A parity bit is used for error checking against an array of data. One byte is made up of eight bits of data, seven that actually contain data and the parity bit. The bit helps ensure that the array is always either odd or even, one or zero. It allows transmitted data to be checked for errors by allowing the receiving computer to make sure a group of bits is either odd or even when the data arrives.
When data is transmitted, the binary digits sent must match the length of bits that the recipient is expecting. If the data expected is even and the data sent with its parity bit is odd, the recipient can reject the data because of an error in parity.
A parity bit helps ensure that each array of data has an even number of ones, for even parity or an odd number of ones for odd parity. Adding the extra bit to the data packet will allow the data to quickly be checked for errors when received. The extra bit acts as an error detecting code or mechanism so unexpected data is not received.
For example, suppose that a document is being transmitted for review and the document is contained within one data packet. The data packet with its parity bit is odd. The packet is transmitted to another computer somewhere on the Internet and the remote server is expecting the data packet to have an odd parity. When the data arrives, the host machine receives a packet with even parity. The host will reject the packet because the parity of the data sent does not match and must be in error.
If a data packet is received and it is out of parity, having the opposite parity bit of what is expected by the receiving computer, the recipient will send a message denoting the transmission error to the host that it did not receive the data as intended. Many times the original sender of the data packet will resend the transmission after recalculating the parity of the data which can sometimes correct the problem.
The parity bit error checking methods are far from perfect. The parity bit error checking scenario is only able to correctly report an odd number of bit errors. If an even number of transmitted bits have errors, the transmission will be reported as correct even though the data within the transmission is corrupt.