1. Shi, Qingmin., Wang, Jian., Zhang, Shaoan. and Lin, Emily."Attitudes toward math, math self-efficacy, math value on math achievement: Mediation role of time on homework" Paper presented at the annual meeting of the 56th Annual Conference of the Comparative and International Education Society, Caribe Hilton, San Juan, Puerto Rico, Apr 22, 2012 <Not Available>. 2019-05-23 <http://citation.allacademic.com/meta/p556639_index.html>

Publication Type: Conference Paper/Unpublished Manuscript Review Method: Peer Reviewed Abstract: Attitudes Toward Math, Math Self-Efficacy, Math Value on Math Achievement:
Mediation Role of Time on Math Homework

The purpose of the current study was to determine the causal relationships among attitudes toward math (ATTM), math self-efficacy (MathSE), value on math (VOM), time on homework (TimeonHW), and math achievement (Math-ACH) across Hong Kong and U.S. samples drawing on the randomly selected 449 participants from Hong Kong and 449 from U.S. eighth-grade students who participated in TIMSS 2007 assessment. The following hypotheses examined were as follows: 1) Student attitudes toward math, math self-efficacy, value on math have positive effects on student achievement in math across Hong Kong and U.S. models. 2) Student attitudes toward math, math self-efficacy, value on math have effects on time students spent on homework; and in turn, time on math homework positively influences their math achievement across Hong Kong and U.S. models. Totally 13 questions related to student ATTM, MathSE, VOM, as well as TimeonHW in TIMSS, 2007 student questionnaire (TIMSS, 2007) were selected as independent variables and student math achievement scores estimated by the five plausible values were the observed dependent variable.
The IEA International Database Analyzer software (IDB, IEA, 2009) was used for data merging and initial analysis (Olson et al., 2008). The hypothesized models were tested with the EQS 6.1 program (Bentler, 2003) by examining the structure of direct and indirect effects on the two sample data for the current study. The initial data screening showed using EQS 6.1 indicated that the normalized estimate for two samples both larger than 5 (Bentler, 2005), a value employed for estimating whether the data are non-normally distributed. Therefore, the maximum likelihood Robust estimation was appropriate for this study and model fits were estimated according to the Satorra-Bentler Scaled chi-square (S-Bχ2), comparative fit index (CFI), non-normed fit index/Tucker-Lewis index (NNFI), and root mean-square error of approximation (RMSEA).
The results of this study indicated that the CFA model and the structural model fit the data well with modification for Hong Kong model. The hypothesized models were tested for both samples starting at the baseline model. The full SEM model fit the data well for the Hong Kong sample, S-Bχ2(69, N = 449) = 132.25, p < .001, CFI = .97, NNFI = .96, RMSEA = .045 (CI: .033 ~ .057), and S-Bχ2(67, N = 449) = 160.43, p < .001, CFI = .96, NNFI = .95, RMSEA = .056 (CI: .045 ~ .067) for the U.S. sample. The results of the current study regarding the effects of ATTM, MathSE, and VOM on TimeonHW, and in turn, on Math-ACH across Hong Kong and U.S. samples indicated that effects existed for the two groups. MathSE has a significantly negative effect on TimeonHW for Hong Kong sample, while a significantly positive effect on TimeonHW for U.S. sample. ATTM and VOM have a positive but nonsignificant effect on Math-ACH for Hong Kong sample, but a negative but nonsignificant effect for U.S. sample. Similarly, TimeonHW for Hong Kong sample has a significantly positive effect on Math-ACH, but a significantly negative effect for U.S. sample. That is, for Hong Kong students who reported spending more time on math homework have higher Math-ACH. For U.S. students, however, who reported more time spent on math homework received lower Math-ACH.
This study was significant in several ways. First, it adds to the literature and supports previous studies about the negative effects of attitudes and homework on U.S. student math achievement, which raises the concern about practices that intend to improve student achievement through assigning student more homework or developing their positive attitudes toward math. Second, it shows that the causal relationships between attitudes, self-efficacy and achievement may be mediated by other factors, such as time on homework or other social contexts. Therefore, it calls for more carefully designed studies to explore such a complex relationship.
References
Bentler, P. M. (2003). EQS 6 Structural Equations Program. Encino, CA: Multivariate Software,
Inc.
Bentler, P. M. (2005). EQS 6 Structural Equations Program Manual. Encino, CA: Multivariate
Software, Inc.
IEA (2009). International database analyzer (version 2.0). Hamburg, Germany: IEA Data Processing and Research Center.
Olson, J., Martin, M., & Mullis, I. (Eds.). (2008). TIMSS 2007 technical report. Chestnut Hill, MA: TIMSS & PIRLS International Study Center, Boston College.
The Third International Math and Science Study (TIMSS, 2007). TIMSS contextual background questionnaires: student questionnaire—math and science (Integrated Science Version). Retrieved September 10, 2010, from http://timss.bc.edu/TIMSS2007/PDF/T07_StudentQ_IS_G8.pdf.

