Page 99 - Second language development of newly arrived migrant kindergarteners - Frederike Groothoff
P. 99

Development of narrative ability 99 However, the NDW is disputed because it is highly influenced by the length of a story. A measure of lexical diversity which seems to reduce the impact of text length is the Guiraud Index Score (GIS; Daller, Van Hout, & Treffers-Daller, 2003). Therefore, in addition to NDW we calculated the GIS by dividing the number of types (different words) by the square root of the total of tokens (total number of words) of the story. The GIS was chosen since it was likely that the population of the present study would, at least in the first narratives, produce stories with a limited text length, the GIS was expected to be a more precise measure than NDW. Lexical richness Another way of looking at lexical items is to assess their richness. Two pupils could have a similar score of NDW or GIS but in practice use different kind of words. For example, one pupil could use the frequently used word “tree” while another pupil would use the less frequent word “oak”. To make a distinction between participants using general, more frequent words and participants using specific low frequent words a measure of lexical richness was added to the analysis. The Measure of Lexical Richness (MLR) takes into account the frequency band of the words. An online program14 which compares the words from the story with a frequency list of Dutch words was used to calculate the lexical richness of the words in the participants’ stories. This program divides the words in nine frequency bands (Vermeer, 2016). Words not recognized by the program were checked and mostly changed from colloquial speech to their written counterpart. For example, “’t” (the reduced form for it) was changed into “het”. Also, incorrectly inflected verbs were rewritten in the correct form, for example “gevliegen” (instead of “gevlogen,” meaning has flown). The words from the higher frequency bands were checked specifically for whether the participant actually meant what the program had assigned as a meaning to the word. Macrostructure Macrostructure was investigated in three parts: through Story Structure, Structural Complexity, and Internal State Terms (ISTs). All three parts of macrostructure are present in the protocol of the MAIN (Gagarina et al., 2012), however, our analysis slightly deviates from that protocol. Table 5.4 illustrates the macrostructure of an example story. Story Structure, Structural Complexity, and ISTs are in column one, four, and five of Table 5.4, respectively. Story Structure For the analysis of Story Structure, the story grammar model (e.g. Mandler, 1979; Stein & Glenn, 1979) was used, but in a slightly adapted version (following Gagarina et al., 2012). 14 https://lukasvermeer.github.io/mlr/  


































































































   97   98   99   100   101