With thin cuts of meat (like chicken breast or a steak), you want to make sure you put the thermometer in to the thin side of the meat so that the whole thermometer goes in to the meat, rather than putting it in through the "top" so that only a little of the thermometer is in the meat. If you're putting it in the top, you can get wildly inaccurate temperatures. Also make sure that the thermometer isn't touching a pan, exposed to air, or touching a bone. Any of those can have a negative effect.
Also, you only need to cook chicken to 74C / 165F. That may be part of the problem in your case.
You may also want to check that your thermometer is accurate. You can stick it in boiling water to check that it reads 100C / 212F (assuming of course it goes up that high).
It will measure the temperature of the glass itself, but not necessarily in an accurate way. Most glass is largely opaque to IR, but will radiate IR of it's own based on temperature.
The trick with IR thermometers is that they are guessing the temperature of an object by making an assumption about how much IR (Infrared Radiation) that object will radiate when it is a certain temperature. Cheaper IR thermometers have what is called "fixed emissivity", which means that they always assume that the object you are measuring is radiating 95% of the radiation that it would theoretically emit if it were a perfect black-body object. An emissivity of 95% (or 0.95) is a good start for most household objects - cloth, painted surfaces, etc.
Where it breaks down is with things like shiny metal, glass, liquids, etc. For those objects, the amount of radiation released at a given temperature is more or less than the expected 0.95. Shiny metal pans, for instance, can be as low as 0.2, meaning the displayed temperature will be much lower than the actual temperature.
You can use tables of emissivity values to correct your temperature reading, like here:
http://www.tnp-instruments.com/sitebuildercontent/sitebuilderfiles/Infrared%20Thermometer%20and%20Emissivity.pdf (Note, for this table, use the 8-14 micron column).
That takes away most of the convenience of the thermometer, but it is a useful reference. The closer an object's emissivity is to 0.95, the more accurate your thermometer will be. You can improve accuracy by painting an area flat black, or even using a piece of masking tape stuck on the object to make a good "reading point". Fancier thermometers allow you to set the emissivity, and even fancier ones allow you to auto-calibrate with a contact-thermometer add-on.
To get back to the original question, glass has an emissivity of around 0.75-0.85, meaning the measured temperature will be a little lower than the actual temperature. Of course, even an accurate glass temperature doesn't tell you a lot about the actual oven temperature. My favorite surface for taking the measurement is my pizza stone, which is just about 0.95, perfect for accurate readings.
Best Answer
I use the same hole if what I am checking is large, like a loaf of bread, and I don't want to poke it full of holes...although with my instant read thermometer, the hole is not particularly large.
If the hole is a large proportion to the item (big hole in the side of a cookie...I know, absurd, but you get the idea) where you think heat can run down that tunnel you've made, then be concerned, but I can't imagine any situations where it is likely that the thermometer hole is going to let enough heat in to alter the cooking.
I CAN imagine a situation where a thermometer in place could help transmit heat to the center and make it, potentially, cook quicker. We used to put a large nail in the center of a potato that we were baking in the coals of a fire so that the steel would help transmit heat to the center of the potato to make sure it got done evenly, but I don't know how much of a difference even THAT made.