2. Germeroth, Carrie., Clements, Douglas., Sarama, Julie., Layzer, Carolyn., Unlu, Fatih. and Fesler, Lily."Math and math + scaffolded play interventions: Analyses of main effects on development of math competence and executive function" Paper presented at the annual meeting of the SRCD Biennial Meeting, Pennsylvania Convention Center and the Philadelphia Marriott Downtown Hotel, Philadelphia, PA, <Not Available>. 2019-05-23 <http://citation.allacademic.com/meta/p961470_index.html>

Publication Type: Presentation Abstract: We evaluated two interventions, Building Blocks math (BB) and the BB synthesized with the Tools of the Mind scaffolding to promote self-regulation (BBSR) (scaffolding of play but also scaffolding integrated into the BB activities), compared to a business-as-usual control (BAU), using a three-armed cluster randomized trial and HLM with 826 children in 84, 4-year-old classrooms across three districts (multi-racial/multi-ethnic, low-income, 27% ELL). Schools were randomly assigned to three conditions using a systematic circular sampling scheme (Lahiri, 1951).

Late pretests complicated analyses and ultimately reduced power. Following Schochet (2008), we compared the magnitude and precision of impact estimates from three estimators, posttest-only, difference-in-differences (DID), and ANCOVA. Unlu, Layzer, Clements, Sarama, Fesler, and Cook (2014) showed that for most outcome measures and both time points, the posttest only estimator yielded impact estimates that had smaller mean-squared error; therefore, we decided to use it for the calculation of the final impact estimates.
Thus, we calculated the BB and BBSR impacts on the achievement measures collected in Spring 2011 using multivariate 2-level HLMs that cluster students in the schools/centers but not conditioning on the late pretest measures. Table 1 presents the corresponding impact estimates, which are expressed in effect sizes using the standard deviation of the BAU group.
Table 2 shows that although most of the impacts for the BB students (BB vs BAU) are positive and larger than 0.1 standard deviations, only one impact attains statistical significance at the p<0.1 level (backward digit, ES: 0.19). The pattern in the impacts for BBSR students is somewhat mixed, with some estimates being positive and some negative. For this group, the only impact estimate that achieves significance at the p<0.1 level is the sentence length measure from the Bus Story assessment (-0.24 SD). Overall, these results do not make a strong case for students being positively affected by either intervention condition, especially not the BBSR, through the end of PreK, the only year when the interventions were implemented.
Follow-up measures were taken at the end of the Kindergarten year (Spring 2012), during which students were exposed only to BAU. Table 2 presents the corresponding results. Results presented for the BB vs. BAU contrast suggests that the differences between the two groups on Spring 2012 measures are larger for almost all measures than those in Spring 2011. In particular, BB impacts on TEAM measures are around 0.2 standard deviations; the effect size for the scaled score is 0.19 and statistically significant at the p<0.05 level. Other statistically significant impact estimates are on Forward Digit Score (effect size=0.2, significant at p<0.05) and Pencil Score (effect size=0.16, significant at p<0.1). Other impact estimates are positive between 0.1 and 0.19 of SD but they are not statistically significant. Compared to the Spring 2011 results, estimates for the impact of BBSR are somewhat larger but none of them reach statistical significance.

The results suggest that implementing mathematics curricula alone may have the dual benefit of teaching an important content area and developing at least some EF competencies.

