# Random assignment

#### Sam12

##### New Member
In order to compare one dyad against two individuals, I wanted to assign two individuals that are randomly selected to notional dyads. So, I wondered what would constitute “randomly selected.” More specifically, my question is, is my arbitrary decision as the researcher to – let’s say – assign the first two participants in the individual condition the notional dyad 1, the 3rd and 4th participants the notional dyad 2, and so forth, “random enough”?

If so, are there any scientific papers I could cite for justifying the approach?

#### hlsmith

##### Not a robit
I wouldn't do it based on your own brain. Even if you don't have access to software you should be able to create a reproducible method. I would say, create a sequence of numbers in MS Excel (e.g., 1-4) in column A. Then in the column next to it (B), use say the normal distribution to simulate 4 random draws from the normal distribution. Then your rule will be that the two lowest random numbers in column represent those assigned to the group of interest.

This is the same concept as using you had, but is less likely to be scrutinized if you present results.

P.S., One note, you will need to set the seed value when randomly pulling from the normal distribution to ensure you get the same results every time you try to reproduce the process. I would image there are examples of this on the world wide web.

Let us know if you have questions.

#### ondansetron

##### TS Contributor
They have tables for random numbers if you don’t have excel. I just don’t see any reason for using the original method. Even a coin flip would be better.