The idea is always to have individual experiments, so when creating experiments, we don’t have the intention to compare them with others.

So let’s say we have the following experiments:

• L1 – Learning with Langoid 1 month 1 hour per day
• D1- Learning with Duolingo 1 month 1 hour per day
• B1 – Learning with Babel 1 month 1 hour per day
• L2 – Learning with Langoid 2 months 2 hours per day
• D2 – Learning with Duolingo for 6 months, 2 hours per day

One person

If we have one person per all 5 experiments, we can compare L1, D1, and B1 with each other.

L2 and D2 can’t be compared, but we can publish results like Hani did in her experiments, so we can say that someone reached the A2 level using the Duolingo D2 experiment.

More persons

If 5 persons done L1, 10 persons D1 and 1 person B1. We can still compare them the same way as when just one person was learning, but now we can say the average result for L1 and D1.

With this approach, every experiment is a group experiment. Still, if we don’t have enough people to repeat the same experiment with multiple people, we can publish them as individual experiments.

Summary of what we are getting

• We have the results of individual experiments, and that is the primary goal we are trying to reach
• If we have multiple Individuals spending the same amount of time and having similar starting point (people knows nothing, for example), we can compare different strategies.
• When we get multiple individuals doing the same experiment, we have more confidence in the experiment results.
• When we get multiple individuals doing different experiments, we can compare them with each other more confidently.