3. Hur, JinHee. and Son, Claire Seung-Hee."Math is Everywhere: The Nature of Home Math Talk and Its Impact on Preschoolers’ Math Skills in Low-Income Families" Paper presented at the annual meeting of the SRCD Biennial Meeting, Pennsylvania Convention Center and the Philadelphia Marriott Downtown Hotel, Philadelphia, PA, Mar 19, 2015 <Not Available>. 2019-05-23 <http://citation.allacademic.com/meta/p958916_index.html>

Publication Type: Individual Poster Review Method: Peer Reviewed Abstract: Introduction: Development of mathematical skills emerges during early childhood (Ginsberg & Pappas, 2004). Early mathematical skills at kindergarten entry are a stronger predictor of later academic achievement (Duncan et al., 2007) and home math interactions are an influential factor promoting early mathematical skills. However, children from low-SES families tend to engage in math interactions less frequently than those from high-SES families (Denton & West, 2002; Vandermaas-Peeler et al., 2009). Given the growing academic gap among children from families with differing SES (National Assessment of Educational Progress, 2011), it is important to examine the development of early mathematical skills of low-income preschoolers and their home experiences. Thus, the current study examines the nature of math interactions at the homes of low-income preschoolers and their associations with early mathematics skills and family characteristics.
Method: Participants include 46 English-speaking low-income Head Start children (age 4-5) and their mothers. At the beginning of the school year, we visited their homes to observe mother-child interactions during cupcake baking as a context to observe home math interactions (Vandermaas-Peeler et al., 2009). The interactions were coded at the utterance level for various types of math talk and task-related talk. Children’s early math skills were tested at school using WJ-III Applied Problems at the beginning and the end of the school year and children’s verbal abilities were tested using WJ-III Picture Vocabulary at the beginning of the year.
Results: Participating mothers provided math talk less often than task-related talk but with great variability. The most frequent math talk was number talk (M = 4.43, SD = 7.06), followed by measurement talk (M = 1.78, SD = 3.51), and the least frequent talk was operation talk (M = .07, SD = .25).
To examine associations between maternal math talk and child math skills, we ran structural equation modeling with STATA predicting children’s fall and spring math skills (Figure 1). Mothers’ number talk significantly predicted children’s fall math skills while mothers’ measurement talk predicted children’s spring math skills.
We ran a hierarchical regression analysis to investigate predictors of mothers’ math talk (Table 1). Mothers with higher education used more math operation talk and employed mothers generally used more math talk, especially measurement talk. Mothers’ measurement talk and total math talk were significantly related to their use of task explaining talk during the activity; mothers’ number talk was related to mothers' use of task organization talk. Interestingly, children’s initial math and verbal skills did not predict any type of maternal math talk.
Implications: Current study showed that mothers’ use of number talk is related to children’s current math skills while mothers’ measurement talk is related to children’s later math skills. Our results suggest that training parents to use math talk could be an intervention targeting children’s math skills. The intervention may need to consider encouraging parents to be more involved in everyday activity with children and to find instances to provide math talk, not just about counting or operation but also about more inferential and applied skills, such as measurement talk.

2018 - Comparative and International Education Society Conference

4. Lutfeali, Shirin."Numeracy boost: What do the results from an early grades math program say about children’s foundational math skills?" Paper presented at the annual meeting of the Comparative and International Education Society Conference, Hilton Mexico City Reforma Hotel, Mexico City, Mexico, <Not Available>. 2019-05-23 <http://citation.allacademic.com/meta/p1353279_index.html>

