Question: The non-uniqueness of it is new to me. Can you elaborate, please?

retrieve their records. A well designed database will handle cases where >>a person doesn’t have an SSN, has multiple SSN’s, or has an SSN used by more >>than one person.

[and more on why we should not assume that the SSN is unique]

… which is all well and good, but you failed to answer the original > question. I don’t see how someone can be using “multiple SSN’s” or > “more than two people using the same number” without either being > a crook or running a great risk of getting a number of important > personal records screwed up. Nor can I imagine why someone might > “change” their SSN unless they are a federally protected witness, maybe.

Do you know for a fact that this occurs, or are you just telling us that > we should assume it MIGHT occur? There’s a big difference.

Answer: Yes it occurs. Of course the Social Security Administration has messed up and issued duplicate SSN’s. There is also a sizeable number of people who don’t understand how the SSN works and use their father’s number, or the sample number on some cards that were distributed. But these make up a very small number.

The mistake you’re making is assuming people intended to end up with multiple SSN’s, or that two people intended to use the same number. The fact of the matter is the quality control in many databases is abysmal.

Most of the database systems do *NO* verification or sanity checks on the SSNs. A typo, or a misread number on a form, whamo, duplicate SSN or record. Since databases are matched against other databases, the incorrect data is duplicated, triplicated, and quadupled. If they do any verification, it tends to be of the “First-used” method. Many assumed SSN’s were unique. But since they can’t verify the SSN was entered accurately in the first place, the claim it is an unique identifier is moot.

Why do you assume that because a person’s SSN shows up on multiple records, or a person has multiple SSNs on their record that it is their fault or they have criminal intentions. Do you blame the people in the elevator when it gets stuck between floors, or do you wonder about the quality control and maintenance record of the elevator company. The same things is true about databases. It isn’t always the fault of the people whose records are messed up in the database. The database designer shares a good part of the blame.

The Social Security Administration did a study of bunch of database systems in use in the U.S. a couple of years ago. They looked how SSNs were used in both government and business databases. They found on average 1 out of 10 records had SSN related information that didn’t match the official SSA records. The prison databases were the worst (nearly 1 out of 2 records had an error). The best were the tax collection agency databases (about 2% in error). Not even the IRS depends solely on your SSN to identify your records. If that’s not a hint, I don’t know what more I can say about relying on the SSN as a unique, permenant indentifier.

An annoying thing is the database owner’s usually put the burden on the individual for finding the mistakes and fixing his or her records. People end up with records under multiple SSN’s, and multiple people under the same SSN often not because of any intent to deceive. But because of sloppy database design or sloppy data entry by clerks. The burden of proof is on you to prove the computer wrong. Of course the computer is never wrong. So the logical conclusion if your name shows up with more than one SSN is you are a crook.

There are also a lot of places that use a SSN-like number. For example, Universities use SSNs as student ids. But not all students have SSNs or some prefer not to use their SSN. So the university register will assign a SSN-like number to these students. The problem with these SSN-like numbers is they tend to migrate out of the university setting in to the business databases. Students will use their university id card as identification when cashing a check. A clerk will see the student id number xxx-xx-xxxx and enter it into the check approval computer. The computer “knows” that any number of the form xxx-xx-xxxx must be a SSN and adds it to somebody’s record whose real SSN happens to match.

“Helpful” clerks are the worst. If you leave the spot for SSN blank on an application, some clerks will “helpfully” complete the form. The clerk will find the first Smith, John (or whatever name is close enough) in the database and use the SSN from that record on the form.

For the last two years, the number one consumer complaint to the Federal Trade Commission was messed up credit histories. Misapplying information to the wrong credit file, merging credit files inappropriately, and so on.

So once again, today’s lesson for database designers

- You should never assume a number of the form xxx-xx-xxxx is a SSN – You should never use SSN as the primary identifier for a record – You should never match records solely on SSN – You should never make SSN a “required” field – You should never “fill” in a person’s SSN based on partial information – You should never “make” up a SSN-like id for records without one – You should never use a self-reported SSN as the sole method of identifying a person

Related posts:

1. student loan victims/please reply
2. Unfair student loan credit policies
3. Student Loan Ombudsperson
4. Question of Student Loan Consolidation
5. Private Student Loan If You Have Bad Credit?
6. Student loan default – foreign migration
7. iMac vs. PC for college / laptop vs. desktop
8. Defaulted DSL Student Loan
9. A question about a student loan default
10. President Bush raises interest rates on student loans starting July 1
11. Can you pay a mortgage with student loan money?
12. Diamonds in the Rough