Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

Ask-A-Linguist Message Details

Subject: Second Language Acquisition
Question: Hello. I am a high school student wanting to conduct a simple
linguistic science fair project concerning Second Language
Acquisition. I have become increasingly fascinated with this scientific

Ultimately, the problem of the science project can be simplified into
the following statement:

To determine which age group of monolingual English-speaking
children (6 -12; 13-19) retains the most amount of Spanish nouns in a
time span of five minutes.

I was planning to make a video with visuals, the written word, and the
spoken selection for the individual to watch.

Could you give me some pointers as to what vocabulary words I
should choose? What time span is optimal? How long and how many
children at a time should be tested? How to test the final material to
see how much of it was comprehended? Is Spanish too familiar for
most English-speaking children?

Your insight and advice is greatly appreciated.

Reply: Erick,

Delighted to know that you're fascinated by linguistics. You're discovering that there
are many variables to control for or at least be aware of in making any scientific
claims about language learning. Your science fair statement is pretty complicated: I
have questions about the variables you're using, and methodology in general.

Here's a general idea to think about: Many psycholinguistic studies have found that
recognition is easier than recall, so take that into account when planning how the
children will indicate that they have heard (or seen) that word before.

For example, one concern about the design of your experiment, ages 6-12 seems
too broad to me, because there are several known developmental stages potentially
captured in that group. Age 6-8 might behave quite differently from 11 & 12 year
olds for language experiments. Similarly, 13-19 is too large a span, where you're
likely to get mixed results and not be able to differentiate why. Perhaps choosing
ages 6-8 and then 12-14 (or an even older group, 15-18) as the comparison groups
might be better to sharply distinguish ages/developmental stages, and avoid overlap
in the populations. How many children in each group is enough? How many boys vs
how many girls? (Age and sex are only a couple of several possible variables about
your participants. You also want to control for those who have another language at
home - even if it's not Spanish - or who have been in some formal language learning
situation, like Chinese after-school programs, or Hebrew school, and unless the
whole group has similar exposure, just rule out those with a Ukrainian-speaking
grandmother living with the family, etc.)

Second concern is about a 5-minute time span between learning and testing. How
will you (or your judges) know that the participants remembered the words initially?
How will you distract them in between the learning time and the testing time? (E.g.,
choose an activity that both groups can do without apparently language use, such as
solving puzzles, like mazes or something else visual. But someone else might object
that they can still be rehearsing the audio while solving a visual puzzle.)

Is this all individual training and individual testing? (Or group training and group
testing?) How will people show you that they already learned this word, or didn't
learn it yet? Think about a way to respond that fits both the 6 year olds and the 14
year olds. Changing the response format might challenge the integrity of the

You've suggested presenting the words as multimodal stimuli (oral with visuals and
written). (I'd urge oral/audio over written language, because I don't have confidence
that 6-8 year olds have the literacy skills to do the task. Then you have to decide
whether it's distracting or instructive to provide the written to the youngest
audience.) Think about what hypothesis you want to confirm or deny when selecting
these conditions.

You don't say where you are, so it's hard for me to guess whether Spanish is too
familiar to English-speaking children. If you're located in California, Arizona, New
Mexico or Texas, I might say yes, choose a different language, because you can't
control for prior exposure. If you're located in the Upper MidWest (Minnesota and
surrounding states), I might say your pool of kids is likely not over-exposed in
advance. (But there are Spanish speakers throughout the US these days.)

What words to choose from? The Swadesh 100 word list is one way to go: those are
vocabulary that any language is likely to have, and presumably would be words
known to kids. (E.g., I wouldn't choose a word like "extinguish", where a 6 year old
might not know it, or its Spanish equivalent, where "fire" is a reasonable word in
both languages.) Do you want all nouns? a mix of nouns, verbs, and other parts of
speech? Do you expect different performance on the basis of part of speech? Think
about the number of syllables in the instruction set and the test set. And it seems
like you're not asking for meaning to be learned - this is not an experiment about
translation - just whether specific words have been shared and are shared again. Be
prepared to defend your choices for the judges.

One way to stick with your plan for multimodal stimuli is to record more examples
than you plan to use for the instructional set, but all with the same presentation
style and voice, background etc. Then you can select which ones to use as initial
stimuli (what to learn) and use many of those (plus some of the unused ones) as the
test set. That way there's no difference in the stimuli, they all have the same editing
breaks between words, and you're relying on recognition instead of recall to know if
they remember the words, not the setting or background visuals.

How many words to present is another detail you haven't mentioned yet? I'd advise
making the activity fun, but just a bit too difficult for both groups, so that you don't
get a ceiling effect (too many people approaching 100% accuracy). However, the
younger group has a shorter memory span (probably), and the older group a longer
one. Think about whether there's a way to control for that.

One way to think about all these questions is to anticipate what outcomes might
happen and how you would explain each of them. I assume your hypothesis is that
older kids can recognize more vocabulary items in a new language than younger
kids can, after brief instruction and a pause. If so, why is this an interesting result,
and not obvious? If you don't get that result, how can you explain it?

Or you can just do the experiment with the younger group, and ask "How many
Spanish words can 6-8 year olds recognize after brief exposure and a distraction
activity?" Does this fit the science fair standards?

Reply From: Nancy J. Frishberg      click here to access email
Date: 11-Dec-2012

Back to Most Recent Questions