Summer learning loss in reading? Not necessarily
The fabled ‘summer learning loss’ may not be cause for concern, according to the results of two new studies in New Zealand.
By James Chapman
Concerns about a summer learning loss (SLL) have been expressed over many years. Atteberry and McMechan (2020) note that “there is a common understanding among policymakers, researchers, and practitioners that during the summer students lose some of the knowledge and skills acquired during the school year” (p. 4). Such concerns have become heightened, especially in the United States, because of the school closures due to the COVID-19 pandemic.
Notwithstanding the devastating effects of COVID-19 across all aspects of life, renewed discussion about summer learning loss warrants some further examination, especially in countries that have shorter school summer vacations (e.g., Australia, New Zealand, UK). In this article, I present data from two New Zealand studies that call into question whether there is a SLL in regard to reading. Neither of these studies was designed to examine summer learning loss. Rather, they provided data that addressed this issue as children transitioned from the end of Year 1 to the start of Year 2, with the summer vacation in between the transition.
New Zealand data on the issue of SLL are relatively sparse. McNaughton, Jesson, Kolose, and Kercher (2012) reported that it is “well known” that a summer learning effect occurs in New Zealand (p. 2). Similarly, Turner and Tse (2015) asserted that there is an SLL effect in New Zealand and implemented summer reading programmes to counter that effect.
Data for the first study were available from five schools in the wider Auckland region in New Zealand. All schools were classified as ‘low decile’ (deciles 1 to 3). Decile rankings are based on the predominant SES status of families in each school’s neighbourhood, with decile 1 indicating a very low SES neighbourhood and decile 10 a very high SES neighbourhood. Four of the five schools were supplementing their Year 1 reading programmes with the Quick60 (Iversen, 2013) programme, which is designed to teach the necessary early literacy skills in an explicit way. The other school used the regular literacy programme which was whole language in orientation.
Reading data were collected in November, towards the end of Year 1, and in February, at the start of Year 2. These data consisted of scores on the Burt Word Test and Reading Book Level, which is determined by means of a running record. The Burt test was administered by a research assistant, whereas book levels were assessed by classroom teachers. Scores on both measures were analysed by means of analyses of variance with repeated measures.
Scores on the Burt test revealed that children in the Quick60 schools (N= 61) increased from a mean of 18.98 points at the end of Year 1 to 23.31 in February of Year 2. Children in the regular literacy programme (N=24) had lower scores, but also increased from 11.88 to 15.41 over the same testing occasions.
Burt scores were also used to group readers as low (less than 11), average (11 to 21), and high (over 21). All three groups showed increases in mean Burt scores at the early Year 2 assessment occasion compared to the end of Year 1 scores: ‘low’ children increased by 1.31 points; ‘average’ children also increased by 1.31 points; ‘high’ children increased by 3.18 points.
For book levels, Quick60 children (N=47) increased from 10.85 to 11.71 over the two testing occasions, whereas children in the regular programme (N=25) showed a slight decrease from 6.34 to 6.04. These changes weren’t statistically significant.
Although boys scored lower than girls on average, both made similar changes on the two measures from the end of Year 1 to early in Year 2.
Of further interest are the findings in terms of home background. Classroom teachers were asked to rate each child’s home background as either ‘normal’ or ‘difficult’. Homes rated as ‘difficult’ involved issues known to teachers such as parental illness, unemployment, drug problems, and relatively high rates of school absenteeism.
Both groups showed increases in Burt scores: 19.33 to 22.67 for children from ‘normal’ backgrounds and 13.55 to 15.91 for children from ‘difficult’ backgrounds. In terms of book levels, changes were from 9.53 to 10.43, and 7.30 to 8.22 respectively for children from ‘normal’ and ‘difficult’ backgrounds. Clearly, children whose home backgrounds are rated by teachers as difficult were achieving at lower levels than those whose backgrounds were considered to be normal.
In sum, data from this study indicate that there is no evidence of a summer slide in reading either for children receiving the Quick60 programme or for those receiving the regular whole language-oriented programme during their first year of schooling. Similarly, there is no evidence in these data that children from ‘difficult’ home backgrounds or those whose word reading was comparatively low at the end of Year 1 suffer from a summer slide in reading performance.
The second study involved children who were participating in an intervention study funded by the New Zealand Ministry of Education (see Chapman et al., 2018a; Chapman et al., 2018b). All children turned five years old during the few months prior to entering school at the start of the school year in February. The study was undertaken in 39 schools in the lower North Island. Schools were randomly allocated to either an ‘intervention’ group or a ‘comparison’ group. The intervention comprised four one-day professional learning and development (PLD) workshops and one two-day workshop during the course of the year for those teachers working with Year 1 children. The workshops focused on providing teachers with the knowledge and skills to adopt explicit and systematic word-decoding teaching strategies in their literacy instruction. Teachers in comparison schools carried on with their regular literacy programme, which was typically whole language in nature. Attrition, the withdrawal of one school, and incomplete data reduced the number of students included in the various analyses.
