The palm comparison touch test gives only a rough idea; it is not a reliable measure. Meat resistance depends not only on doneness but also on thickness, fat content, cut, whether it has rested, and how cold it was when cooking began. The same feel can therefore correspond to two different temperatures. Only a thermometer gives a definite result. The test is a clue, not a decision.
The method is described this way: touch the thumb to the index finger, then press the fleshy part of the palm beneath the thumb. The softness corresponds to rare meat; touching the middle finger corresponds to medium, the ring finger to a more well-done level, and the pinky finger to well done. The idea is sound because, as meat cooks, its proteins shrink and firm up, increasing its resistance. The problem is not the idea but that this resistance cannot be converted into a single temperature on its own. There is also a subtle drawback: when the test happens to be right several times in a row, it can persuade someone to stop using a thermometer, even though that accuracy is coincidental.
The same resistance can correspond to two different temperatures. A thick cut and a thin cut feel different even at the same internal temperature. Highly marbled meat can feel softer than it really is because of the melting fat. Meat that starts cold can firm up on the outside while remaining raw inside, which can be misleading to the touch. Areas near the bone are always firmer. The test also varies from person to person, because not everyone's palm is equally firm, so the point of comparison changes.
The same is true of color, and the risk of being misled is greater. Pinkness in a cut of meat is not always proportional to doneness: meat that has been frozen and thawed can turn gray earlier, the type of meat used in ground meat affects its color, and poultry can remain pink near the bone even at a safe temperature. So there is only one thing to say: color is not a doneness test; a thermometer is the test, and they are not interchangeable.
This does not mean the touch test can never be used. For an experienced cook, it is a useful preliminary indicator: it gives a rough idea before checking the meat and helps determine when to insert the thermometer. Someone who cooks the same recipe many times can establish their own reference for their pan and meat. But that reference immediately stops applying when used for a different cut, a different thickness, or a different stove. The same is true of cutting the meat open to check it: cutting makes it lose its juices, and the cut cannot be undone.
An instant-read meat thermometer is enough for a reliable reading and does not have to be expensive. Insert the thermometer from the side into the thickest part of the meat, without touching bone or fat. Check two or three spots and use the lowest reading. The target temperatures are clear: whole cuts of red meat 63°C / 145°F and a three-minute rest, ground meat 71°C / 160°F, and poultry 74°C / 165°F. You can also test the thermometer from time to time by checking whether it reads zero in ice water.