Desarrollo de la escritura para la codificación y decodificación en niños de primaria de Mérida, Yucatán

The objective of this work was to analyze the development and consolidation of writing in children. 546 children between the age of 6 to 12 years were evaluated using writing and reading tests in order to measure the number of spelling errors. In the results, it was found that children in lower grades frequently make natural spelling errors (modification that alters the phonology of the word, heterophone errors); and that all children make arbitrary spelling errors (substitution of a homophone letter). In particular, it was observed that second grade children do not identify misspelled words and that they are specialists in the use of dominant spellings to represent their respective phoneme. Heterophone and homophone errors decreased in fourth grade children, here there was a greater dominance of the phonemegrapheme relationship and vice versa; furthermore, the spelling stage was observed at this age, since the children identified a greater number of misspelled words, compared to the second-grade children. The sixth graders showed an automation for writing and identifying lexical units as -ción, which means that they can read them as a whole and not as a sequence of grapheme-phoneme units. Finally, learning to write in the Spanish language is fast, the transition from the partial alphabetic stage to the spelling one takes two years. In some cases, however, individuals show difficulties or problems (dyslexia), so they continue to make natural spelling errors, despite being exposed to the same amount of text as their peers. These individuals with learning disabilities were noticeable in sixth grade.


Introduction
Reading and writing involve a complex system that relates to words, meaning continents of different levels of abstraction from the individual's concrete environment, with a graphic expression organized under a certain configuration. It is a skill that is acquired based on the intellectual development of the individual and neurologically implies adjustments related to the acquisition of concepts, the memorization and the management of spelling patterns of words, the establishment of associations of meanings with organized signs and the interpretation of the information.
Models of literacy learning explain the systematic process of acquisition and consolidation of spelling in individuals. In this regard, the author Ehri (2005) carried out a review of all the models that continue to be used. And she immediately proposed a series of phases of reading development. The first phase, known as the pre-alphabetic phase, consists of identifying the words or letters as visual cues. In the second phase, partial alphabetic, individuals know the name of the letters and identify them; however, the subjects still lack decoding skills. In the third phase, complete alphabetic, people develop the decoding strategy, forming connections between graphemes and phonemes, and vice versa. Finally, the orthographic phase consists of storing sequences of combinations of graph-phonemic units, including morphemes, rhymes, monosyllabic and polysyllabic words.
However, the process of learning to read is not the same in all children (Matute, González and Guajardo, 2012). This means that not all individuals have the same facilities or skills for the process.
Correct writing depends on visual memory mechanisms. The storage of lexical units facilitates the recognition of words and spelling patterns (Ferroni, Mena & Diuk, 2016). Therefore, either to write a word well or to identify it in print, as well as for a good development of literacy, the same visual-orthographic imprint stored in long-term memory is called upon. (Gómez et al., 2014).
It should be noted that if the automation mechanisms of an individual do not occur with the expected efficiency according to their age, this will lead to some complications, for example, requiring a longer time to read words than someone without difficulties (Afonso, Suárez and Cuetos, 2020;Suárez and Cuetos, 2013); also the inability to identify spelling errors, which leads to the acceptance of words with homophone spelling errors (E. Hm) (substitution of one letter for another with the same phoneme) as real words (González,  Barrios, Gómez & Zarabozo, 2017 ;González, Gómez and Rodríguez, 2013;Guàrdia, Peró, Zarabozo, González and Gudayol, 2015). Therefore, the objective of this work is to analyze the process of acquisition and consolidation of literacy in 546 individuals of 2nd (164), 4th (184) and 6th (198) grade of primary school in the municipality of Mérida , Yucatán, through writing and reading tests.
The number of spelling errors written was correlated with the variables: the use of inconsistent spellings and the detection of spelling errors when reading.

