Page 255 - IJET_July2021_final

section is called the ‘Basic Statistics’ the distinctive indicative measures
where the total words, total pruned of Complexity, Accuracy, Lexical and
words and total time is listed down. Fluency as defined in the ‘constructs’
The rest four sections are Complexity, section of the paper. Under the
Accuracy, Lexical and Fluency measure complexity section, the two measures
sections. The complexity section based that are taken into consideration for the
on the coding done on the speech data analysis are number of AS Units (T- units)
produces information such as number and the ratio of clauses per AS unit. The
of AS units, total number of clauses, accuracy of the speech data is assessed
number of independent and dependent by the ratio of error- free clauses
clauses, number of before and after and number of errors per 100 words
clauses, ratio of clauses per AS unit, which is comprehensive of syntactic,
number of pruned words per AS unit morphological, lexical and phonological
and number of words per clause. The errors. The F- score stands as a whole
Accuracy section provides data on the representative of the lexical measure
ratio of error- free clauses, errors per and the raw speech rate stands for the
100 words, syntactic, morphological, fluency in the speech.
lexical, phonological discourse and Reliability of CALF tool
pragmatic errors per 100 words.
The CALF tool being a measurement tool
The lexical section of the CALF system of the complexity, accuracy, lexis and
provides data on the significant F- score fluency of the language data, the data
and lexical density. The fluency measures was coded twice at an interval of two
that are derived out of the output are weeks and the results were tabulated.
raw speech rate, pruned speech rate, Since the nature of the measured data
mean length of run, phonation time, is of continuous type and because the
number of mid-clause pauses and speech data sample was the same in
silences per 100 words, number of two different administrations, Intraclass
independent and dependent clause Correlation Coefficient (ICC) was used
pauses, clause silence and average to establish the reliability of the CALF
clause length. It also provides data on tool in SPSS (see Table-1). The model
number of filled pauses, pseudo filled adopted in the correlation is model 3 as
pauses, false starts, reformulations, the raters of interest were of only the
repetitions and replacements.
researcher and it was two-way mixed.
For this study, the researcher has chosen

Table-1: Details of Respondents

95% Confidence Interval F Test with True Value 0
Intraclass
Correlation b Lower Upper Value df1 df2 Sig
Bound Bound

Indian Journal of Educational Technology 245
Volume 3, Issue 2, July 2021