Raising Early Achievement in Math With Interactive Apps: A Randomized Control Trial.

Math Development

Math development requires the acquisition of different component skills and processes that range in level of difficulty ( Goswami, 2006 ; Holmes & Dowker, 2013 ). Components of math knowledge can be grouped into four broad categories: factual knowledge (e.g., number bond combinations and properties of shape and patterns) and conceptual understanding (e.g., identifying and applying mathematical procedures), which both encompass basic math skills. In contrast, mathematical reasoning (e.g., making deductions and inferences from mathematical information), and problem solving (e.g., combining and applying different areas of mathematics to solve a problem in a specific context) reflect higher-level mathematical skills that require the application of basic math knowledge to find a solution ( Rutherford-Becker & Vanderwood, 2009 ; Thurber, Shinn, & Smolkowski, 2002 ). Research shows the acquisition and automatization of basic math skills facilitates higher-level mathematical development ( Codding, Archer, & Connell, 2010 ; Gersten & Chard, 1999 ; Mayfield & Chase, 2002 ; VanDerHeyden & Burns, 2005 ). For example, longitudinal research shows early factual and conceptual knowledge, including, number bond combinations, counting, pattern knowledge, and calculation ability predict more complex skills, such as problem solving later in development ( Björn, Aunola, & Nurmi, 2016 ; Fuchs et al., 2006 ; Rittle-Johnson, Fyfe, Hofer, & Farran, 2016 ). In contrast, poor fluency in basic math skills is shown to be commonly associated with mathematical difficulties ( Geary, 1993 ; Jordan, Hanich, & Kaplan, 2003 ). Together, this evidence supports theories of cumulative learning that propose the mastery of basic math facts and concepts are an essential foundation for the acquisition of more complex math skills ( Gagné, 1968 ). It emphasizes the vital, foundational role of strong basic math skills in successful mathematical development.

However, many children struggle to acquire basic math facts and conceptual knowledge ( Geary, 2011b ), which makes them vulnerable to persistent underachievement throughout their education ( Duncan et al., 2007 ; Jordan, Kaplan, Ramineni, & Locuniak, 2009 ). Efficiency in basic math skills can be facilitated through targeted practice, which emphasizes task repetition for effective skill acquisition ( Daly, Martens, Barnett, Witt, & Olson, 2007 ; Imbo & Vandierendonck, 2008 ) and direct instruction ( Chodura, Kuhn, & Holling, 2015 ; Kroesbergen & Van Luit, 2003 ; Swanson & Hoskyn, 1999 ), which is characterized by deliberately sequenced small units of information taught explicitly ( Kirschner, Sweller, & Clark, 2006 ). Intervention studies show that individualized training that places the child at the center of their learning with learning activities that incorporate targeted practice and direct instruction can enhance the development of targeted basic math knowledge and generalize to other more complex math components not included in the intervention ( Fuchs et al., 2009 ; Kidd et al., 2013 ; van der Ven, Segers, Takashima, & Verhoeven, 2017 ). This evidence provides further support for theories of cumulative learning ( Gagné, 1968 ) and the importance of developing a strong foundation in basic math skills. It also suggests well-designed individualized early interventions that include targeted practice and direct instruction are needed to provide all children with the necessary learning opportunities to develop a strong foundation in math ( Stacy, Cartwright, Arwood, Canfield, & Kloos, 2017 ). Such approaches may be particularly beneficial in the first years of schooling ( Clements, Baroody, & Samara, 2013 ) when children show the fastest rates of math development ( Hill, Bloom, Black, & Lipsey, 2008 ).

App Technology

Educational math apps delivered on touch-screen tablets offer an opportunity for individualized math practice targeted to children’s needs. Apps that are grounded in learning science theory ( Hirsh-Pasek et al., 2015 ) and incorporate the principles of universal design and play ( Burgstahler, 2012 ) can provide a blended learning approach ( Naismith, Lonsdale, Vavoula, & Sharples, 2004 ). Specifically, apps that embody the principles of active, engaged, meaningful, and socially interactive learning with a specific learning goal ( Hirsh-Pasek et al., 2015 ) can combine benefits of direct instruction ( Kirschner et al., 2006 ), for example, feedback, repetition, and rewards with features of free play ( Gray, 2015 ) particularly, self-regulation and control. This can help provide an efficient child-centered but scaffolded learning environment ( Mayer, 2004 ; Mayo, 2009 ) tailored to individual needs ( Slavin & Lake, 2008 ) enabling individualized and structured instruction ( Gulliford & Miller, 2015 ) without additional, time-consuming, teaching demands ( Hilton, 2016 ; Kucian et al., 2011 ).

