Is younger really better?
In 1978, Catherine Snow and Marian Hoefnagel-Höhle published an article based on a research project they had carried out in Holland. They had studied the progress of a group of English speakers who were learning Dutch as a second language. What made their research especially valuable was that the learners they were following included children as young as three years old as well as adolescents and adults. Furthermore, a large number of different types of language use and language knowledge were measured and analysed.

Pronunciation was tested by having learners pronounce 80 Dutch words twice: the first time immediately after hearing a native speaker say the word; the second time, a few minutes later, they were asked to say the word represented in a picture, without a model to imitate. Tape recordings of the learners were rated by a native speaker of Dutch on a six-point scale. In an auditory discrimination test, learners saw pictures of four objects. In each group of four there were two whose names were minimal pairs, that is, alike except for one sound (an example in English would be 'shi'' and 'sheep'). Learners heard one of the words and were asked to indicate which picture was named by the word they heard.

Morphology was tested using a procedure which required learners to complete sentences by adding the correct grammatical markers to words which were supplied by the researchers. Again, to take an example from English, learners were asked to complete sentences such as 'Here is one boy. Now there are two of them. There are two _______.'
,
The sentence repetition task required learners to repeat 37 sentences of increasing length and grammatical complexity.
For sentence translation, learners were given 60 sentences to translate from English to Dutch. A point was given for each grammatical structure which was rendered into the correct Dutch equivalent.
In the sentence judgement task, learners were to judge which of two sentences was better. The same content was expressed in both sentences, but one sentence was grammatically correct while the other contained errors.
In the Peabody Picture Vocabulary Test, learners saw four pictures and heard one isolated word. Their task was to indicate which picture corresponds to the word spoken by the tester.

For the story comprehension task, learners heard a story in Dutch and were then asked to retell the story in English or Dutch (according to their preference).
Finally, the storytelling task required learners to tell a story which corresponded to a set of pictures they were given. Rate of delivery of speech mattered more than the expression of content or formal accuracy. The learners were divided into several age groups, but for our discussion we will divide them into just three groups: children (aged 3 to 10), adolescents ( 12 to 15 years), and adults ( 18 to 60 years). The children and adolescents all attended Dutch schools. Some of the adults worked in Dutch work environments, but most of their Dutch colleagues spoke English well. Other adults were parents who did not work outside their homes and thus had somewhat less contact with Dutch than most of the other subjects.
The learners were tested three times, at four- to five-month intervals. They were first tested within six months of their arrival in Holland and within six weeks of their starting school or work in a Dutch language environment.