A Systematic Review of Parent–Child Communication Measures: Instruments and Their Psychometric Properties

Parent–child communication represents an important variable in clinical child and family psychology due to its association with a variety of psychosocial outcomes. To give an overview of instruments designed to measure the quality of parent–child communication from the child’s (8–21 years) perspective and to assess the psychometric quality of these instruments, we performed a systematic literature search in Medline and PsycInfo (last: February 25, 2022). Peer-reviewed journal articles published in English with a child-rated instrument measuring the quality of parent–child communication were included. Initial screening for eligibility and inclusion, subsequent data extraction, and quality assessment were conducted by couples of review team members. Based on the screening of 5115 articles, 106 studies reported in 126 papers were included. We identified 12 parent–child communication instruments across the studies. The Parent-Adolescent Communication Scale (PACS) was used in 75% of the studies. On average, the evidence for psychometric quality of the instruments was low. Few instruments were used in clinical and at-risk samples. Several instruments are available to rate parent–child communication from the child’s perspective. However, their psychometric evidence is limited and the theoretical foundation is largely undocumented. This review has limitations with regard to selection criteria and language bias. Registration PROSPERO: CRD42021255264. Supplementary Information The online version contains supplementary material available at 10.1007/s10567-022-00414-3.

Reviews have shown that the child's own perspective and children's active involvement in research about their psychosocial situation is largely under-utilized (e.g., Facca et al., 2020;Larsson et al., 2018). Hence, a fourth reason to focus on child-rated measures is that this may promote the use of children as informants in research. The final reason relates to relevance for the practice field. Self-report questionnaires are more accessible and less resource-demanding to administer than observer-rated measures. Providing an overview of easy-to-administer child-rated measures will thus have high relevance for the practice field.
A fourth methodological issue to consider is the scope or focus of the parent-child communication measure. Definitions of parent-child communication are wide and varied (Vangelisti & Caughlin, 1997), which opens up several measurement angles. Measures can be focused on topics [e.g., sexuality (Sales et al., 2008), health behavior (Miller-Day & Kam, 2010), conflicts (Peterson, 1990)], and/or situations/settings [e.g., home, laboratory (Hadley et al., 2013)], and/or refer to the general quality of parent-child communication (Barnes & Olson, 1982). In addition, measures can address dyadic communication between the child and one parent or triadic communication between the child and both parents. Furthermore, measures can focus on communication quality, frequency, or a combination of these (Miller-Day & Kam, 2010;Xin et al., 2021). A related issue is the time perspective of the measure, i.e., concurrent, prospective, or retrospective. In the current review, our interest lies in measures of parent-child communication that are widely applicable, especially with regard to child mental health and development. Therefore, we focus on the quality of current general parent-child communication and, if subscales are provided, their specific features.
Finally, measures can be tailored for different populations. Whereas some measures are meant for the general population, others are tailored for ethnic groups, nationalities, or age groups. In the current review, we focus on measures for the general child population that can be applied to clinical and at-risk populations as well. "Clinical" indicates that the child has been diagnosed with a mental health or somatic disorder, whereas "at-risk" indicates that the sample was selected according to criteria that are considered as a transitory or continuous risk for child mental health such as being a minority or being bereaved.
The aim of the current study is to provide the field with an overview of existing instruments that measure the quality of parent-child communication from the child's (8-21 years) perspective. We will consider the psychometric properties of the scales using criteria based on De los Reyes and Langer (2018). We investigated the following research questions: Which child-report questionnaires exist to measure parent-child communication, what kind of samples have they been applied to, and what is their psychometric quality? We will also consider the instruments' availability, including translations and norms, to ease the decision-making processes for practitioners and researchers who aim to measure parent-child communication.

Methods
We conducted this systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). We searched the PROSPERO database initially to ensure that no similar studies had been started or planned and published a protocol for this study under the number: CRD42021255264.

Eligibility Criteria
The eligibility criteria were as follows: original, peerreviewed journal articles published in English language and assessing the quality of general communication between parents and their children via multi-item scales for child selfreport were included. In the context of this review, communication included verbal, nonverbal, cognitive, and affective aspects of the interaction between parents and their children, but not the physical ability to communicate. Studies assessing broader concepts such as general family communication or studies using single questions or ad-hoc measures to assess parent-child communication were excluded, just as studies on specific topics of communication such as healthrelated behaviors (e.g., sex, alcohol, tobacco use). Studies reporting only parent ratings were excluded.
The age range of the study population was set at 8 to 21 years of age, including older children, adolescents, and emerging adults. We included studies examining parent-child communication in general, clinical (both somatic and mental health), and at-risk populations. In terms of study design, we included all types of empirical studies (crosssectional, longitudinal, interventional, validation studies). Qualitative studies and case reports were excluded.