Methodology
With the written informed consent of the school principal and the teacher of each group, 546 children participated voluntarily: 164 from 2nd grade, 184 from 4th grade and 198 from 6th grade. All the children came from public schools in urban areas and from different socioeconomic levels: low, medium and high.
The schools were randomly selected. As verified, the study plans of the selected educational centers follow the national education plan issued by the Ministry of Public The requirements for participation were the following: that the children know how to read and write; miss a school year; that the mother tongue was Spanish; absence of diagnosis of any disorder or condition of neurological and psychiatric disease, and the consent of the child.
Consequently, 33 2nd grade children (could not read and write), 10 4th grade children (one was a repeater and nine refused to participate) and four 6th grade children (one suffered from neurological problems and three refused to participate). Table 1

Writing battery
Two writing tasks were designed: word and non-word dictation. In the word dictation test, 19 words were chosen that appear at least three times in the Spanish textbook. First grade (Dávalos, González, Kriscautzky & Omaña, 2018) and that have the letters b, c, g, h, j, ll, s, y, z and the accent sign.
In the non-word dictation test, words, morphemes and lexical units were chosen from the spelling rules in Spanish (for example, after the letter n the letter v is written: nv) that appear more than three times in the textbook already mentioned (Dávalos et al., 2018). Later, they were combined with other words, morphemes and lexical units; for example, histo-town: histo is a morpheme and town is a word, and thus 10 pseudowords were formed.
Homophone spelling errors (E. Hm) were evaluated: substitution of one letter for another with the same phoneme; and heterophones (E. Ht): substitution of one letter for another with a different phoneme.

Spelling error detection test
The short story by Agustín Yáñez (100 words) was taken from the Spanish book. First grade (Dávalos et al., 2018). A spelling of 21 words was changed. The substitute spelling had the same phoneme as the substituted spelling: c-s-z (eight), b-v (six), g-j (one) and y-ll (four); in the case of the letter h (two), the letter was omitted. The test lasted five minutes.
In this test, the spelling errors detected (ED) and the false detections (FD) were evaluated, that is, the correctly spelled words selected as misspelled.

Process
The children were evaluated in their classrooms with the presence of the grade level teacher, and on some occasions with the presence of the school director, with a duration of between 10 and 15 minutes. The author of the project was the one who applied all the evaluations.
The order of presentation of the evidence was counterbalanced. Each task was preceded by several examples in order to ensure that the child understood the instructions. Table 2 shows the descriptive results of the word dictation task. The percentage of children who made spelling errors, both heterophones and homophones, is shown.

Dictation of words
All children were found to commit at least one E. Hm. Something equally relevant was that most of the 2nd grade children committed at least one E. Ht (see Table 2). In 2nd grade children, it was found that, in the phoneme / s /, the letter s was better written than the letter c (U = 3859, z = -7.841, p <0.001) and the letter z (U = 675 , z = -13.706, p <0.001); likewise, letter c obtained fewer errors than letter z (U = 5303, z = -6.942, p <0.001). In the phoneme / b /, the children were more assertive in the use of the letter b compared to the letter v (U = 6634.5, z = -3.243, p = 0.001). In the phoneme / j /, the children wrote the letter j better than the letter g (U = 6376, z = -3.740, p <0.001). And in the phoneme / y /, the letter ll was better written than the letter y (U = 2521, z = -10.487, p <0.001) (see figure 1).
In fourth grade primary school children, it was found that, in the phoneme / s /, the letter s obtained less probability of error than the letter z (U = 11300.5, z = -4.192, p <0.001); in the same way, the letter c was better written than the letter z (U = 11723, z = -3.758, p <0.001). In the phoneme / y /, the children were more assertive with the letter ll than with the letter y (U = 6581, z = -9.646, p <0.001). There were no significant differences in the letter s  In the phoneme / j /, with the g there were fewer errors than with the letter j (U = 17131.5, z = -2.078, p <0.05). Finally, with the letter ll, more assertiveness was registered than with the y (U = 10951.5, z = -8.407, p <0.001) to represent the phoneme / y / (see figure 3).
With regard to the silent letter h and the accent, it was observed that children in lower grades use it less, which is why they have a high probability of error; opposite case of the children of superior grades. Table 3 shows the percentage of the number of children who used a spelling to represent the phoneme before non-words; in the same way, the probability of using the spellings to represent the phoneme is observed. Subsequently, the relationship between the number of homophone errors (NEHm) made by the children, from the dictation test, and the probability of use of spellings, from the pseudoword dictation test was searched (see Table 4). For the relationship, the Spearman correlation statistic was used, since the variables did not comply with the Kolmogórov-Smirnov normality test.

