Harpin's type-token ratio of the first 100 words indicates vocabulary diversity, but the rate of TTR drop after the first 100 words does not differentiate attainment levels¶

Claim · All claims
Evidence · 1 study (2 entries) · associational r1 · q2 · i0 negligible
Currency · newest study 2002 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? The TTR of the first 100 words gave the clearest indication of vocabulary diversity, broadly rising with attainment level (0.54 at Level 2c to 0.75 at Level 5+). → C.E. Holdich 2002 q2 i0 Analysing the rate of drop in TTR after the first 100 words did not, as Harpin hypothesised, contribute to differentiating between the levels; little discriminated between levels other than the Level 2 story. → C.E. Holdich 2002 (2)

Evidence¶

C.E. Holdich 2002¶

C.E. Holdich, R.G. Holdich and P.W.H. Chung. (2002). Assessing aspects of children's written grammar: automating the process. Computers and Education, Vol 39, pp37-50. http://creativecommons.org/licenses/by-nc-nd/2.5/

q2 · i? · design · r2

Harpin's (1976) type-token ratio technique applied to the first 100 words of each of the 12 stories; stories 10 and 11 (TTR 0.73 and 0.75) were especially praised by examiners while stories 1 and 2 (0.54 and 0.60) showed noted repetition.

"Broadly, the lower levels of attainment resulted in low TTR ratios whilst the higher levels of attainment resulted in higher TTR ratios, alt hough the results inevitably deviated on occasion depending on the strengths and weaknesses of the particular young writer and on the theme and organisation of the story."

C.E. Holdich 2002 (2)¶

C.E. Holdich, R.G. Holdich and P.W.H. Chung. (2002). Assessing aspects of children's written grammar: automating the process. Computers and Education, Vol 39, pp37-50. http://creativecommons.org/licenses/by-nc-nd/2.5/

q2 · i0 · associational · r1

The normalised TTR-drop curves plotted in Figure 1 showed only the Level 2 story dropping sharply; the article reports the drop-rate measure was not useful for quantitatively differentiating levels, a null result for Harpin's extension.

"Whilst the TTR for the story achieving Leve l 2 dropped sharply, th ere was little to discriminate between the other levels. Although Harpin was correct in his hypothesis that the rate of fall in the TTR would reveal the extent of variety of a child's vocabulary, the results did not prove useful in a quantitative analysis of the various levels."

Discussion¶

Related Claims¶

• The ratio of mean sentence length to commas was the most useful computer test for discriminating maturity level in children's writing — related
• Story length correlates with National Curriculum attainment level in children's Key Stage 2 narrative writing, with roughly 300 words needed for Level 4 and 400 or more for Level 5 — related
• Feature operationalization can reverse the direction of the relationship with essay score, as with document length and TTR variants — related
• The most predictive linguistic features differ across datasets, overlapping only in lexical diversity measures — related