To examine evidence for a summer reading loss, data from the Burt test collected in November of Year 1 were compared with Burt scores collected during February of Year 2. The Burt test was administered by trained research assistants. Book levels were not available. Complete data for the two testing occasions were available for 522 children.
There was an overall increase in mean Burt scores from 17.96 at the end of Year 1 to 19.94 early in Year 2. Mean score changes were similar for both the Intervention (N=270) and Comparison (N=252) children, with the Intervention children obtaining slightly higher gains than Comparison children: Intervention = 17.38 to 19.08; Comparison = 17.25 to 19.44. These results were combined for two separate cohorts of ‘intervention’ teachers.
The second cohort of teachers received a modified PLD workshop programme based on changes made to the programme for the first cohort. Results for children whose teachers were in the second cohort were better than for those in the first cohort: intervention (N=104) = 19.41 to 20.42; comparison (N=57) = 14.04 to 15.79. Of particular interest were results for Intervention children who were in the low band of Burt scores: Intervention = 6.57 to 7.43; Comparison = 4.24 to 4.92. During the course of Year 2, the Intervention children went on to outperform Comparison children on a range of measures (phonological awareness, alphabetic coding, language processing, word reading and spelling).
Consistent with the results for Study 1, children in each of the three decile bands of schools showed increases in Burt scores between the end of Year 1 and early in Year 2: low = 13.23 to 18.76; middle = 18.15 to 20.36; high = 20.93 to 22.55. The greater increase for children in low decile schools was due primarily to higher Burt scores obtained by children in the Intervention group in contrast to those in the Comparison group.
Although boys tended to obtain lower Burt scores than girls in both the Intervention and Comparison sample, roughly similar gains of around 1 to 1.5 score point differences between the two testing occasions were made for boys and girls.
The purpose of this article was to identify data from two studies to show whether or not there was evidence of a summer slump in reading performance in New Zealand children. Compared to other studies, these data do not reveal such a slump. Rather, there was a general tendency for children to increase word reading and reading book level scores between the end of Year 1 assessments in November, and the start of Year 2 assessments in February.
There is no obvious answer to the question as to why no summer slump in reading was found in these two data sets. Being part of an intervention study or not was not associated with a slump; being in a low decile school and/or having low Burt reading scores at the end of Year 1 was not associated with a slump; and being a boy (or a girl) was also not associated with a slump. Perhaps importantly, coming from a home background considered by teachers to be ‘difficult’ also was not associated with a slump. That said, there are ongoing disparities in reading achievement between children from low compared to high SES backgrounds. And in line with many other countries, boys tend to perform less well on reading assessments than girls.
It’s hard to believe that New Zealand children engage more with reading-related activities over the summer break than children in other countries. Consider that the summer break in Southern Hemisphere countries coincides with the Christmas vacation. In New Zealand, most likely in line with other southern countries, people typically engage in family time and holidays over the Christmas/New Year period. Perhaps literate cultural capital is enhanced for some children with home-based literacy activities and trips to the library. But not all children have access to such resources.
Further, it is unlikely that early childhood experiences in New Zealand provide a better literacy foundation for children prior to school entry than other countries, thereby mitigating the risk of a summer slump in reading. Systematic pre-reading literacy activities in most New Zealand pre-schools and kindergartens are discouraged in favour of informal play-based programmes with a holistic approach to curriculum planning. Policies and curriculum for early years in New Zealand do not favour explicit instruction in early reading-related skills.
Despite the lack of a ready explanation for the results of these two studies, summer reading ‘clinics’ can provide children with the opportunity to further enhance their skills in this area. And parents who are able to can also assist children further develop their literacy skills over the summer vacation. Children who do not have these opportunities have to rely on teachers to provide quality literacy instruction.
Over the last four decades most New Zealand teachers have adopted a whole language approach to literacy instruction, with a strong reliance on the three-cuing system of early word identification. That is how teachers have typically been trained in education colleges to teach reading.
Despite there not being an obvious summer slump in reading, much remains to be done in New Zealand in terms of adopting contemporary scientifically based approaches to literacy instruction in the early years of schooling. Significant changes are underway as a result of recent and current research to change the predominant, whole language approach to literacy instruction. Hopefully these changes will benefit all children and ensure that the results of the studies in this article showing no summer slump in reading are widespread and persist.
James Chapman is a Professor Emeritus in the College of Humanities & Social Sciences at Massey University, in Palmerston North, New Zealand. He has published over 150 journal articles, book chapters and books on learning disabilities, special education, literacy learning difficulties, early literacy development, reading intervention, and self-system factors in academic achievement.
This article appeared in the December 2020 edition of Nomanis.