Family Intervention for Empowerment through Reading and Education (FIERCE)

In this RCT, we will evaluate the the We Love Reading (WLR) program. We Love Reading is 12-week non-formal education programme developed in Jordan by the Taghyeer Foundation. We Love Reading is locally embedded, community-based, and has a high uptake among adults and children living in northern Jordan. WLR was developed in response to needs identified by the local community and involves training local volunteers to read stories to children (ages 4–12) at least once a week.

We Love Reading involves people volunteering for, being trained in, and then delivering a shared book-reading intervention. Training in the We Love Reading intervention is conducted by Taghyeer Foundation staff over two days, with groups of 20-25 people attending each training session. The training equips local volunteers with the skills they need to implement the read aloud method, as well as builds capabilities in social entrepreneurship. Every module is highly interactive and involves a range of exercises with the participants. These volunteers are given books by the Taghyeer Foundation; these books were specifically designed and printed for use in the We Love Reading programme. The people taking part in the training deliver the intervention immediately following the training. Each person gathers groups of 6-10 children once per week for 12 weeks to read a children’s book to the children and discuss together in safe spaces, such as homes, mosques, and community centres. Children are recruited into groups from the adult’s networks: their children, family members, family friends, and neighbours. Each We Love Reading volunteer delivers the We Love Reading programme on their own, with one adult per group of children.

2021-01-01

2022-04-30

We have two primary outcomes: children's literacy and attitudes toward reading.

We will measure literacy using a tool designed for use with refugee children, the Holistic Assessment of Development and Learning Outcomes (HALDO). HALDO is designed to be administered to children ages 4-12 and has an adaptive structure to account for the wide age range. We will be using the literacy component of the HALDO, which assesses common letter identification, uncommon letter identification, expressive language, and reading with comprehension. Child literacy will thus be measured solely in the children (no parent report). This measure is available in Arabic and has been previously used with refugee children in Jordan and across the MENA region.

We will measure the child's love of reading using both child- and mother-report. The children will complete the following survey measures:
1) Two subscales of the Young Reader Motivation Questionnaire (YRMQ): the 4-item efficacy for reading subscale and the 3-item reading orientation subscale)
2) 8 items from the reading attitude subscale of the Reader Self-Concept Scale (RSCS).
Given the age of our sample, that many Syrian refugee children in Jordan do not attend school, and levels of poverty among potential participants, we have chosen subscales and items from established measures which do not refer to reading in class, reading with a teacher, or books in the home.
We will ask the mothers about the children's love of reading using 7-item parents’ perceived child attitudes towards reading scale. None of these scales are currently available in Arabic.

To prepare the measures for use in this study, all of the attitudes toward reading measures will be translated to Arabic, back translated, and then evaluated by experts for relevance and consistency with the original construct being measured before being implemented. In addition to the one literacy measure, we plan to have three attitude to reading outcomes: one variable which combines the two subscales of the YRMQ, one variable which consists of child responses on the RSCS, and one variable which consists of mothers' responses to the parents' perceived child attitudes towards reading scale.

We have two secondary outcomes: 1) family functioning and 2) cognitive and social stimulation in the home.

Family functioning will be measured using both observation of mother-child structured and unstructured interactions as well as by the mother-reported 14-item Arabic Revised Dyadic Adjustment Scale, the 15-item Pianta Child-Parent Relationship Scale, and the 20-item Arabic Father Involvement Questionnaire. The Pianta Child-Parent Relationship Scale has two subscales: closeness and conflict. The Father Involvement Questionnaire has three subscales: 1) fathering involvement with the child, 2) fathering involvement with the mother, and 3) fathering involvement in the community.

Cognitive and social stimulation in the home will by assessed using observations of the mother-child structured and unstructured interaction tasks. The observational task has two components: one reading task of up to 10 minutes and up to 5 minutes of unstructured play. For the reading task, mothers and children will be given the choice of four books (the same four options for all participants) and the mother will be asked to read the story to the child. For the play task, the mothers and children will be given a set of toys (the same set for all participants) and told to play freely with the toys for up to 5 minutes. Coding of the observational task will be conducted by trained research assistants.

Qualitative component:
A postdoctoral research assistant will conduct one-on-one semi-structured interviews (either in person or over the phone, COVID-dependent) with adults and children who have taken part in We Love Reading as well as with those who have not, to understand how We Love Reading influences education and learning. These interviews will be conducted with participants living in Amman and surrounding areas – where a substantial proportion of Syrian refugees in Jordan live – and in Za’atari refugee camp or Mafraq (depending on COVID restrictions). The postdoctoral research assistant will conduct an inductive thematic analysis on this interview data. They will transcribe the interviews, generate initial codes, and search for and review themes, resulting in an understanding of the role of non-formal education programs such as We Love Reading in their learning and education. To ensure credibility and dependability, we will use reflective commentary, second coding, and peer feedback. The total sample for this qualitative component will be ¬30 adults and ¬30 children.

