The effects of self-assessment

An article published in Educational Research Review examines the effects of self-assessment on self-regulated learning (SRL) and self-efficacy in four meta-analyses.

To understand the impact of pupils’ assessment of their own work, Ernesto Panadero and colleagues from Spain analysed 19 studies comprised of 2,305 pupils from primary schools to higher education. The meta-analyses only included studies published in English that contained empirical results of self-assessment interventions in relation to SRL and/or self-efficacy, had at least one control group, and had been peer-reviewed.

The findings indicated that:

  • Self-assessment had a positive effect on SRL strategies serving a positive self-regulatory function for pupils’ learning, such as meta-cognitive strategies (effect size= +0.23).
  • Self-assessment had a negative effect on “Negative SRL”, which is associated with negative emotions and stress and is thought to be adverse to pupils’ learning (effect size=-0.65).
  • Self-assessment was also positively associated with SRL even when SRL was measured qualitatively (effect size = +0.43).
  • Self-assessment had a positive effect on self-efficacy (effect size= +0.73), the effect being larger for girls.

The authors suggest that self-assessment is necessary for productive learning but note that the results have yet to identify the most effective self-assessment components (eg, monitoring, feedback, and revision) in fostering SRL strategies or self-efficacy.

Source: Effects of self-assessment on self-regulated learning and self-efficacy: Four meta-analyses (November 2017), Educational Research Review, Volume 22.

Does school entry age matter?

In the UK, children usually start primary school in the academic year in which they turn five. However, because entry rules vary across local authorities, some schools may defer entry for children born later in the year until the second or third term.

A new study at University College London looks at what impact an earlier versus later entry into Reception has on pupils’ cognitive and non-cognitive skills up until age 11 (their final year of primary school).

Christian Dustmann and Thomas Cornelissen analysed information on more than 400,000 children born in 2000-01 who attended state schools in England and whose records are included in the National Pupil Database. This was combined with information on more than 7,000 children born in 2000-01 who took part in the Millennium Cohort study.

The researchers found that receiving an extra month of schooling before age five increases test scores in language and numeracy at ages five and seven by about 6–11%. But by age 11, the effects on test scores have largely disappeared. For boys from low socioeconomic backgrounds, the benefits of an earlier school entry are even greater. An additional term of schooling before age five reduces the achievement gap between boys from low and high socioeconomic backgrounds at age seven by 60-80%.

Source: Early school exposure, test scores, and noncognitive outcomes (March 2019), CReAM Discussion Paper Series CDP 03/19, Centre for Research and Analysis of Migration

How do young children develop agency, literacy, and numeracy

A new resource from Deans for Impact summarises current cognitive-science research related to how young children – from birth to age eight – develop skills across three domains: agency, literacy and numeracy.

It aims to give guidance to anyone working in education who is interested in understanding the science of how young children develop control of their own behaviour and intentions, how they learn to read and write, and how they develop the ability to think mathematically.

For each domain, the report identifies key questions about learning and provides a short list of the principles from learning science that inform the answers to these questions. The resource then connects these principles to a set of practical implications for specific teaching strategies. The original research is clearly referenced for anyone wishing to find out more.

Source: The science of early learning: How do young children develop agency, literacy, and numeracy? (2019), Deans for Impact.

Ethnic minority pupils disproportionately identified with special educational needs

Pupils from ethnic minority groups are over-represented for some types of special educational needs (SEN) and under-represented for other types compared to white British pupils, according to new research led by Steve Strand and Ariel Lindorff at the University of Oxford.

Using data from the England National Pupil Database from 2005–2016, the report looks at all children age five to 16 in England who have been identified with different types of SEN. As well as identifying ethnic disproportionality, the report also considered whether socio-economic factors, such as poverty and neighbourhood deprivation, or children’s early attainment, had any impact on pupils being identified as having SEN.

The key findings of the report suggest:

  • Black Caribbean and mixed white and black Caribbean pupils are twice as likely to be identified with social, emotional and mental health needs as white British pupils.
  • Asian pupils are half as likely to be identified with autistic spectrum disorders as white British pupils.
  • Indian and Chinese pupils are half as likely to be identified with moderate learning difficulties as white British pupils.

While similar research has been done in the US, it is the first time a study with this detail has been conducted in the UK.

Source: Ethnic disproportionality in the identification of special educational needs (SEN) in England: Extent, causes and consequences (December 2018), University of Oxford

Positive results for early years maths apps

A randomised controlled trial of two new maths apps to support young children’s early maths development has shown positive results. The apps, “Maths 3–5” and “Maths 4–6”, are based on core mathematical concepts in number and shape, and space and measure, which are covered in the Early Years Foundation Stage, and also start to introduce children to topics covered in Key Stage 1.

Laura Outhwaite and colleagues conducted the randomised controlled trial of the apps with 389 children aged 4–5 years from 12 schools in the UK. The trial took place over 12 weeks in the last weeks of their Reception school year before pupils moved to Key Stage 1. Pupils were randomised to either use the apps in addition to standard maths teaching activities (treatment); use the apps instead of a regular small group-based maths activity (time-equivalent treatment), or continue with usual maths teaching activities (control).

The results showed that pupils in the treatment group made more progress on standardised assessments of maths performance over 12 weeks than pupils in the control group (effect size = +0.31). Similarly, pupils in the time-equivalent treatment made more progress in maths performance than pupils in the control group (effect size = +0.21). There was no significant difference in maths performance between pupils in the two treatment groups (effect size = +0.08).

A randomised controlled trial of apps developed for primary children in Malawi, which we covered in a previous issue of Best Evidence in Brief, also showed positive results for maths achievement.

Source: Raising early achievement in math with interactive apps: A randomized control trial (February 2019), Journal of Educational Psychology, 111(2)

Test anxiety and performance in high-stakes testing

A study published in Contemporary Educational Psychology suggests that the relationship between test anxiety and performance in high-stakes tests is positive, but the relationship varies for pupils with different achievement levels.

Yao-Ting Sung and colleagues at the National Taiwan University used data from 1,931 Taiwanese ninth grade (Year 10) pupils from 37 schools. The Basic Competence Test (BCTEST) was used to benchmark their achievement. The BCTEST is a high-stakes test for Taiwan junior-high school pupils, determining to which high schools with different levels of prestige and tuition fees they will be admitted. Subjects in the test included Mandarin, English, mathematics, social studies, science and writing. Test anxiety was measured by the examination stress scale.

Findings include:

  • The overall relationship between text-anxiety and learning achievement in high-stakes testing was positive (r =+0.18).
  • Lower levels of test-anxiety were found among high-achievement and low-achievement pupils while higher levels of test-anxiety were found among moderate-achievement pupils.
  • For higher achievement pupils, the relationship between text-anxiety and achievement in high-stakes testing was found to be negative (r = -0.16), while for the group of pupils with lower achievement, a positive relationship was found (r= +0.22).

Source: Reexamining the relationship between test anxiety and learning achievement: An individual-differences perspective (July 2016), Contemporary Educational Pyschology, Volume 46