Historically, individuals knew each other. The family unit knew whether they could loan money to Uncle Fred or if he was a drunkard who would squander it all. Church members knew if a farmer had been down sick and needed help getting their crops in at harvest time, or if the farmer had just been too lazy to do the work.

In the current era of people wanting to borrow money from everyone else they have just met (in the form of store credit cards, general credit cards, auto loans, and mortgages), we need a measure of trustworthiness that has some history to it, even when our own personal relationship is brand new.

Enter the credit history, and more specifically a credit score. A credit history is the details of payment from the individual; but since it still takes time to review a credit history, and interpret it, the credit industry has boiled that down to a score that they calculate with one of innumerable proprietary and every changing algorithms. Even one number is often too complex for most people, they want that all rolled into whether the score is above or below a certain threshold.

Unfortunately, as a biostatistician, I can tell you that reducing numerical complexity this way may be potentially useful in the aggregate, but certainly not always, and definitely not on the scale of the individual.

Example –

Having a speedometer on my car is too complex (just like thermometers and even tachometers have been removed from many cars these days). Instead, I substitute the speedometer with a yellow light and a green light (avoiding red/green for the sake of my color-blind husband who might also drive the car sometimes). If the yellow light is on, I’m speeding. If the green light is on, I’m not.

Sounds good, right? Let’s see how this plays out in reality.

• I’m a conservative driver, with lots of experience using a speedometer. I work hard to not be a slow poke on the road, and to keep the light green, balancing going fast enough to keep up with traffic while still staying a non-speeder.
• My (now deceased) grandma was losing her cognitive abilities before she got moved into long term care. If she borrowed my car, at best she would have focused on keeping that light green, probably by the expedient of driving 5 miles per hour, no matter what the weather or surroundings.
• My children, when old enough to learn to drive, won’t have any experience with a speedometer. They will have no idea how fast they are going, only if the light is green or yellow. They won’t be a good judge of whether they are going 26 mph in the 25 mph school zone, or 60 mph in that same school zone, they will only know that the light is yellow.

.

Now let’s look at what happens when data are taken out of context. In today’s era of data, once data are collected, they can be used in all sorts of out of context ways. For example:

• A credit score is about how good you have done about repaying money in the past, it has nothing to do with your savings level or ability to earn money to pay future bills.
• Medical data are collected to treat patients. They are also used for medical research (with HIPAA forms signed in some places), to look for biostatistical relationships. That data is also used (without any patient consent) for looking at insurance reimbursement and determining future pricing.
• If you collect data on patients with Condition A, analyzing them to kingdom come won’t tell you anything about patients with Condition B. Or even about patients without condition A.

Credit scores are being used in all sorts of out of context ways now. Some examples include:

• As an employment criteria for large employers.
• As a criteria for eligibility for renting an apartment.
• As a criteria for determining your auto insurance premiums.
• As a criteria for eligibility for renting a car.

There is other trickiness involved in your credit score calculation. If you happen to believe in removing all your credit history, be aware that in the world of credit scores, you are now not having a missing value, you have a value of zero. To understand how bad a zero can be instead of a missing value, think about recording a blood pressure reading. If I come to the doctor’s office and they forget to take my blood pressure one visit, that’s not alarming; if the computer fills that in as a zero, suddenly all sorts of alarms may be triggered. Or if we’re calculating my average blood pressure over time, skipping just a few readings (filling them in as zeros) could take a hypertensive patient down into appearing to be a normotensive one.

New uses for your credit history are being developed constantly. In England, lack of a sufficient credit history is keeping people from being able to get access to at-home COVID19 testing.

It’s worthwhile to understand what goes into your credit history, and in general how that information is used to calculate your credit score.