Quantitative component:
A project manager and 5 fieldworkers based in Jordan will be conducting data collection for the cluster randomized controlled trial with Syrian refugee mothers and their 4-8-year-old children participating in We Love Reading groups. We will collect data from 180 mother-child dyads from the children taking part in their We Love Reading sessions, as well as 180 control women and their 180 4-8-year-old control children. Where a mother has more than one eligible child within the 4-8-year-old age range, the child closest in age to 6 years old will be sampled. Data will be collected at two timepoints (immediately before taking part in and immediately after the intervention), allowing us to evaluate the effects of We Love Reading on children’s love of reading and literacy skills. Participants will be sampled from We Love Reading training sessions in Amman and Za’atari (or, if COVID-restrictions are still in place in Za'atari, in Mafraq). We will use a cluster randomized wait-list design: we will contact community based organizations, will assign them as either wait-list control or experimental organizations (without informing them), and then will have information sessions for We Love Reading at their organizations and sample women who have expressed an interest in We Love Reading sessions at those sessions.

Sampling will be done as follows: We will use coin flip randomization to determine whether the community based organization will be in the treatment or control condition. 3-5 of the organizations will immediately receive training and the other 3-5 will be wait-listed for training. Women sampled from information sessions held at each of these community based organizations will then be asked inform us which children they will have in their reading groups once they are trained, and fieldworkers will contact the parents of those children to see if they would be interested in taking part in our study. Following this, we will have baseline data collection with the 180 treatment mother-child dyads and the 180 wait-listed control dyads. After approximately 12 weeks, we will conduct endline data collection. We have received additional funding and so will follow up a proportion of the participants for a T3, which will take place approximately a year after the first timepoint data collection. Following the end of data collection, the women interested in being trained in We Love Reading from the control organizations will take part in We Love Reading training and thus all of the children will have access to this intervention.

Jordan-based fieldworkers will go to participants’ homes in groups of two, with one fieldworker collecting data from the mother and the other collecting data from the child. Data collection will consist of surveys, behavioural /cognitive tasks, and a structured mother-child observational task. This combination of psychometric measures (both child- and mother-report), behavioural measures, and direct observation will allow us to gain a holistic understanding of children’s love of reading and literacy. These measures will also allow us to establish whether the effects of We Love Reading are direct, or whether the impacts occur through changes to children’s cognitive stimulation, social stimulation, or through family relationship quality. Using this multi-method approach will provide multiple ways of assessing intervention efficacy, reducing reliance on participant self-report. To increase understanding and avoid issues arising from potential low literacy of participants, fieldworkers will ask all survey questions aloud and input the child/mother response into a laptop loaded with the survey software Kobo Toolbox.

We will collect fidelity information from the volunteers leading the We Love Reading sessions, asking for weekly updates of what book they read, which children attended, where the session was held, and how long the We Love Reading session lasted.

We will use cluster randomization, where the children will be randomized into either treatment or control dependent on which community based organization the person leading their We Love Reading group was sampled from. We will do determine whether community based organizations are treatment or control using a coin flip randomization, such that ~50% will be in the treatment and ~50% in the wait-listed control group. We will sample 6-10 community based organizations to get the needed sample size.

Our unit of randomization is the community based organization from which we draw volunteers to lead We Love Reading sessions. Taghyeer is affiliated with a number of community based organizations who would be willing to host information sessions on We Love Reading. We will gather a list of 6-10 of these who could hold the sessions in January-May 2021 and then will randomize the organizations into treatment or wait-listed control.

In the information sessions, we will sample women who are interested in being trained to lead We Love Reading sessions. If they are sampled from a treatment organization then they will receive this training right away. If they are sampled from a control organization then they will not receive this training until after endline data collection (i.e. after 12+ weeks). The children who would take part in these women's We Love Reading sessions are thus randomized to treatment or wait-listed control on the basis of whether the potential leader of the We Love Reading session is sampled from an organization which is either treatment or control.

Yes

We intend for there to be 6-10 clusters. 3-5 of these will be randomized to the treatment condition who are trained in We Love Reading immediately and 3-5 will be randomized to the wait-list control condition and will be trained in We Love Reading following endline data collection. This randomization is done at the level of the community based organization and will involve coin flip randomization by the study PI.

We will sample 360 4-8 year old Syrian refugees and their 360 mothers (N = 360 mother-child dyads, 720 participants in total) at baseline. We expect attrition at a rate of 25%, for an endline sample of 270 mother-child dyads. As far as clusters in which the randomization will be conducted, we expect to have 6-10 clusters, approximately 4 of which will be treatment and 4 of which will be wait-listed control clusters.

We intend to have the sample approximately evenly split between the treatment and wait-listed control groups. This will entail a sample of 180 mother-child dyads in the treatment group (who take part in We Love Reading right away) and 180 mother-child dyads in the wait-listed control group (who take part in We Love Reading after endlne).

Power analyses suggest with a sample of 265 mother-child dyads at endline, we will be able to detect effects of small to medium effects (f = .20, in line with Dowdall et al., 2019) with 90% power at an alpha of .05 in a mixed ANOVA. To achieve this sample at endline, we plan to sample 360 mother-child dyads. We are expecting retention rates of at least 75% based on previous research by our team with refugees in Jordan (360 x .75 = potentially 270 participants at endline).

Analysis plan - 27 October 2021