Parents’ Involvement in High-Stakes Language Assessment: A Review of Test Impact on Parent Behavior

This paper reviews a total of 20 empirical research studies concerning parents’ behavior under the context of high-stakes language assessment, aiming to reveal the impact of the assessment on parents’ behavior. The results show that (1) parents are typically involved in high-stakes language assessment process; (2) their involvement practice includes general (such as hiring tutors for children) and extreme involvement behavior (such as participating in movement against high-stakes testing); (3) no unanimous conclusion is reached concerning the effectiveness of parents’ involvement in high-stakes language assessment; (4) multiple factors that affect parents’ involvement in high-stakes language assessment are identified, including parents’ perceptions of tests, their educational background, and the time they spend with their children. This study concludes that tests might influence the ways parents are involved in children’s education. However, not all parents might be influenced by testing, and testing might have a positive impact on some parents but a negative impact on others. This synthesis has several practical implications. Firstly, it indicates that parents’ involvement behavior in the context of high-stakes language assessment deserves to be further investigated. Secondly, it points that various intervention programs should be provided for parents to help them better support their children’s learning and test preparation. The paper also offers several suggestions for future research.


Introduction
A wealth of research has shown that parents are typically engaged in their children's learning. This behavior is called parent academic involvement or parents' involvement (PI) in children's education, which is defined as "parents' interactions with schools and with their children to promote academic success" (Hill & Tyson, 2009, p. 741). Mounting evidence suggests that PI is effective for advancing student academic achievement, and the effect sizes are small to moderate (e.g., Fan & Chen, 2001;Hill & Tyson, 2009). Although many researchers are committed to exploring the factors that affect PI (e.g., Hoover-Dempsey & Sandler, 1995, 1997, few have treated assessment as an essential variable. It seems that PI has been widely investigated within the non-assessment field (Cheng, Andrews, & Yu, 2010;Dawadi, 2019).
Similarly, in the field of language assessment, studies investigating the impact of assessment on parents are scarce, specifically in what ways parents help their children to be successful in high-stakes assessments (Cheng et al., 2010;Tsagari & Cheng, 2017). To date, most of the existing research focused on parent perceptions of language tests (e.g., Desforges, Hughes, & Holden, 1994;Scott, 2007;Vandergrift, 2015) rather than parent behavior that is likely to be affected by assessments.
It is essential to explore test (Note 1) impact on PI behavior (Cheng et al., 2010;Dawadi, 2019) for several reasons. First, not only are parents one of the key test stakeholders (Scott, 2007), but they know their children best (Martínez, Martínez, R., & Pérez, 2004;Rutland & Hall, 2013). Second, tests might have an impact on PI in assessment-the behavior parents adopt to improve their children's academic achievement (Cheng et al., 2010). Third, it is argued that PI in language assessment might help children succeed in the tests (Reta, 2017). Yet, scarce empirical evidence has been provided to support the prediction that tests affect PI in assessment, which in turn impacts students' academic attainment.
Despite the importance of the topic, however, no systematic review of the literature has been conducted concerning test impact on PI. The review by Harris (2015) was the only one that can be found until now, but her study focused on parents' understandings of and attitudes towards testing rather than parent behavior. This research attempts to fill this void by conducting a synthesis of test impact on parent behavior, interpreting parents' involvement behavior under the context of high-stakes language assessment. In so doing, the following research questions are addressed: RQ1: Are parents involved in the assessment process? If yes, how are they involved? RQ2: Is parents' involvement behavior in the assessment process effective for promoting students' academic achievement? If yes, what is the effect size? RQ3: What factors affect parents' involvement behavior in the assessment process?
This synthesis begins by describing the procedure of searching, screening, and analyzing literature. Then, it reports the findings and discusses the test impact on parents' behavior. Finally, implications are expounded.

