**2. Methodology**

#### **2.1 Participants**

For the purpose of this research we needed the collaboration of students and their parents. The sample was made up of 112 primary school children, with ages ranging from 10 to 13 (mean = 10.46 years), distributed according to study groups, grade and sex as shown in Table 1.


Table 1. Sample of students according to their experimental group, course and sex

All participants attended school regularly and none of them had any specific educational support needs. The children of the different groups did not initially differ, according to our pretest in any of the assessed variables.

As for families, the following table describes the main characteristics of the parents that were actively involved.


Table 2. Characteristics of EFP and PAD family groups

In the other two study groups family collaboration was not explicitly required although their main features may be of interest. Their ages ranged from 32 to 52 years, with a mean age of 42.9. In terms of levels of education, 21% had reached primary, 50% secondary and 29% higher education. Finally, 77% were working, while 23% were stay at home parents.

#### **2.1.1 Sample selection**

The process of selection and distribution of the sample across different groups followed a procedure which aimed to achieve a correct and comparable distribution of participants to experimental conditions.

First, to select the schools, we took several criteria into account with the aim of achieving the maximum possible similarity between them, both with regard to their structure and organization, and with regard to the characteristics of students and families. Specifically, we selected state schools in which the teachers' profile, teacher-student ratios and the availability of human and material resources or infrastructures were similar. In addition, these centers hosted middle-class families with traditional structures and were mostly Spanish. Finally, based on the objectives of our the research and considering the interests and availability of schools and families we established the need to involve four schools. As noted, we tried to control their differences, thereby trying to overcome the possible handicaps resulting from the impossibility of making a completely random distribution of participants to experimental groups, as the involvement of parents and teachers was voluntary.

Of the institutions addressed, two of them declined the option of increasing family collaboration, and were thus assigned to the PAD and EFP experimental groups. In these cases, only the parents who showed interested in family training were enrolled in it, which determined the group of students in EFP. Other parents of those schools, unable to attend the training sessions, chose the option of assisting their children more actively with writing assignments and were assigned to the PAD group. Therefore, children in groups FP and PAD were enrolled in the same schools, were classmates and had the same teachers, so that one group served as control for the other. PRO was carried out at a third school. The faculty of the third cycle of primary was responsible for the direct implementation of the intervention program in written composition in 5th and 6th year of primary. The last school took part in the CO group. It therefore only carried out the relevant assessments of students in grades 5 and 6, maintaining the ordinary curriculum regarding teaching of writing and regular family help with homework, thus also acting as a control for the PRO group.

## **2.2 Instruments**

108 Learning Disabilities

For the purpose of this research we needed the collaboration of students and their parents. The sample was made up of 112 primary school children, with ages ranging from 10 to 13 (mean = 10.46 years), distributed according to study groups, grade and sex as shown in

 *Boys Girls Total Boys Girls Total Boys Girls Total Boys Girls Total* 

grade 11 11 22 8 8 16 5 8 13 7 6 13

grade 7 6 13 6 4 10 6 6 12 7 6 13

Sx/Gr 18 17 35 14 12 26 11 14 25 14 12 26

All participants attended school regularly and none of them had any specific educational support needs. The children of the different groups did not initially differ, according to our

As for families, the following table describes the main characteristics of the parents that

 *Men Women Mi Ma M 5th 6th Primary Secondary University Active Non* 

EFP 0% 100% 36 52 42 62% 38% 12% 41% 47% 56% 44% PAD 8% 92% 33 51 41 52% 48% 36% 48% 16% 72% 28%

In the other two study groups family collaboration was not explicitly required although their main features may be of interest. Their ages ranged from 32 to 52 years, with a mean age of 42.9. In terms of levels of education, 21% had reached primary, 50% secondary and 29% higher education. Finally, 77% were working, while 23% were stay at home parents.

The process of selection and distribution of the sample across different groups followed a procedure which aimed to achieve a correct and comparable distribution of participants to

Parent's highest level of education achieved

Labor market participation

*active* 

Table 1. Sample of students according to their experimental group, course and sex

grade

Table 2. Characteristics of EFP and PAD family groups

PRO EFP PAD CO

**2. Methodology 2.1 Participants** 

Table 1.

5th

6th

Total

pretest in any of the assessed variables.

Sex Age Child's

were actively involved.

**2.1.1 Sample selection** 

experimental conditions.

#### **2.2.1 Performance assessment in written composition: product and process measures**

To evaluate the written performance of students we sought to cover their written products as cognitive processes activated in word processing, using different tools previously validated by the research team.

