The Wilkinson splitter was invented in 1960 by Ernest Wilkinson and formed the basis of high frequency splitter design still used to this day. Its advantages are low loss and high isolation within a limited bandwidth. With proper design, it can handle high levels of power, to hundreds of watts. The resistive splitter, often in a tee or star configuration, offers theoretically infinite bandwidth, at the expense of high signal loss and poor isolation. It cannot handle high power because of the star topology, and is only suited for splitting or combining small signals (typically 0.1 watts or less).

The bandwidth refers to the operating frequency range of the device. Narrowband often refers to an octave or less. Wideband typically means several octaves, and ultrawideband can be 10 or more octaves. This is not a hard rule, and there are tradeoffs of performance vs bandwidth. Typical performance is near-ideal within a narrowband, excellent wideband, and good ultrawideband.

At small signals (typically under 0.1 watt but always consult the datasheet), splitters may be used in reverse direction for combining signals. At higher power levels, care must be considered to the internal design and structure of the combiner. A splitter that was not designed to combine high power signals is subject to damage at high power when used in reverse.

A hybrid is a special type of splitter that introduces a 90-degree phase shift between output terminals. They are often used in amplification to combine two amplifier modules for redundancy. If one of the amplifiers fails, the system can still operate with reduced gain until it can be serviced.