Bodong Chen

Crisscross Landscapes

Towards A Longitudinal Understanding of the Growth of Productive Vocabulary in Knowledge Building


“My theory is that it’s all about gravity. The groundwater stays down by gravity!” (Grade one, Water unit)

“Because the plane is so big and heavy it is hard to reduce gravity because gravity can be taken away better if there are lighter, stronger materials.” (Grade two, Flight unit)

“How does gravity work? Is it a force in the ground that pulls you down or something in the air that pushes you down?” “I think gravity comes from the core of the Earth or the core of other planets.” (Grade 56, Astronomy unit)

Students write, and LOVE to write, in collaborative knowledge building. These excerpts about gravity above are from writing of a student cohort when they progressed from Grade 1 to 6. They appear to be naturally engaged with vocabularies – as well as their related scientific concepts – when striving to explain all sorts of phenomena they were trying to understand. However naive their understanding was, they were able to incorporate vocabularies productively (for communicating ideas) in their writing.

Explicit vocabulary instruction, which emphasizes teaching new vocabulary through dictionary definitions and spelling lists, has been traditionally embraced by classrooms around the world. However, evidence shows that explicit vocabulary instruction leads to a decontextualized understanding of words. Rather than explicitly instructing vocabulary, people have been arguing for learning new words through emergent use in authentic contexts to support a more holistic understanding of them. A social-pragmatic view of language development has been proposed, which argues that language acquisition is driven by social interaction and the child’s need to connect with others.

The practice of vocabulary development present in knowledge building embraces such a view – by immersing students in a literate environment extensible to the broader world on the Internet and beyond, in which they advance understanding collectively by building and improving explanations together. The Knowledge Building proposition is that immersion in complex literate worlds from early ages of schooling will lead to advances in both basic and advanced competencies. In knowledge building, students write frequently in Knowledge Forum, an online space for them to communicate ideas. However, little research has done to understand vocabulary development in knowledge building, especially from a longitudinal angle (not surprising given longitudinal work is so rare in CSCL).

A recent study put together by me and colleagues, to be presented at CSCL2015 in Sweden, started to attend to this gap. By tracking writing by a cohort of students over 6 years, we were trying to construct the picture of their productive vocabulary growth in knowledge building. Below are some results in and not in the CSCL paper.

Students Write, Read, and Make Edits

Writing, reading and revising activities were happening over the course of six years. There was a trend of increase during the first four years but activities dropped in Grade 5 and 6. This was not surprising because as students get older, a lot of their writing normally happens outside of Knowledge Forum (especially in busy graduation seasons).

Not surprisingly either, the total number of words produced by each student in a year was positively correlated with the frequencies of their reading, writing, and revising activities. The more students use Knowledge Forum, the more words they produce. It makes sense.

Students Wrote More Complex Words

More interesting things started to emerge when looking at student writing more closely. We used a measure called P_Lex to measure the extent to which students use more advanced vocabulary in their writing. Four levels of vocabulary – drawn from Laufer and Nation’s (1995) work – were distinguished in the analysis of student writing: first 1000 word families, second 1000 word families, the Academic Word List, and words beyond these three lists. A higher P_Lex score of a piece of text basically means more higher-level words are present in the text.

Trend analysis identified a significant trend of increase, even though P_Lex dropped in later grades (probably because of the decrease of writing activities in Grade 56). Surprisingly, reading activities, not writing, was identified as the strongest predictor of the P_Lex score.

Different Levels of Words Grew at Varied Speed in Different Years

When tracing each student’s acquisition of words in these four lists in each year, we found the growth of word types from four lists were uneven. Noticeably, the growth of the first 1,000 words peaked in Grade 1 (from 0 to approximately 80 on average), whereas the Academic words and words out of the lists grew more rapidly in Grade 5 and 6 (even though students wrote less).

Additional correlation analysis indicated that students who wrote and read more in Knowledge Forum were likely to demonstrate greater growth in productive vocabulary. Interestingly, in this case note modification appeared to be the most significant predictor for vocabulary growth – more strongly correlated with the growth rate of vocabulary (r = .35, p < .001) when compared with the other two KB measures.

Caveats and Thoughts

Overall, this study highlighted some potential linkages between knowledge building and vocabulary growth. However, there has been some caveats with the current analysis. First of all, we had challenges with tracking students across years. One of the things was some students switched to schools at some point and some others joined. Also, we could not give enough consideration to different teaching styles and sociocultural dynamics in different years. Another challenge was to attribute this challenge to knowledge building; even though there were significant correlations between vocabulary growth and knowledge building activities, it remains hard to distinguish them from “natural” growth of students.

A possible direction to take is to align multiple cohorts of students together so that comparisons could be made within a same grade (same teacher, different cohorts). It would also be fun to explore possible learning analytics to help students collectively develop lexicons. The following (intricate) graph shows the use of Academic words (y axis) by students (x axis) in each year (divided by the vertical green line). While it is clear that most students are comfortable with using some words, e.g., “theory”, it is also evident that different students have demonstrated better access to different words. Can we build tools to capture the appearance of some “big” words within a student community and engage students in talking about them in a meaningful context of knowledge building? It would be fun to help students “light up” the grey areas – not though explicit instruction but exchanges among themselves!

A Wired Thing…

Nothing really relevant – but I was bored enough to map out the locations of notes in Knowledge Forum. For those who are not familiar with Knowledge Forum, it provides a 2D space, called view, to place and organize notes in meaningful ways, normally starting from the upper-left corner of a screen. It was not surprising most notes sat in the area within 0 - 1000 pixels in both x and y axises. But it was surprising some notes appeared in the far upper-left corner. I guess we can probably go treasure hunt in Knowledge Forum some time :)