Methods
This research synthesis conducted a literature search in major electronic databases such as the Educational Resources Information Center (ERIC) and Google Scholar, using the following keywords: parent(s) and assessment(s)/ test(s)/testing/ high-stakes test(s)/ high-stakes testing, the impact/ consequence/ effect of high-stake assessment(s) /test(s)/ testing on parent(s), involving parents in children's assessment(s) /test(s)/ testing, and parents' involvement in assessment(s)/ test(s)/ testing/ high-stakes test(s). Finally, 33 research studies were identified, which were carefully reviewed and screened based on the following criteria.
• This synthesis mainly focused on high-stakes language assessments, particularly English tests. Given the small number of studies specifically centering on the impact of high-stakes language assessment on parents, this synthesis also considered the studies investigating the entire high-stakes assessment program that includes language assessment; • Due to the limited number of empirical research, this synthesis was not limited to peer-reviewed publications (Harris, 2015). However, in order to guarantee the reliability of the results, this synthesis only considered the studies with rigorous research design, which means that the studies reported how data were collected and analyzed, the research findings were consistent with the data collected, and the conclusions were aligned with the research findings (Henderson & Mapp, 2002); • For the sake of focus, this synthesis did not include the assessment designed for special education. Finally, 20 studies were included and they represented the following three categories: • Research on parents' general involvement behavior (N = 8); • Research on parents' extreme involvement behavior (N = 7); • Research on parent intervention programs (N = 5). These studies are listed in Tables 1-3. The main information retrieved from these studies included macro-and micro-context (educational background and target tests), research type (qualitative or quantitative; longitudinal or cross-sectional), methods of data collection and analysis, parent participants, sample size, research purpose, main findings, the content of the parent intervention program, effects and effect sizes reported.  (1) Most of the Opt-Out movement participants were parents of public-school students.
(2) Social media played a major role in mobilizing participants. (3) The reason for participating in the movement was not just their opposition to high-stakes standardized tests. Some of the parents disagreed with teaching to the test and judged teachers' performance by using students' scores on high-stakes tests. (1) Parents viewed the policies of using high-stakes testing to sort students, evaluate schools and teachers as "government overreach and the corporatization of public schools" (p.8).
(2) They thought the educational policies were injustice to students and teachers. (3)   To investigate the effect of the program on (1) parents' knowledge of educational system and testing; (2) parent involvement in children's education; (3) students' academic achievement (1) The program helped parents better understand the educational system and testing, which resulted in greater PI.
(2) Intermediate school students' parents who participated in the program had children who significantly increased from Basic to Proficient levels in math, language arts, and science, whereas parents who did not participate in the program had children who remained at a Basic level in all subjects.
Not reported.

Results
The next section presents the results of the analysis, which shows parents' involvement behavior under the context of high-stakes language assessment. The magnitudes of associations between constructs and effect sizes are reported based on Cohen's (1988) benchmark.

The Answer to RQ 1
RQ1 asks "are parents involved in the assessment process? If yes, how are they involved?" Tables 1 and 2 illustrate the studies investigating parents' involvement in the assessment process. All of the 20 studies provided empirical evidence for parents involved in testing, indicating that parents do involve in the high-stakes language assessment process and they are involved in two ways: general and extreme involvement behavior.
Except for these general PI practices, some studies reported strategies parents specifically employed to help children prepare for the test. Namely, without a test, parents might not take these strategies, which thus can be named parent involvement in test preparation (PI_TP). A range of PI_TP strategies are identified in the existing literature: • Purchasing test-related materials, such as examination guidance books (Barksdale-Ladd & Thomas, 2000); elt.ccsenet.org English Language Teaching Vol. 14, No. 12; • Sending children to cram schools or hiring tutors (Dawadi, 2019(Dawadi, , 2020Ferman, 2004;Martínez et al., 2004); • Teaching children test-taking strategies and sharing the test-taking experience with children (Dawadi, 2019(Dawadi, , 2020; • Urging children to learn for the test (Dawadi, 2019(Dawadi, , 2020Ferman, 2004); • Cooperating with neighbors and relatives to support children' test preparation (Dawadi, 2019(Dawadi, , 2020); • Helping children analyze the reason why they did not succeed in the test (Martínez et al., 2004); • Familiarizing children with test questions through doing practice questions (Eizadirad, 2020). In addition, research shows that not all parents involve in children's test preparation. Barksdale-Ladd and Thomas (2000) reported that some parents in the north state were involved in children's test preparation such as purchasing test-related materials, while other parents in the north state and all parents in the southern state did not involve in children's test preparation and they only took general PI strategies such as helping children with homework.
Further, the literature indicates that although parents prefer to involve in the assessment process by supporting their children at home (Desforges et al., 1994), researchers want more from parents. For example, Wolfendale (1998) highlighted the importance of involving parents in the assessment process and used the term "joint assessment" (p. 355) in an attempt to invite parents to be more involved in the assessment process. Fredericks and Rasinski (1990) fully described the principles and practice of involving parents in the reading assessment process and illustrated two assessment tools that were successfully used in an American elementary school. Rutland and Hall (2013) proposed that parents should be not only the information consumers and informants, but also be "a collaborative member of the assessment team" (p. 8) to identify "strategies and approaches for assessment that will work best for their child" (p. 8). However, these researchers mainly focused on the assessment designed for early childhood and elementary students. Additionally, there is a lack of empirical data to prove the feasibility and effectiveness of joint assessment projects.

