There are two basic types of runout in a rotary tool; static and dynamic. As their name implies, one, static runout (static TIR), is the result of problems with the physical dimensions of, or arrangement of the components of the tool/collet/spindle system. The other, dynamic runout (dynamic TIR) might also result for dimensional infidelity, but can include other factors such as anisotropic (uneven) material density, worn out spindle bearings, poor collet to spindle coupling, loose bits and spindle motor resonances.
Of the two, dynamic TIR is generally much harder to measure. Luckily, it is also usually smaller than purely static TIR. Static TIR can be easily measured by mounting a bit in a spindle and measuring the concentricity of the mount using a test indicator. It is often a combination of angular (azimuthal) TIR and radial (offset) TIR.
Angular TIR is the result of a misalignment (skew) between the rotational axis of the cutting tool and the central axis of the collet/spindle system. Causes include improper use of set screws in a two-point collet (like the CarveWright adapter), poorly aligned central collet bore, worn spindle taper, and debris between the collet and spindle bore tapers.
Radial runout is the result of a lateral (parallel) offset between the rotational axis of the tool and the central axis of the collet/spindle system. The most common causes are an offset collet bore and mounting a tool with a shank smaller than the minimum diameter of the collet gripping range.
To make life more interesting, measurements of static TIR virtually always reveal a combination of both angular and radial runout. Wouldn't you know it?