This will go against most conventional wisdom on the Internet, but here we go...
If this is a modern rotating disk, a simple pass of dd with /dev/zero is enough to foil almost any attempt at data recovery, even from a professional data recovery house. It might be possible to extract some data with extremely expensive specialized equipment (e.g. a government lab), but that is out of reach of pretty much anyone that isn't willing to spend $millions on you. (Note this will not comply with any official-sounding government standards for data disposal, but it works.)
The problem with most of the wisdom you read on the Internet about this topic, is that it is more urban legend and actual fact. If you look for an actual source on this topic, most people refers back to a paper that was published in 1996, and was referring to MFM/RLL drives (pre-IDE). Additionally, most of the government standards on these topics are decades old as well.
The logic behind multiple passes to erase data boils down to the idea that residual information can linger in the space between sectors on a platter. On older drives, the density of sectors was relatively low, and there was lots of empty space on the platters where this residual data could linger. Since 1996, hard drive capacities have increased by orders of magnitude, while platter size has remained the same. There simply is not that much empty space in a platter for data to linger anymore. If there was usable extra space in the platters, drive manufactures would be using it and selling you a higher-capacity disk.
The wisdom of these secure erase standards has been picked apart, and papers have been published that say a single pass is enough.
A few years ago, someone issued the Great Zero Challenge, where someone overwrote a drive with dd and /dev/zero, and issued an open challenge for someone to extract the data. There were no takers as I recall. (Disclaimer: The original web site for this challenge is gone now.)
But what about Solid State Drives? Because of the flash wear leveling, bad sector remapping, and garbage collection, traditional overwrite methods may not actually overwrite the data (although it will appear overwritten to the host PC). A single pass of dd with /dev/zero will stop a casual user from reading back any data. However, a dedicated attacker with a logic analyzer can crack open the drive and extract data from the flash chips inside.
This problem was identified a while ago. So, a standard command called Secure Erase was added to the ATA standard. The firmware in the drive will securely erase all of the flash cells. Any modern SSD will support this command. I beleive it also works with rotating drives. Note that this command can sometimes be tricky for an end user to access. You typically need a special utility to use it, some BIOSes implement a "security freeze" that can get in the way.
What you should take away from this:
- If you have a rotating drive, just use dd with /dev/zero, or the single pass option in DBAN.
- If you really want, you can use a multi-pass erasure method. It won't hurt anything, but it will take longer.
- If you have an SSD, you should use the ATA Secure Erase Command.
- If you are required to erase the drive to a certain standard (e.g. a contract says the data shall be erased per DoD 5220.22-M), just do it.
- If the data is so sensitive that its value exceeds the cash value of the drive itself, just physically destroy it (be creative). If you are really paranoid, make sure the remains of the drive are scattered over a wide area (e.g. multiple public trash cans in multiple parts of the city).