Parents' Extreme Involvement Behavior
In some countries that administered high-stakes accountability systems, standardized testing aroused strong opposition. Parents adopted extreme behavior to express their resistance to high-stakes standardized testing, which was described as an unwelcomed PI strategy by schools (Schroeder, Currin, & McCardle, 2020a).
Some parents united to boycott high-stakes testing, and the Opt-Out movement in the US is a typical example. Since the administration of the No Child Left Behind Act, the US has stepped into an era of accountability via high-stakes standardized testing (Schroeder et al., 2016). The excessive power of high-stakes testing aroused public resistance, anger, and despair (Schroeder et al., 2016). Such background triggered "a grassroots coalition of opposition to high-stakes tests that are used to sort students, evaluate teachers, and rank schools" (Paladino, 2020, p. 14), and parents were leaders or main participants of the Opt-Out movement.
Most studies in Table 2 mainly explored the reason why parents participated in the Opt-Out movement. The findings of these studies indicated that the main reason for parents' opposition to high-stakes tests was that the high-stakes accountability system was not conducive to students' educational development (e.g., Abraham, Wassell, Luet, & Vitalone-Racarro, 2018), which might harm children's physical and mental health, as well as students' learning (e.g., Schroeder et al., 2016Schroeder et al., , 2020a. In addition, parents argued that the high-stakes testing accountability system was "government overreach" (Schroeder et al., 2016, p. 8), which was unfair for both teachers and students (e.g., Pizmony-Levy & Saraisky 2016;Schroeder et al., 2016Schroeder et al., , 2020a.

The Answer to RQ 2
RQ2 is about the effect and the effect size of parents' involvement behavior in the assessment process for promoting students' academic achievement. No direct evidence was identified to answer this question. However, the literature in Table 3 provided implicit evidence for the effectiveness of PI in language assessment. Many intervention programs offered a series of sessions to help parents familiarize themselves with tests and teach them how to support their children's test preparation. In Araque et al.'s (2017) and Ashbaugh's (2009) research, the intervention programs successfully improved PI_TP, which in turn promoted child academic achievements. On the other hand, according to Auerbach and Collier (2012), although the program increased PI_TP, children's academic attainments were not advanced. Moreover, Dawadi (2019) also provided indirect evidence that although parents tried their best to support their children's test preparation, children did not perform well on the test. These conflicting findings suggest that consensus has not been reached about the effectiveness of PI in the language assessment process.

The Answer to RQ 3
RQ3 asks "what factors affect parents' involvement behavior in the assessment process?" Only one study provides direct evidence to answer this question. Cheng et al.'s (2010) research showed that parents' perceptions of test quality, test knowledge, and opportunities to know about tests contributed significantly to PI_TP (R 2 = .314, large effect size). Parents' perceptions of test impact on students' learning motivation also made a significant contribution to PI_TP (R 2 = .185, moderate effect size). Further, parents' levels of education and the time they spent with their children also had a positive effect on PI_TP (R 2 = .116, small effect size).
Besides, this question can also be answered through indirect evidence. For example, the studies on the Opt-Out movement (see Table 2) indicated that parents' negative viewpoints about high-stakes assessment might impact their behavior. Martínez et al. (2004) also reported that parents' knowledge about assessment and their levels of education might affect their involvement, but they did not report the effect sizes.

Discussion
The next section firstly discusses the answers to research questions. Then, it proposes issues that deserve further investigation. Finally, the test impact on parents' behavior is delineated.