Educational apps delivered on touch-screen tablets are also particularly suited for young children, because they typically find them motivating ( Flewitt, Messer, & Kucirkova, 2015 ) and intuitive to use ( Cooper, 2005 ). Touch-screen tablets are mobile, light weight, and do not rely on dexterity-based motor skills that are needed to use a computer keyboard or mouse ( Kucirkova, 2014 ). Furthermore, access to mobile devices in educational settings is increasing. For example, in the United Kingdom, 70% of elementary schools have access to touch-screen tablets ( Clarke, 2014 ).

Previous Research

Despite the prevalence, popularity, and potential benefits of using app technology to support math development, the current evidence-base is fragmented ( Haßler, Major, & Hennessy, 2015 ) and suffers from a paucity of rigorous scientific investigations ( Cheung & Slavin, 2013 ). Concerns have also been raised about the impact of technology based screen time on early child learning and development ( Greenfield, 2015 ; Palmer, 2007 ; Sigman, 2012 ). To evaluate the impact of app technology in educational settings, practical and high-quality research is needed ( Cheung & Slavin, 2013 ) and should focus on the quality of the educational app content ( Blum-Ross & Livingstone, 2016 ; Falloon, 2013 ).

Emerging experimental evidence demonstrates the effectiveness of different high-quality math apps with early years pupils in a classroom setting ( Outhwaite, Gulliford, & Pitchford, 2017 ; Pitchford, 2015 ; Schacter & Jo, 2016 , 2017 ; van der Ven et al., 2017 ) and increasing time spent on learning math through using educational apps at home positively benefits children’s achievement in school ( Berkowitz et al., 2015 ). All of the math apps evaluated in these studies were grounded in evidence based learning theory, embodying the principles of active, engaged, meaningful, and socially interactive learning with a specific learning goal ( Hirsh-Pasek et al., 2015 ). Common features in these apps include explicit instruction, repetitive and cumulative training in mathematical concepts, immediate feedback, challenge and early reward, and individualized, self-paced learning, which are important components of effective math interventions ( Baker, Gersten, & Lee, 2002 ; Fuchs et al., 2008 ; Gersten et al., 2009 ).

Effectiveness and Implementation

First, there is a need to understand how educational apps are best implemented in a classroom setting. Typically, researchers have implemented app-based interventions (e.g., Schacter & Jo, 2017 ) as a supplementary teaching aid, in addition to standard math practice (e.g., Berkowitz et al., 2015 ). This entails greater instructional time on learning mathematics compared with the comparison groups, rendering it difficult to disentangle the effects of the intervention from the effects of extra time learning math ( Foster, Anthony, Clements, Sarama, & Williams, 2016 ; Ginsburg & Smith, 2016 ). To address this threat, experimental study designs need to include a time-equivalent control group (e.g., Holmes & Dowker, 2013 ). Furthermore, it is critical that teachers implement the intervention, to ensure high ecological validity and to support the generalizability of the intervention beyond the research context ( Clements, Sarama, Wolfe, & Spitler, 2015 ). To address this issue in the current study, two forms of teacher-based implementation of an educational math app intervention were compared to standard math practice. As illustrated in Table 1 , children in Group 1 (treatment) used the math apps in addition to all other standard math activities and so had increased exposure to math instruction. In contrast, children in Group 2 (time-equivalent treatment) used the math apps instead of a daily small group-based math activity, so time spent learning math was equivalent to the children in Group 3 (control) receiving standard teacher-led math instruction that included a daily small group-based activity. Thus, in the current study, all children received whole class math instruction delivered by the teacher, which was embedded into play-based learning, as is standard practice for early years classrooms in the United Kingdom.

In summary, this study asked, do children make more progress when the math app intervention is implemented by teachers in addition to regular math instruction (Group 1) or when implemented instead of a daily small group-based math activity (Group 2) compared with children receiving standard instructional practice (Group 3)? Based on previous research ( Outhwaite et al., 2017 ; Pitchford, 2015 ), it was predicted that children who used the math apps (Group 1 and Group 2) would progress more than children receiving standard math instructional practice (Group 3), and children who received the math apps in addition to their regular math instruction (Group 1) would have the strongest learning gains.