Error detection
In the task of detecting spelling errors, it is observed that the 2nd grade children detected few misspelled words and many FD; this means that some children selected words that are not misspelled as misspelled. In the 4th and 6th grade children, EDs increased and FDs decreased (see Table 5). No correlation was found between the NEHm committed by the 2nd year children and the detection of misspelled words (ρ = -0.157, p = 0.075), nor was found between the NEHm committed by the 6th year children and the FD (ρ = 0.061, p = 0.396).

Discussion
In the present investigation, the process of acquisition and consolidation of literacy in children of 2nd, 4th and 6th grade of primary school was analyzed through writing and reading tests.
In general, the descriptive results show a lower number of misspellings, a greater use of dominant letters, and a greater detection of misspelled words in children in higher grades compared to those in lower grades. This difference in degrees is also found in the works of Defior, Martos and Herrera (2000), Defior, Jiménez and Serrano (2009)  The word dictation test is intended to identify the development of spelling knowledge and the age at which it begins to matter, in addition to identifying possible literacy learning problems. In this test two results are obtained: the commission of E. Ht and E. Hm.
It will not be detailed in depth about the E. Ht because a set of phonemes was used (/ t /, / m /, / r /, / l /, / a /, / e /, / i /, / o / y / u /) that result in very few misspellings. Although it would be interesting to continue with this research and find out what are the main causes that lead the individual to transfer the phoneme to a spelling that does not correspond to the moment of writing, as was done in the work of Leal, Matute and Zarabozo (2005).
A cursory review of E. Ht's results show that over 90% of 2nd graders make this type of misspelling. Researcher Ehri (2005) mentions that, being new to the field of reading and writing, individuals lack the skills to decode and encode. So E. Ht's commission in writing would be considered normal at this school level.
The opposite is the case for children in 4th and 6th grade of primary school, who have had a greater exposure to printed texts, at least the books of the previous grades they have studied, for which the E. Ht commission could be considered a problem of the literacy learning (Baron et al., 2018;Morken and Helland, 2013;Suárez and Cuetos, 2015) This effect is detailed in the work of Landerl, Frith and Wimmer (1996), who point out that the problem in individuals with dyslexia is explained by a weak link between phonological and orthographic representations. This means that seeing a written word does not automatically evoke the internal sound of the words; nor does the sound of a word automatically evoke the internal orthographic representation.
With respect to the E. Hm, it is observed that all the children represented the phoneme, in most of the times, by the same letter, known as the dominant spelling. This shows that children are specialists in the use of dominant spellings in writing, both in words and in nonwords. The most used letters to represent the phoneme / s / is s, the phoneme / b / is b, the phoneme / j / is j, and the phoneme / y / is ll.
These results are similar to many research works Gómez et al., 2014;Jiménez et al., 2008). Therefore, it is interpreted that the participants have a good phonological command, but have poor spelling knowledge Signori and Borzone, 2003). The specialization of spellings in writing decreases as the school grade increases (Defior et al., 2000;Gaintza, 2005), but it does not disappear, since individuals of adulthood continue to make this type of spelling error (Gómez et al., 2014;Moojen et al., 2020).
A relevant result of the word dictation test is the increase in the probability of error of the letter s in 6th graders of primary school (see Figure 3). The increase in error is considered to be due to the invocation of small lexical units, known as morphemes. In the word dictation test, words with the phonological ending / sion / were used, normally the words that end with this sound are represented in printed form with the spelling c, as -ción; but in this work the words were represented with the spelling s, for example: precision, version and invasion.
In transparent language works, in Spanish (Suárez, Martínez and Cuetos, 2017) and in Italian (Angelelli, Marinelli, De Salvatore and Burani, 2017), the use of morphemes in "normolector" and "dyslexic" children was evaluated. The results show that, regardless of the group, children are more exact in writing the words and pseudowords that have frequent morphemes as roots.
