Impacts of access to book corners and reading diaries on cognitive and non-cognitive outcomes: A field experiment in Nan, Thailand

The study aims to evaluate the impacts of access to book corners that are equipped with grade-appropriate children’s books and reading diaries to primary school students on both cognitive and non-cognitive outcomes and in both short and long runs. We randomly divide randomly chosen 62 schools with altogether 5,708 students from Grades 1 to 6 in Lower Nan Province into three groups. The first group of 21 schools each receives a book corner with a set of 129 children’s books. The books have been evaluated by children’s books experts and academics. They cover children’s literature, popular science, character-building children’s books, and creativity-and-imagination-enhancing children’s books. The book corner consists of an inviting, colourful and visible shelf, sitting mats and a Japanese table. Also, for the first treatment group, each individual student gets a grade-tailored reading diary, designed by the Programme, to match the children's books for each grade. The second treatment group of 21 schools is also provided with a book corner but the students get no reading diaries. The two treatments help distinguish whether there are differential cognitive and non-cognitive impacts between reading-for-purpose (Treatment 1) and reading-for-pleasure (Treatment 2) among primary school students. The rest of the 20 schools constitute the control group. The treatment schools in both groups get to keep the book corners even after the completion of the experiment.

2017-06-19

2023-02-11

End point 1: Short-run impacts after an academic year
(i) Library-using and non-textbook reading behaviours
(ii) Cognitive outcomes in terms of literacy of Thai language tested by the Programmes’ own devised Thai reading and writing tests for each grade
(iii) Non-cognitive outcomes that can be related to reading – attentiveness
(iv) Non-cognitive outcomes that can be related to reading -- empathy

End point 2: Medium-run impacts after a summer holiday and before the students learn new literacy material of the new academic year according to their curriculum
(i) Cognitive outcomes in terms of literacy of Thai language tested by the Programmes’ own devised Thai reading and writing tests for each grade
(ii) Non-cognitive outcomes that can be related to reading – attentiveness
(iii) Non-cognitive outcomes that can be related to reading -- empathy

End point 3: Long-run cognitive outcomes after three years from the start of the intervention.
Using the reading and writing Thai language tests for each grade, based on the nationalised tests by the Ministry of Education.

End point 4: Long-run cognitive outcomes when students from each grade reach grade 6.
Using the nationalised O-NET exam on Thai language, Mathematics, Science, and English to test the long-term effects whether reading can have a spillover impact on the proficiency of other subjects that are also based on the ability to read, understand and analyse, and concentrate.

Note that the O-NET, which stands for "Ordinary National Educational Test," is a standardised, nationwide exam in Thailand that assesses the knowledge and abilities of students in grades 6, 9, and 12, primarily used to evaluate the quality of education at a national level and inform teaching practices within schools; it is administered by the National Institute of Educational Testing Service (NIETS) and is based on the Basic Education Core Curriculum of Thailand.

End point 1:
(i) Library-using and non-textbook reading behaviours
Constructed from the questionnaires asking
1. “Do you know where the library is?”
(a) Yes
(b) No
2. “Do you find library as a relaxing reading space?”
(a) Yes
(b) No
3. “Do you find library as a peaceful place you can focus?”
(a) Yes
(b) No
4. “How often do you read non-textbooks?”
(a) Every day
(b) Some days
(c) Never
As reading every day is a small fraction of the answers, we combine (a) and (b) to constitute the variable of reading some days or every day and assign the value 1. Never reading non-textbooks, in contrast, is assigned to the binary measure of 0.
5. “Do you enjoy reading non-textbooks?”
(a) Yes
(b) No
(c) Just so-so
As the number of students who answer NO is very small, we combine Just so-so and No to create the binary measure of 0. For those who enjoy it, we assign the value of 1.

(ii) Cognitive outcomes in terms of literacy of Thai language tested by the Programmes’ own devised Thai reading and writing tests for each grade.
Constructed from the raw scores of each student, we construct a z-score based on the normalisation of subtracting the mean and dividing it by the standard deviation of the treatment group of the same grade. The tests were also piloted in a school in the neighbouring Payao Province to test the validity and the distribution of the raw scores.

(iii) Non-cognitive outcomes that can be related to reading – attentiveness
The construction of the binary variable is based on the questions in Eysenck test (Eysenck, 1984) that are relevant to school-domain attentiveness.
1. “Do you usually work quickly without bothering to check answers?”
(a) Yes
(b) No
2. “Do you sometimes put down the first answer during the test and forget to check it later?”
(a) Yes
(b) No

(iv) Non-cognitive outcomes that can be related to reading – empathy
The ordinal variable construction is based on the following questions.
1. “When reading books or watching animation, I feel bonded and understand the feelings of the characters.”
(a) Yes, that is very true.
(b) Yes, but just sometimes.
(c) Never.
2. “Whenever I have a chance, I always help others.”
(a) Yes, that is very true.
(b) Yes, but just sometimes.
(c) Never.

