,000 stitches; you tell me if skimping on a couple hundred of them at the outset sounds like a smart idea.
Use the needles and the process for your project. It's fairly common to find that your gauge shifts just a little bit when you go from one needle material to another; some of us tighten up when working on slippery metal, for instance, while others relax as they get into a speedy flow. Not everybody will experience this, but until you know that you don't, it's probably wise to at least do the majority of your swatching with the needles you'll use for the majority of your project, and then perhaps a smaller swatch with any other needles you might be using to be sure you don't have a visible shift. It's even more important to make sure you're swatching flat for projects knit flat, and in the round for projects in the round. When you knit stockinette flat, you're alternating rows of knits and rows of purls; when you knit stockinette in the round, you have only rows of knits. If you have a difference between the size of your purls and the size of your knits, and very many people do, you will likely have a gauge difference between the two methods. If you want to see whether you've got a difference, and by how much, make two swatches: do the first with about 20 stitches and 30 rows of garter stitch, all knits, and then do another, the same width and length, but this time purl every stitch of every row. The swatch that's larger will be the stitch that's larger for you, and how much larger will give you some indication of how big the difference can be. Unless there's essentially no difference at all, it's probably a good idea to do two swatches for a project that includes both round and flat knitting (for instance, a sweater that's knit in the round to the underarms, and then split and knit flat for front and back), and make sure you're matching gauge in both methods. You may find it helpful to find the needle size that gives you gauge for the round portion, and then change only the needle for the purl rows for the flat portion, so that you're holding the knit stitches the same size as the round portion and matching the purls to those, rather than matching average gauge against average gauge, which could still give you a visible break.
How do you do a round gauge swatch, anyway? There are several possible methods. One is to just knit a tube (Elizabeth Zimmermann recommended making a hat as a swatch for a sweater), but in order to have a 4" flat measurable area on a tube, you need to make it at least 10" around, so that it can be about 5" when folded and you can measure 4" without getting into either fold, or you need to make it more like 6" and then cut it and lay it out flat, which works fine but does give you issues with reusing the yarn. Another choice is to do a faux in-the-round method, which simulates the knit-every-row aspect of working in the round, but allows you to make a flat swatch. The most common of these is the "giant i-cord" method: using DPNs or a circular, you work a row, and then instead of turning, you slide the stitches to the other end of the needle, loosely strand the working yarn across the back of the piece, and work another row in the same direction. With this method, you need to be sure to leave the back strands loose enough that they don't restrain the piece, and edge stitches are absolutely essential, as the end stitches will definitely have odd tension things going on and you don't want to measure anywhere near them. The method I prefer to use is a variant on this. Again, using DPNs or a circular, you work a row, and then slide the stitches to the other end of the needle; instead of drawing the yarn across the back of the piece, however, you leave a big loop (for a 6" wide swatch, you'll want this to be at least 2 feet of yarn) and shift your hold so you're feeding yarn to the needles from the big loop instead of from the ball; you'll be leaving the strand that goes to the ball attached to the right edge of the fabric. Work this row, using up most of the yarn from the loop, and then slide the stitches to the other end again, and work a row normally using the yarn from the ball which you left at the right edge. When you finish, you'll have a flat swatch with a series of loops on the left edge; if you like, you can tie each of these in an overhand knot snug against the edge of the fabric, to make sure the edge stitches behave themselves. There's a certain temptation with this method to try to finely adjust the amount of yarn in the big loop so you run out as you reach the left edge; resist it. If you do that, you'll necessarily be having to change the way you hold the yarn so it's different from the way you normally hold it, and this is likely to cause more gauge shift than what you're avoiding by working a faux-round swatch in the first place. This method does use a bit more yarn than the giant i-cord method, since the final loops are generally a little bit longer than strands straight across the back would be, but it's a little bit nicer looking as a finished product, and keeps the edge stitches under better control, which is why I like it.