In the Spanish language, -ción is a frequent morpheme and -sion is considered an infrequent morpheme, since its use depends on knowing some spelling rules of the Spanish language, so these results give indications that children of 6th grade they begin to write through what is stored in long-term memory and not through the phonological, in other words, with the orthographic or lexical knowledge.
Thus, there is a storage of small and large lexical units in long-term memory that contribute to writing correctly (Ferroni et al., 2016;Gómez et al., 2014, Suárez et al., 2017 and identifying words ( Suárez and Cuetos, 2013;D'Alessio, Wilson and Jaincheco, 2019).
However, in this case, and according to our results, it interferes with correct writing.
Despite not being one of the objectives, nor of the hypotheses raised, the result of the The results of the spelling error identification test are similar to the research works of some authors (González et al., 2013;Guàrdia et al., 2015;González et al., 2017;Moojen et al., 2020), who indicate that people with a good lexicon detect more words with spelling errors and have different electrophysiological responses (González et al., 2013) and neuroanatomical activations (González et al., 2017;Guàrdia et al., 2015) than individuals with a low lexical level.
Something that reinforces the previous point is that in this study there is no correlation between NEHm and ED in 2nd grade children: having little reading experience, the children guess what the misspelled word is (see Table 5). This means that children do not recognize the spelling part of words.
Word recognition is seen in 4th grade children, with an increase in the detection of misspelled words and a decrease in false detections, in addition to a decrease in dominant spellings. This agrees with the authors Jiménez et al. (2008), who mention that Spanish children in 4th grade begin to write with spelling knowledge.
The learning process of writing is similar to the learning model of the author Ehri (2005). Writing begins in the partial alphabetic phase, where individuals begin to use the sounds of some letters to form partial connections, syllables; however, they lack coding skills, as seen in this research paper. Thus, children resort to different strategies such as guessing, using partial phonetic clues or similar letters.
The entire alphabetic stage is visualized in 4th grade children. In this phase, the ability to decode and encode is developed, the subjects resort to the grapheme-phoneme strategy and vice versa. In addition, the alphabetical consolidation or spelling stage is displayed, where storage in long-term memory begins, which facilitates the automation of words to later recognize or read them at a "glance": morphemes, rhymes, monosyllabic words and spelling of most frequent syllables in polysyllabic words, to name a few examples.
As readers learn to read words that share letter patterns, those that symbolize the same combination of phonemes of different words, eg, song, publication, communication, a consolidated unit is formed. Knowledge of -tion as a consolidated unit means that readers can read it as a whole and not as a sequence of grapheme-phoneme units.

Conclusions
Spelling errors are common in the literacy learning process and decrease as there is greater exposure to books. However, they do not disappear.
In tune with the acquisition and consolidation process, similar to that described by current models, despite the difference in languages, 2nd grade children begin the development of phonological awareness, phoneme-grapheme relationship and vice versa, but they have poor spelling knowledge, which allows them to write most of the words according to sound, characteristic of the partial alphabetic phase.
Whereas, in 4th grade children a greater domain of phonological awareness is observed and they begin to develop the storage of lexical units, both monosyllabic and polysyllabic. Although the storage of letter combinations is not always favorable. Therefore, this grade level is considered to be a transition between the full alphabetic phase and the spelling phase.
Taking the above into consideration, we can say that the learning process of writing in the Spanish language is similar to languages considered opaque, such as English, with the only difference that the transitions between stages are faster.