A/B EXPLORATORY PROTOTYPING
Topic
Serial position effect: the tendency of a person to recall the first and last items in a series best, and the middle items worst.
Hypothesis
If people will read a series of words they are most likely to remember the first and last words.
Testing protocol
15 seconds to read a list of 10 countries followed by another 15 seconds to write down the countries they remember, without following any specific order.
Prototype
From the left test A and test B
Variable
In prototype A the list starts with “Singapore” and ends with “Austria”. In prototype B the list starts with “Japan” and ends with “Singapore” which has been moved from first to last place.
Findings
The test was conducted on two groups of 5 people. The first group was subjected to test A while the second group to test B. To compare the results obtained I gave a score to each word, 0 if the word was not remembered, 1 if the word was the tenth word written by the tester and so on up to 10 if the word was remembered first. The maximum score obtainable is 50, that is, the country was written in the first position by all 5 testers in the group. The graphs obtained show that the first country in the list is the one that is remembered the most, but this does not happen with the last one in the list. The Netherlands in both cases scored high, probably due to the fact that the test was conducted in this country, showing that there is a bias. Some testers stated that they remembered the countries they visited best, one tester wrote Denmark in the second position because she was going to visit it the following week. Also note that some countries were remembered incorrectly as South Korea, which was written as Saudi Arabia or North Korea, or Austria which was confused with Australia. The user who wrote Saudi Arabia said she confused the name because of the formula 1 race that took place in Saudi Arabia the day before the test.
Reflection
The test partially confirmed the initial hypothesis showing that the most remembered word in a list is the first one. What I didn't foresee is that it would have been more appropriate to choose a list with different content since a list of countries involves biases from the user, compromising the final result. I would redo the test using names with which the tester cannot have any personal connection.