Observe your swatch in its natural habitat. Er, I mean, measure your swatch without distorting it in the process. There are a few schools of thought on how to go about this. First, do you use a measuring tape or a hard ruler? Some people like a tape, because with a ruler it can be a bit tempting to just sort of... nudge things a little. I like the straight edge of a hard ruler, though, and I trust myself to resist that temptation; if you do use a tape, you want to hold it snugly extended, rather than letting it lay on the surface of the fabric where it may ripple a bit. The one thing I don't like is the kind of needle gauge that has a window in it: in the first place, these are usually only 2" long, and we already discussed how smaller areas tend to multiply error; in the second place, it just seems way too easy to play tricks on yourself with these, angling the edge to scoot the fabric in or spread it out just a bit. The next question is whether you mark the stitches to be measured or not, and if so, with what? Some authorities advocate placing pins at the points to be measured; others think that this tends to cause some distortion of its own, and I'm inclined to agree. I like to mark off the measurement area, but I use contrasting sewing thread, loosely basted in place, instead of pins. And then, if you mark, exactly what are you marking? Again, there's two ways to do this: you can mark each edge of the group of stitches that should be 4" wide and tall, and then measure to see if it is, or you can mark the left and bottom edge, and then measure to the right and up 4", mark those, and then count. I like to do the first version when I'm matching someone else's gauge, and the second when I'm designing and trying to determine what my own gauge is. Whichever method you determine, it's important to make sure you don't distort the swatch when you measure it. Don't carefully smooth the swatch out, patting it into shape -- you're not going to have someone following you around smoothing your sweater all day. Instead, pick it up and give it a good shake or three, and then drop it flat on the table and measure it as it naturally falls; if you basted in marking strings, you can pull those straight if they got a little rumpled, but don't fuss with the swatch itself. Put your ruler on it, lining the edge up properly, and then let go, so you're not putting pressure on the stitches, and count. Remember what I said about getting away from your borders -- take your horizontal measurement away from the top and bottom borders, and your vertical measurement away from the side ones. Count partial stitches as well as full ones; quarter stitches can make a difference over an adult-sized sweater, and half stitches definitely do. It's probably not a bad idea, either, to pick the swatch up and shake it a second time, and measure it again; you should be able to get the same numbers twice running. Finally, write the numbers down, on something you're fairly sure you can hang onto.
Give it a good wash. Do you really have to wash your swatch? Well, do you intend to wash your finished item? It does you no good at all to make a sweater that fits you perfectly right off the needles, if it's going to grow or shrink several inches the first time you wash it, and it's a whole lot better to find this out with a swatch that took you an hour to make than with a sweater on which you've proudly lavished several weeks of your free time. Some yarns, of course, are more prone to change than others; cotton, for instance, is likely to shrink, and superwash wools are notorious for growing to an astonishing degree. You want to wash your swatch as you plan to wash your finished garment; it's not that informative to know how the swatch behaves with gentle hand-washing and flat drying, if what you really mean to do with the sweater is toss it in the washer with your jeans and dry it in the dryer on hot. Block your swatch once it's washed, if you mean to block the finished item (if it's intended to be a wash-and-wear item, like a kid's cardigan that you know will never get used if you have to do anything more than washer-to-dryer-to-drawer, then skip straight to the post-washing measurements -- those are the ones you're going to have to live with). With some yarns, especially wool, you can play a few little tricks of your own with gauge in this step, blocking a piece to the measurements you want, but you need to use care so you don't introduce distortion here: block the piece and let it dry or cool, but don't measure it in that state -- pick it up and shake it out well and drop it flat, just as you did at the pre-wash stage, and then measure it. This gets rid of any "bounce back" that's going to occur from tension put into the fabric in the process of blocking it out; the measurement you get after this will be one you can rely on. Write these measurements down too.
Let it all hang out. This step won't be necessary in all situtations, by any means, but sometimes it's a good idea to test out what's going to happen to your swatch under the influence of gravity. Thicker cotton yarns, in particular, are notorious for growing under their own weight, and some of the tape and ribbon yarns are also notably bad about this. Designs with enough ease that they're going to hang free of the body are also more prone to stretching; those that lie close to the body also get some support from it. To check out how gravity is going to affect your swatch, hang it up. (Be sure you did the "native state" measurements first!) Clip the top of it to a hanger, or run a knitting needle through the edge, and let it hang; if it's going to be a particularly heavy item (like a cabled cotton cardigan with an intended baggy fit), you may want to weight the bottom as well -- try running another needle through the bottom edge. See how much vertical difference this makes, after leaving it several hours or overnight, and also check to see if there's any horizontal change (there may or may not be). Again, write the numbers down.
Ta da! Once all the washing and the blocking and the hanging are done, you've got a set of measurements that you can rely on. If your final numbers match the gauge you're supposed to get, and you knit to the pre-wash gauge without variation, then after you wash and block your finished item (and if it had a growth issue, let it hang up overnight), barring pattern or knitting mistakes, it should be exactly the advertised size.
