How do you measure the temperature of a Globe/Sphere ?

Temperature is a funny choice of observation anyway. I am not sure what it tells you.

When you start using temperature anomalies it tells you even less and when you start taking averages of averages it tells you even less than that.

A better thing to measure would be heat.

Mathematically, if the global average surface temperature anomaly increases then the radiation to space should increase but it could decrease as well. You can't tell because of the fourth power in the calculation.

If the air in one place has a higher specific heat than another then the same amount of heat will result in different temperature increases in the two places.

If you have a class of 20 people and 19 get 100% and 1 gets 0% then (max-min)/2 =50%

If 19 get 0% and 1 gets 100% then the answer is still 50%.

Having said all that, I think they use an area derived from the latitude and longitude lines. So the area used for the average varies over the surface of the globe. If they use a 5° grid then the area at the equator will be about:

5/360*24,000 miles square. That is, about 111,000 square miles. This will reduce to less than 5,000 square miles at the poles.

Depends on the nature of the globe/sphere. If it is uniform, then taking a measurement at one place is sufficient.

For the planet, you generally take measurements at different locations, split the planet into gridcells, try to get a temperature for each gridcell based on the measurements taken, then average over all gridcells.

Wet-bulb globe temperature. Indoors, or when solar radiation is negligible, the following formula is often used: The WBGT index was developed in 1956 by the United States Marine Corps at Parris Island to reduce heat stress injuries in recruits; it has been revised several times.

That depends on the composition and size of the sphere.