A Thinking Aloud or Think Aloud test is generally an integral part of the user experience. During such tests, users of a website or product are encouraged to think aloud when using individual functions. These expressions, mostly unconscious, are registered and evaluated. Thinking Aloud tests are very popular because they can be implemented with little financial and logistical effort. In most cases, these tests are still in the beta stage of a software, website or product.
In the early 1980s, psychology dealt with the complex internal processes that occur in human beings when they perform a certain action. This ushered in the era of cognitivism, which deals intensively with human thought. A primary question of research in this area was how to access the internal processes of a person's brain.
One of the proven methods for this was in short found in "Thinking Aloud". People are supposed to be able to provide accurate guidance on their actions if they talk about it out loud throughout the buying and selling process. For example, respondents were asked to express their thoughts when solving math puzzles or tasks.
However, this method was quickly criticized because many researchers thought that the data collected using the Thinking-Aloud method was too subjective to be able to verify or falsify hypotheses. Finally, with Ericsson and Simons, two researchers took up the subject with the work "Verbal report as data" in 1984, a kind of basic work. In this scientific work, three levels of verbalization of actions are defined:
- Level 1: simple representation, for example, through sequences of numbers or enumerations.
- Level 2: Non-linguistic skills are verbalized, for example, with illustrations or photos.
- Level 3: Actions are justified
The researchers in summary assigned the first two stages to short-term memory. Consequently, they are also relevant to testing, whereas level 3 verbalization is derived from long-term memory and is thus subjective.
Another important step towards Thinking Aloud Testing in practice was Clayton Lewis's paper "Using the Thinking Aloud method in cognitive interface design," which he published in 1982 as an IBM employee. Describes for the first time how the Thinking-Aloud method was used for testing. In the following years many other publications were created on the same subject.
Right now, Thinking Aloud tests are used by industry, advertisers, and product developers alike. Unlike Ericsson and Simons, test results now start at Level 3. Because these are typically exactly the tracks that UX designers want to hear from their test people.
Compared to other methods such as eye tracking, a Thinking Aloud test it does not require a huge technical team. For this reason, this test operation can be used relatively quickly.
First, it describes the special action scenarios that are expected of an average user when it consists of a product or web portal. Test users are assigned tasks to perform. At the same time, you must say out loud what they think when they use, what questions they have and what they do not understand. All account statements are recorded and evaluated afterwards.
Jakob Nielsen, CEO of the group of the same name, reduces the procedure of a Thinking Aloud test to three steps:
«Recruit representative users.
Give them representative tasks to do.
Shut up and let the users do the talking.
Using the method
With the help of a Thinking Aloud test, advertisers and product developers can almost follow "live" how a potential customer interacts with a product. This is because the considerations for each individual action step, for each new web view or product feature are provided directly and can be compared to the expected actions. In this way, software programs, websites or consumer products can be designed with greater precision, not by intuitively anticipating customer actions, but by predicting them quite accurately with the help of empirical tests. The larger the data set analyzed in a Thinking Aloud test, the greater the likelihood that user actions will occur.
The Thinking Aloud tests have the following advantages:
- Low cost process: no experts are needed and no need to buy expensive equipment.
- Only a few topics are needed to get relevant results.
- Customers receive spontaneous and authentic reactions from users.
- Misinterpretations can be practically ruled out, as test users do not have the opportunity to re-review their comments.
- Simple combination with other user experience related test procedures.
- Flexible apps.
- The tests are never authentic application situations, but mostly artificial laboratory conditions.
- There are delays in the app, because users have to express their thoughts out loud.
- Small samples are often not enough to transfer large quantities.
- Thinking out loud can influence the actions of the people being tested.
- Investigating more complex action steps requires a lot of attention.