## 2020-04-15

Categories: Blog post Tags: Privacy in research

I’m working on a project that might be exempt from IRB review…if I can convert a particular ID code to an anonymous code. The ID codes are an eight digit number. Here’s a simple approach that should work for this project.

The steps are

1. Create a random number between 100,000,000 and 900,000,000.

2. Add the same random number to each eight digit id. This will produce a new nine digit id.

3. Link the eight and nine digit ids to the raw data.

4. Store the merged data with the nine digit id, but not the eight digit id.

5. Shred the random number.

The process of shredding the random number will prevent anyone from converting the nine digit id back into an eight digit id.

### 1. Create a random number between 100,000,000 and 900,000,000.

There are several easy ways to create a random number with a specific range. In Microsoft Excel, you would place the formula

=randbetween(100000000, 900000000)

Count your zeros carefully, there should be eight of them for both numbers.

This will produce a value. Mine is 480035489, but yours will be different.

### 2. Add the same random number to each eight digit id. This will produce a new nine digit id.

Here’s some data with some made up id values

x
12345678
12345678
12345678
12345678
23456789
23456789
23456789
34567890
34567890

Here’s what those ID numbers look like when you add my random number

x y
12345678 492381167
12345678 492381167
12345678 492381167
12345678 492381167
23456789 503492278
23456789 503492278
23456789 503492278
34567890 514603379
34567890 514603379