Can the Existing Literature Fully Answer the Research Questions
RQ 1 was fully answered. Even though parents know little about language and assessment-related information, parents are still involved in the high-stakes language assessment process (Dawadi, 2019), and the ways of involvement include general and extreme involvement. Typically, extreme involvement strategies can only be found in a few countries implementing high-stakes testing accountability systems. It might thus be concluded that the most frequently adopted PI strategies under the context of high-stakes language assessment include general PI and PI_TP strategies.
RQ 2 was not fully answered. Most studies merely described what parents did under the context of high-stakes language assessment, while few researchers specifically examined the effectiveness of PI_TP practice. Although some studies provided indirect evidence indicating the effectiveness of PI_TP behavior, their findings were contradictory. Additionally, scarcely did researchers report the effect sizes, meaning that little is known about the magnitude and importance of the effect of PI_TP strategies.
RQ 3 was also not fully answered. Although a few potential factors were identified, direct evidence was only garnered from Cheng et al.'s (2010) research, and most of the evidence was indirect that calls for further investigation.

Suggestions for Future Research
Based on the answers to research questions, this paper proposed the following issues that deserve further investigation.
• How should parents get involved in the language assessment process? Is it necessary to invite parents to engage in the assessment process through intervention programs?
• What is the conceptualization of PI_TP? What is the definition of PI_TP? Can PI_TP be just taken as a dimension of PI?
• What are the specific strategies of PI_TP under various educational backgrounds? Will the effect or frequency of these strategies change with other factors such as time or grade?
• What factors affect PI_TP? The existing studies implied that parents' perceptions of testing might affect PI_TP, and parents' perceptions of testing might consist of several dimensions. Thus, which dimension affects PI_TP? What is the nature and extent of the effect? Are there mediating factors in the process of parents' perceptions of testing affecting PI_TP behavior? Does this process get moderated by other factors?
• What is the effectiveness of PI_TP? Does PI_TP promote students' academic achievement or other aspects of learning? If yes, what about the effect size? Among various PI_TP strategies, which is most effective?

Test Impact on Parents' Behavior
Taken together, there are three features concerning the impact of high-stakes language assessments on parents' behavior: • High-stakes language assessments might influence the ways parents are involved in children's learning. Specifically, parents tend to adopt various PI_TP strategies. For example, sending children to cram schools or hiring tutors is a strategy that parents most frequently used in the test preparation period (Dawadi, 2019(Dawadi, , 2020. In order to help children prepare for tests, parents will learn test-related knowledge through various channels, such as communicating with teachers (Holyk, 2011) or attending parent intervention programs.
• High-stakes language assessments might have a positive impact on some parents but a negative impact on others. Evidence shows that when a test provides accurate information that parents need, it will encourage more active parental involvement in children's test preparation (Vandergrift, 2015). On the other hand, tests may make parents feel extremely anxious, which in turn drives them to adopt inappropriate PI_TP strategies such as forcing their children to stay up late to prepare for tests at the cost of reducing their sleep time (Dawadi, 2019(Dawadi, , 2020. • High-stakes language assessments might influence some parents, but not others. For instance, in Desforges et al.'s (1994) research, some parents were engaged in test preparation, while others were not.

Conclusion
The present study has several limitations. Firstly, due to the limited number of empirical studies regarding the impact of high-stakes language assessments on parents' behavior, it did not consider the standard of peer review when screening literature, which may weaken the reliability of this study. Additionally, there is a shortage of longitudinal research, making it difficult to describe the dynamic process of the impact of high-stakes language assessments on parents' behavior. Furthermore, a few studies (e.g., Ferman, 2004) did not collect data directly from parents, rather, they gather data from other test stakeholders such as teachers and students, which might also weaken the reliability of this study. Finally, a lack of quantitative data makes it difficult to statistically describe and synthesize study outcomes.
Despite these limitations, this meta-analysis makes several contributions to test impact research. Firstly, it might be the first systematic synthesis that reveals test impact on parent behavior. Through analyzing the findings of extant literature, it identifies both direct and indirect evidence to expound whether and how parents are involved in the assessment process, the factors affecting PI in the language assessment process, and the effectiveness of such involvement. Moreover, it summarizes test impact on parents' behavior, while most of the previous research merely focused on test impact on parents' perceptions.
This research synthesis has practical implications. Firstly, it highlights the importance of exploring the impact of high-stakes language assessment on parents' involvement behavior and proposes several issues for future research. Secondly, it suggests that parents want to be involved in children's test preparation, but they do not know how to do it (Ashbaugh, 2009) so that it might be helpful to provide colorful intervention programs and materials that meet their needs.