Okay now, really, do you have to do all of that every time you knit anything? Nope. That's just the all-out no-holds-barred method. A lot of times it's safe to do less; the trick is in knowing when, and how much less. First, you'll notice that I've referred a lot of times to the making of adult sweaters, and there's a reason for that. Sweaters are both fairly large in scale and fairly sensitive to fit. That means that there's more scope for small errors to be multiplied, and that it's more likely to matter if they are. For a lot of other items, this isn't going to be the case -- for instance, a blanket is big, but does it really matter if it's 6 inches wider than you meant it to be? A hat is somewhat fit-sensitive, but it's generally only about half as big around as a sweater, and has enough negative ease that it may overwhelm an error. It may be okay there to do a smallish swatch, or one that's knit flat even if you're going to work in the round. On the other hand, I wouldn't skip the washing step even for a small item, if it's one of the fibers that tends to shrink or grow a lot -- I have a friend whose superwash wool tam, intended to be blocked on a dinner plate, grew so dramatically it would have fit her biggest Thanksgiving turkey platter. For baby sweaters, you may want to go ahead and do a sleeve, and wash and block that -- it's not so much larger than a decent-sized swatch anyway, and if it turns out the right size, then you're one sleeve ahead of the game. For socks (small and with a lot of negative ease), I'll readily confess that I do a mini-swatch, maybe 20 stitches worth and only a couple of inches of fabric and I knit it flat, although I do knit back backwards rather than turn and purl, which I know to give me personally a closer match to my round gauge. However -- and this is absolutely the key to successful swatch skimping -- I do not have issues with denial, and I do have a willingness to admit that an item really isn't the right size and to rip right back to the start. If you aren't willing to do that, then you'd better check your gauge properly in the first place. For sweaters, the point at which it becomes apparent that there's a problem is further in than I'm willing to risk (and that's not even counting any post-washing issues), so I really do go through that entire process. Yes, it takes some time and effort -- but if I'm going to have an issue, I'd a lot rather find it out after using half a ball of yarn and a day or two, than 15 balls of yarn and a month or two.
What about those cases where I suggested you might want to do two swatches, such as one with circs and one with DPNs, or one in the round and one flat -- do you need to do the full process with both? I don't think so. Usually in this case, what you're trying to be sure of is that you don't have a visible break or measurable size change between the two methods. Therefore, I do my full swatch process with the method that's going to be used for the majority of the project, and then for the secondary method I do a smaller swatch, and check it for apparent compliance with the size and appearance of the larger swatch; if it matches visually in the pre-wash stage, then I assume that it will behave similarly in the post-wash stage, and so on.
One swatching shortcut that I haven't talked about yet is the multi-size swatch. This is a tool to use when you know that you may need to go through several sizes to zero in on the appropriate needles, especially if you expect a size change in the washing, and you want to get all your swatching over with in one session. To do a multi-size swatch, you'll start with a normal swatch, but then instead of binding off, do a short border section, and then change needle sizes, and do the next section of swatching. Your border section can be the same as the lower border stitch, but I like to do a row of drop-stitching (k, YO,YO, across, then drop the YOs on the next row) -- the loose row thus created makes sure that the smaller or larger stitches of one section don't influence the next section. Repeat this process until you've tried out all the needle sizes you wanted to try, and then bind off normally, and wash and block the whole swatch strip. One place where this technique really comes in handy is when you use a particular yarn a lot, at a variety of gauges; you can prepare a "master swatch" this way, and if you've kept careful notes, when you go to use the yarn in a project, you can just consult your master swatch and know what needle size to use for the project at hand. Some people, when doing a multi-size swatch like this, like to build right into the swatch an indicator of what needle size was used -- for instance, when swatching with size 6 needles, on the first row they might do (k2tog, YO) 6 times, leaving 6 eyelets, to indicate that number. Personally, I just use paper price tags attached after the washing process, but do whatever lets you keep it straight.
There is one other place where gauge goes astray that I haven't talked about, and that's gauge drift in the course of a knitting session. Some people experience this as a regular thing -- they may start off tight with every session, and gradually loosen up over the first 15 minutes or so, for instance. For those people, I'd suggest keeping something really simple to start each knitting session with, such as a garter-stitch dishcloth or a simple afghan square; switch to your real project (and do your real swatching) after you've gotten into the groove. Other people may have gauge drift only occasionally, probably due to emotional factors -- a friend once noted that she could tell exactly which parts of a scarf had been knitted during the chase scene of the movie she was watching, and Elizabeth Zimmermann recommended keeping one tight-gauge project, like a sock, expressly for stressed-out knitting. If this is a big factor for you, you probably want to try to do your swatching when you're not stressed, and maybe even make a second, unwashed swatch, to check your knitting against as you work, to make sure you're not experiencing a drift.
Finally, what do you do with your swatch once you've made and measured it? I suggest keeping it, at least while you're working on the project. Having it on hand can be reassuring, if you're not certain that your fabric is behaving as intended, or if you need to remind yourself that it's going to be fine once you wash it. If you need to scavenge the swatch yarn to finish the project, you can always do that at the end. If you don't need to, then saving the swatch gives you a nice record of the project, and if you do another project with the same yarn, a head start on that project.