Components of Math Development

Second, there is a need to examine which components of math development are supported by educational apps. Previous research evaluating app interventions has frequently used assessments closely aligned with the intervention content, which typically focuses on specific aspects of math knowledge (e.g., Schacter & Jo, 2017 ). Studies are required to take a broader view of mathematics and consider how educational apps support the acquisition of targeted components of math knowledge and whether this generalizes to higher-level skills. This will help elucidate how math development is supported by interactive, individualized, educational apps. To address this, the math apps evaluated in this current study primarily targeted basic math facts and concept knowledge (see Table 2 ) and a standardized assessment of early mathematical skills that comprised measures of the four components of early mathematical development outlined above was given to all children in the trial, before and after the intervention period. This enabled learning gains for each mathematical component, including targeted basic skills and higher-level knowledge not included in the intervention to be compared across the three intervention groups. Therefore, this study also asked, for each of the four components of math development, do children make more progress with the apps when used in addition to regular math instruction (Group 1) or when implemented instead of a daily small group-based math activity (Group 2) compared with children receiving standard instructional practice (Group 3)?

Participants

The CONSORT (2010) data in Table 3 summarizes the study sample at each stage of the RCT. In total, 461 children aged 4–5 years were randomly allocated to one of the three groups. There were 153 children assigned to Group 1 (treatment) and received the math app intervention as well as all daily standard math practices. There were 152 children randomly allocated to Group 2 (time-equivalent treatment) and used the math app intervention instead of a daily small group-based math activity that is given as part of standard math practice. The remaining 156 children were assigned to Group 3 (control) and received regular math teaching practice only.

There were 452 children from the 12 schools pretested on the PTM5 ( Math Assessment Resource Service, 2015 ). Nine children were absent at pretest but were still randomized to group. Of the 452 children that were pretested, 389 children from 11 schools were available at posttest and were given the same math assessment immediately after the 12-week intervention period. In total, 63 children who were pretested did not complete the posttest; one child left school, two children were removed from the study by their teachers for reasons unknown, and 60 children were absent at posttest, including 30 children from one school because of a school fieldtrip. It was not possible to follow-up on children that were absent on the day of the posttest because posttesting took place during the last week of the school year. Table 4 details descriptive data for the final sample of 389 children.

Math App Intervention

The intervention consisted of two math apps, “Maths 3–5” and “Maths 4–6,” developed by onebillion , an educational not-for-profit organization ( www.onebillion.org.uk ). These math apps are based on core mathematical concepts in Number and Shape, and Space and Measure, covered in the Early Years Foundation Stage (EYFS) Profile ( Department for Education, 2013 ; see Table 2 ). The apps also start to introduce children to topics included in the U.K. National Primary Curriculum for Key Stage I ( Department for Education, 2014 ). The apps primarily target factual knowledge and basic conceptual understanding, for example, simple numerical operations, such as addition and subtraction. Table 2 details the topics covered in each app and how the app content maps onto the math curriculum and the components of math development.

Features of the math apps and how they map onto the principles of active, engaged, meaningful, and socially interactive learning are discussed in detail above. Overall, the apps are designed to deliver child-centered tuition through interactive picture, audio, and animation formats with clear objectives, instructions, and immediate formative feedback, consistent for all users. Children work through the apps individually with headphones, at their own pace, and have the opportunity to repeat instructions and activities as often as needed. To complete a topic, children need to achieve 100% pass rate on an end of topic quiz included in the software. The quizzes are designed to assess children’s knowledge of the mathematical concepts covered in the topic activities.

For example, in topic 1 in Maths 3–5 children are taught the concepts of sorting and matching through a range of activities involving sorting and matching different items by type, shape, size, and color. Screenshots of example activity items and task instructions for Topic 1 are displayed in Figure 1 (courtesy of onebillion ). After completing seven sets of activities, children reach the end of topic quiz that includes 10 questions from the previous activities. When children pass the quiz they are awarded a certificate and progress to the next topic.

As shown in Table 1 children in Group 1 (treatment) and Group 2 (time-equivalent treatment) received the daily math app intervention for 30 min each day over the 12-week intervention period.

Small Group Math Instruction

Small group math instruction was consistent with the Number and Shape, and Space and Measure content covered in the EYFS Profile ( Department for Education, 2013 ; see Table 2 ). Group-based activities were delivered by the class teacher and focused on a particular mathematical concept from the EYFS Profile. Example activities were obtained through observations made during school visits by the first author. For example, shape recognition was taught by the teacher drawing different shapes on the whiteboard and asking the small group of children, “what shape am I?” Children responded by calling out the answer and receiving corrective feedback from the teacher before moving onto the next item. In an activity focused on understanding the concepts of more and less, the teacher utilized a number line visual aid and physically demonstrated “1 more than 18.” The teacher then asked the small group of children, “what is 1 more than 10?” and children responded by writing their answer on an individual mini whiteboard and showing it to the teacher. The teacher would then give corrective feedback before moving onto the next item. As highlighted in Table 1 , children in Group 1 (treatment) and children in Group 3 (control) received instruction through daily small group math activities as part of standard math practice.