Glass thermometers are usually (unless otherwise specified) adjusted for full immersion. This means that they indicate the correct temperature only when the whole thermometer is exposed to this temperature. That is, including the scale. In stills, however, usually only a small part of the thermometer is exposed to the hot vapor. As a result, the thermometer then indicates a temperature that is too low. This is called the "stem error" because the liquid in the capillary is also called "stem".
How large this error is depends largely on the material of the stem. And this is one of the reasons why high-quality laboratory glass thermometers always have a mercury or gallium filling. These materials expand much less depending on the temperature than, for example, petroleum (see the listed material constants). This means that the thermometer needs a much larger quantity of liquid, i.e. a much larger storage vessel at the bottom, but this then has the great advantage that a larger proportion of the liquid is placed in the vapor and a smaller proportion outside the still. So the stem error turns out to be much smaller.
Example: A mercury thermometer indicating 90°C, immersed in steam up to the scale mark at 0°C, and whose emergnt stem has an average of 40°C, indicates 0.8°C too low. But if it is a thermometer filled with petroleum, the error is 4.1°C, i.e. about five times as much.
An estimate of the average temperature of the emergent stem can be obtained with a second thermometer in the middle of the emergent stem. But there are also other, much more elaborate methods with so-called "auxiliary stems" or with so-called "stem thermometers". More information is available in this pdf:
"The Correction for Emergent Stem of the Mercurial Thermometer" von E. Buckingham 1911
There are also thermometers that are adjusted for a specific immersion depth. They are often sold together with a laboratory glass distilling bridge and a silicone stopper. With these, however, the laws of physics are of course not suspended. The scale has simply been changed somewhat. However, since the average stem temperature is very dependent on the outside temperature, and this can vary greatly depending on the room temperature and also on the insulation of the still, this is not very accurate. If it is then as usual a petroleum thermometer, it makes a lot of difference.
And most importantly, then it is also different for each distillation. For example, if the outside temperature has changed. Or even if the setup of the distillery changes: In a potstill, the emerging part of the steam thermometer is near the hot boiler, so it's very warm; in a reflux still, it's not. So it is no longer possible to compare current readings with older ones very well.
That is why we recommend using electronic thermometers.