#### **Product Measures**

The written products were evaluated using two types of measures, some based on objective evidence collected in the text or text-based measures (MBT). Others used subjective criteria, based on the overall interpretation of the text by the reader (MBL). Table 3 lists the parameters included in the MBT, as well as the criteria to be observed by the reader to estimate their subjective ratings.

We would like to emphasize that since all these measurements, imply an opinion, assessment or interpretation on the part of the evaluator, corrections were made by two experts who worked independently. We then calculated their rates of agreement, which ranged overall between quite high (between kappa 0.6 and 0.8) and very high (kappa greater than 0.8 points).

Preventing Children's Writing Difficulties Through Specific Intervention in the Home 111

carried out when writing (e.g. Luis thinks "What things can I tell about my city?"). Students had to identify to which of the 7 categories each of the 25 items belonged. When we compared the children's result with that of an independent expert we obtained an average agreement of .957 (kappa index). Students were then asked to write their own text and

A second group of scales were intended to meet different motivational-affective aspects of how students relate to writing, specifically to examine their attitudes, attributional patterns

The survey of attitudes toward writing, included in the test evaluation of planning and writing psychological factors (EPP and FPE, Garcia, Marbán, & de Caso, 2001), allows us to evaluate students' attitudes toward writing tasks. It consists of ten items - statements to which children have to respond "yes", "no" or "neither agree nor disagree" depending on

The questionnaire Motivation to Write II (MOES II, Garcia, Marbán, & de Caso, 2001), examines the actions performed by students in essay writing tasks. It allows us to see whether students attribute their success/failure to their own work, effort, ability or luck. It consists of thirty-two claims in eight scales based on the attribution component considered. Children must respond according to their degree of agreement with each statement on a

The Self-Efficacy in Written Composition Questionnaire (developed by the research team) considered all the high and low cognitive level factors that are involved and influence the writing of texts and over which a person can exert voluntary control. It consists of twenty items evenly distributed among low and high cognitive level processes. Ten items measured mechanical processes such as calligraphy, layout, spelling and grammar and the remaining ten covered substantive processes categorized into information generation, prior knowledge, organization, monitoring, and draft review. The scale asks the student to what extent s/he thinks s/he can achieve/include each item in a text (e.g.: To what extent can I write my text with correct spelling? To what extent can I include many ideas in the text?). Each response is graded on a scale of 0-100, where 0 means "very sure not to" and 100 means "very sure to". The questionnaire was explained and applied immediately before the

**2.2.3 Instruments to assess family involvement: Family Opinions (FAOP)** 

PRAES), which are widely described elsewhere (see chapter in this book).

For the perception of children about parental involvement in education and in written composition of specific academic subjects we used the scales *Family opinions: parental implication in education* (FAOP-HI-IM) and *Family opinions: writing practice* (FAOP-HI-

**2.2.2 Instruments for the evaluation of motivational-affective elements** 

record their progress on a writing log.

and perceptions of self-efficacy.

how well each claim applies their own real attitudes.

scale of 1 (strongly disagree) to 5 (strongly agree).

**Attitudes** 

**Attributions** 

**Self-Efficacy** 

start of the writing assignment.


Table 3. Textual measures based on the text and on the reader

#### **Process Measures**

To measure the activation and deployment of cognitive processes involved in writing tasks we used a variation on Kellogg's triple task (Olive, Kellogg, & Piolat, 2002) that had been used previously by the research team in several research projects (García & Fidalgo, 2006; Torrance, Fidalgo, & García, 2007). This technique works as follows: as students are writing their texts, they hear an audible signal (beep) distributed randomly over time (with a mean onset interval of forty-five seconds), when they must make a direct and immediate retrospection on their thoughts and actions and choose between seven response categories that assess the major cognitive processes involved in writing. For planning, the categories used were: read about the subject, think about the content and develop an outline. As for the editing process, the category included was writing a text. Finally, to assess the review process we considered categories of reading and changing text. Finally, we added a seventh category called "not related to the task".

Prior to the self-reports, students were trained through example to identify and memorize the seven categories. After this, in order to verify the reliability of the process, we presented them with a case study with a total of 25 possible items that Luis, our example of a student, carried out when writing (e.g. Luis thinks "What things can I tell about my city?"). Students had to identify to which of the 7 categories each of the 25 items belonged. When we compared the children's result with that of an independent expert we obtained an average agreement of .957 (kappa index). Students were then asked to write their own text and record their progress on a writing log.