End point 2:
(i) Cognitive outcomes in terms of literacy of Thai language tested by the Programmes’ own devised Thai reading and writing tests for each grade.
Constructed from the raw scores of each student, we construct a z-score based on the normalisation of subtracting the mean and dividing it by the standard deviation of the treatment group of the same grade. The tests were also piloted in a school in the neighbouring Payao Province to test the validity and the distribution of the raw scores.

(ii) Non-cognitive outcomes that can be related to reading – attentiveness
The construction of the binary variable is based on the questions in Eysenck test (Eysenck, 1984) that are relevant to school-domain attentiveness.
1. “Do you usually work quickly without bothering to check answers?”
(a) Yes
(b) No
2. “Do you sometimes put down the first answer during the test and forget to check it later?”
(a) Yes
(b) No

(iii) Non-cognitive outcomes that can be related to reading – empathy
The ordinal variables are constructed from the following questions.
1. “When reading books or watching animation, I feel bonded and understand the feelings of the characters.”
(a) Yes, that is very true.
(b) Yes, but just sometimes.
(c) Never.
2. “Whenever I have a chance, I always help others.”
(a) Yes, that is very true.
(b) Yes, but just sometimes.
(c) Never.


End point 3: Long-run cognitive outcomes after three years from the start of the intervention
Using reading and writing Thai language test based on nationalised tests by the Ministry of Education.
From the raw scores of each student, we construct a z-score based on the normalisation of subtracting the mean and dividing it by the standard deviation of the treatment group of the same grade. For Grades 4 to 6 students, who were in Grades 1 to 3 at the start of the intervention, the scores cover four categories, (i) reading aloud, (ii) reading comprehension, (iii) total reading, and (iv) writing. As some schools have a secondary-school-level extension, we also test Grades 7 to 9 students, who were in Grades 4 to 6 at the beginning of the intervention. Their scores span two categories, (i) reading comprehension and (ii) writing.

End point 4: Long-run cognitive outcomes when students from each grade reach grade 6.
Using the nationalised O-NET exam on Thai language, Mathematics, Science, and English to test the long-term effects whether reading can have a spillover impact on the proficiency of other subjects that are also based on the ability to read, understand and analyse, and concentrate. For each subject, From the raw scores of each student, we construct a z-score based on the normalisation of subtracting the mean and dividing it by the standard deviation of the treatment group of the same batch.

The trial is designed to test the significance of access to children’s books’ resources or resource constraint hypothesis. As such, we have the treatment groups that are provided with book corners. To further illustrate the mechanism inside the children’s reading process, we have two distinct treatment groups; one with reading dairies designed particularly for each grade’s books and distributed to individual students, and one without. This helps us distinguish whether there is a differential impact between reading-for-purpose and reading-for-pleasure. The trial is conducted upon the 62 randomly chosen state primary schools with altogether 5,708 students in Lower Nan Province, which represents primary schools in Thailand. We obtain the names of the schools in the Regional Education Area One of Nan Province which is in Lower Nan and then randomise to select schools into the sample and then randomise again to assign the schools into three groups. Then we contact the schools to see whether they would like to participate in the experiment. All of them agree. We do not target any group of children but all children in primary schools that participate in the Programme.

Randomisation was done in the office by a computer over all state primary schools in Lower Nan Regional Education Area.

School

Yes

Number of schools: 62 schools in total
Treatment 1: 21 schools
Treatment 2: 21 schools
Control: 20 Schools

Number of students across all treatment arms: 5,708 students altogether from Grades 1 to 6.

Treatment 1: 21 schools and 2,319 students
The schools in Treatment 1 receive a book corner with a set of 129 children’s books. They cover children’s literature, popular science, character-building children’s books, and creativity-and-imagination-enhancing children’s books. The book corner consists of an inviting, colourful and visible shelf, sitting mats and a Japanese table. Also, for the first treatment group, each individual student gets a grade-tailored reading diary, designed by the Programme, to match the children's books assigned for that grade.

Treatment 2: 21 schools and 1,617 students
The second treatment group of 21 schools is also provided with a book corner but the students get no reading diaries.

Control: 20 schools and 1,772 students
The control schools receive nothing.