Publication Type: Panel Paper Abstract: Purpose: The purpose of this paper is to present math outcomes and trends in Save the Children’s early grades math initiative, Numeracy Boost, across five countries: Egypt, Ethiopia, Pakistan, Malawi and Bangladesh.
Background: Numeracy Boost is Save the Children’s innovative, research-based toolkit to support the development of math skills in young children. Recent research has shown that early exposure to math concepts and activities positively impacts later school achievement (Jordan 2007). In fact, early math knowledge is an even greater predictor of later academic success than early literacy abilities (Duncan 2007). More broadly, mastery of mathematics skills and concepts is essential for children and adults to function in communities, work and daily life.
The program was piloted in the 2012-2013 school year in Malawi and Bangladesh, and has now grown to include Egypt, Ethiopia, and Pakistan. Numeracy Boost supports children’s learning at three different levels: the student, the teacher and the community. It aims to not only strengthen children’s basic math skills in the early grades, but also to illustrate that math is useful and has relevance to everyday life.
Design/Methodology/Approach: The Numeracy Boost assessment is modeled on the Early Grades Math Assessment (EGMA) and children participating in the Numeracy Boost approach are assessed at baseline and endline. The assessment tests children’s knowledge and skills in three core domains: Number and Operations, Geometry and Measurement. Some of the sub-tasks are timed. The assessment also includes a Home Numeracy Background section, which asks about the student’s use of and exposure to math outside of school to provide information about how non-school related activities contribute to learning outcomes.
Using data collected from the Numeracy Boost assessments over the past five years in five countries, this presentation will present trends in early grades math learning outcomes. The presentation will highlight trends related to gender, equity, and whether certain skills are ‘linked’, like the ability to calculate quickly and successfully solving multi-step word problems. The presentation will also share project implementers’ perspectives on the research and trends and their implications for programming.
Results: Numeracy Boost has been implemented in over five countries, including humanitarian settings since 2012. In each of these countries, children participating in the Numeracy Boost intervention have outperformed peers in control schools, with results being statistically significant. The Numeracy Boost intervention has shown to support girls and those at the bottom of the socio-economic spectrum.

5. Libertus, Melissa., Odic, Darko., Feigenson, Lisa. and Halberda, Justin."Verbal number estimation predicts math ability and mediates the relation between numerical approximation and math ability" Paper presented at the annual meeting of the SRCD Biennial Meeting, Pennsylvania Convention Center and the Philadelphia Marriott Downtown Hotel, Philadelphia, PA, <Not Available>. 2019-05-23 <http://citation.allacademic.com/meta/p958761_index.html>

Publication Type: Presentation Abstract: An increasing body of research has demonstrated a small but significant link between children’s math abilities and their non-verbal approximate number discrimination abilities (i.e., children’s Approximate Number System; ANS). In the preschool years, this relation appears specific to certain types of mathematical processing: ANS precision is consistently linked with informal math skills, that are learned before or outside of school and do not require an understanding of formal mathematical conventions (e.g., counting, informal calculation using fingers or tokens), but appears uncorrelated with formal math skills (e.g., numeral literacy, written arithmetic, number fact retrieval; Libertus, Feigenson & Halberda, 2013). However, few studies have investigated the link between children’s math abilities and their number estimation abilities (e.g., saying approximately how many briefly flashed dots were on the screen). The precision of number estimation depends in part on the precision of the ANS, but also additionally on the precision of the mapping between the ANS and number words (often measured as coefficient of variance). The goals of the present study were twofold: We wanted to 1) assess whether number estimation mediates the link between the ANS and math abilities, and 2) investigate the degree to which the ANS and number estimation are linked to informal and formal math abilities in kindergarten and early elementary school. To this end, 5- to 8-year-old children (N = 51) completed an ANS comparison task (judging whether there were more blue or yellow dots in a display), a number estimation task (stating the approximate number of dots in briefly flashed arrays), and a standardized math test covering informal and formal mathematics (Test of Early Mathematics Ability; Ginsburg & Baroody, 2003).

We found significant correlations between ANS precision, estimation ability, and overall math ability (all ps < .05, see Figure 1). Importantly, the correlation between estimation and overall math ability remained significant when controlling for ANS precision, suggesting that verbal estimation ability accounts for additional variance in math ability, beyond that accounted for by ANS precision. Moreover, using bootstrapping to test the significance of the indirect pathway between ANS and overall math ability via estimation ability, the direct path between ANS acuity and overall math ability was no longer significant when the indirect path via number estimation was added (p = .08) and the 95% confidence interval ranged from -44.41 to -0.67 suggesting a significant mediation effect (p < .05).

Finally, ANS precision was correlated with informal (R2 = .09, p < .05) but not formal math abilities (R2 = .03, p = .23), whereas verbal estimation ability was correlated with both informal (R2 = .11, p < .05) and formal math abilities (R2 = .09, p < .05, see Figure 2). These findings demonstrate that the ANS and number estimation play unique roles for math ability, and that the precision of the mapping between the ANS and number words is a critical component for formal math abilities.