Information Sources and Search Strategy
We conducted the main search and selection process between May 2021 and October 2021, identifying original studies by searching the electronic databases APA PsycInfo (Ovid) and MEDLINE (Ovid). On February 25, 2022, an updated search for papers published after the initial search was conducted and resulted in the addition of nine reports. The references of all selected publications were searched for additional studies. We included additional sources on psychometric data in our assessment of psychometric quality if it was referred to in one of the publications and available in English. Table 1 presents the search strategy used via the

Quality Assessment
In our evaluation of instrument quality, we relied on the criteria set forward by Hunsley and Mash (2007, 2008, complemented by Youngstrom et al. (2017), and summarized in De los Reyes and Langer (2018). This system is used to rate the psychometric properties of assessment instruments across nine categories: (a) norms, (b) internal consistency, (c) interrater reliability, (d) test-retest reliability, (e) content validity, (f) construct validity, (g) validity generalization, (h) treatment sensitivity, and (i) clinical utility. Each category includes a description of the quality of evidence required for a rating of adequate (minimal level of scientific rigor), good (solid scientific support), or excellent (extensive, high-quality support). Youngstrom et al. (2017) later added repeatability, discriminative validity, and prescriptive validity to the original system. Since the original system was intended for clinical measures, not all categories apply to parent-child communication scales. Thus, in the current review, we rated the following nine quality categories: norms, internal consistency, test-retest reliability, content validity, construct validity, factorial structure, discriminative validity, validity generalization, and treatment sensitivity (See Supplement Table).
In terms of norms, there are no clear cut-offs for sample size, but we considered community samples N > 400 and clinical samples N > 100 as representative. In terms of reliability, the system applies the following criteria for Chronbach's α: 0.70-0.79 is adequate, 0.80 to 0.89 is good, and > 0.90 is excellent, based on the median of reported numbers. The quality assessment was conducted by KF, SO, SC, YH, and JB, as well as cross-checked by SC and HZ. If members of the review team had co-authored a paper under consideration, the other team members did the quality assessment.

Study Selection
A total of 6147 hits were retrieved from the databases, and 1032 duplicates were removed. Based on the screening of the remaining 5115 titles and abstracts, 499 full-texts were retrieved. Based on these we included 118 papers. In addition, reference lists were searched for eligible literature, resulting in 32 papers of which 28 were retrieved. Eight additional papers were included, resulting in 126 papers in total. Figure 1 shows the corresponding PRISMA flow chart. Table 2 provides an overview of included articles.

Study Characteristics and Identified Instruments
Twelve different instruments were identified (see Table 3 for an overview and Table 4 for quality assessment). The Parent-Adolescent Communication Scale (PACS; Barnes & Olson, 1982) was used in 100 papers based on 85 studies. The Parent-Child Communication Scale (PCCS; McCarty, McMahon and Conduct Problems Prevention Research Group, 2003) was used in seven papers based on four studies. The Parent-Adolescent Communication Inventory (PACI; Bienvenu, 1969) and the Revised Family Communication Patterns Instrument (RFCP; Ritchie & Fitzpatrick, 1990) were used in three studies each. The Parent-Child Communication Questionnaire (PCCQ; Yang & Zou, 2008), the Parent-Child (Parent-child* or parent-adolescent* or parentteen* or mother-child* or father-child* or mother-adolescent* or father-adolescent* or caregiver-child* or caregiver-adolescent*). ti,ab,hw,kf,mh 2 Communication.ti,ab,hw,kf,mh 3 (Functioning or well-being or mental health or stress or psychopathol* or adjust* or relationship or internali* or externali* or valid* or psychometr* propert*).ti,ab,hw,kf,mh 4 1 and 2 and 3 5 4 and 1991:2022.(sa_year) 6 Remove duplicates from 5 Communication Scale (PCCS; Loeber et al., 1998; also known under the name Revised Parent-Adolescent Communication Form (RPACF); Loeber et al., 1991), and the Parent-Child Communication Scale (PCCS; Krohn et al., 1992) were all used in two studies each. The Father-Adolescent/ Mother-Adolescent Communication Scale (FACS/MACS; Shek et al., 2006), the Parent-Child Communication Scale (PCCS; Chi, 2011), the Perception of Parenting Communication Scale (COMPA; Portugal & Alberto, 2014), the Parent-Adolescent Communication Inventory (PACI; Schmidt et al., 2010), and the Family Communication Patterns Scale (FCP;McLeod et al., 1972) were all used in one study each.

The Parent-Adolescent Communication Scale (PACS)
The PACS (Barnes & Olson, 1982) was the most widely used instrument and has been translated from English into nine other languages (Spanish, Dutch, Chinese, French, Malay, Italian, Khmer, Korean, and Swedish). The instrument comprises two subscales (open communication and communication problems, 10 items each). Items are measured on a 5-point Likert scale, and parent and children versions are identical apart from changing referents (my mother/ father/daughter/son). For the quality assessment, Barnes and Olson's (1982) study was considered as source in addition to the studies retrieved by the systematic literature search.
In terms of content validity, the original authors specified the conceptual foundation within the framework of the circumplex model of family functioning: parent-child communication was conceptualized as an additional dimension facilitating adaptive change in family functioning (Barnes & Olson, 1985). However, in most studies found in this review, the PACS was used as a stand-alone measure of dyadic parent-child communication. The factorial structure found in the initial study (Barnes & Olson, 1982) was corroborated by a principal component analysis in a Dutch study (Jackson et al., 1998). In a Spanish sample, however, three factors were found in another principle component analysis (open communication, communication problems, and avoidant communication;Estevez et al., 2005). In an American sample, only one factor was found in exploratory and confirmatory factor analyses, with two items from the problem scale not loading on the same factor as the other items (Wu & Chao, 2011). The PACS has shown good internal consistency in most of its' versions. The means and standard deviations are available for large samples (i.e., n > 400) in six versions (see Table 2). For the English version of the PACS, some evidence on test-retest reliability and treatment sensitivity was found. The included studies did not report sufficiently on construct and discriminative validity. The PACS has been published originally in Barnes and Olson (1982) and is available online (for example: https:// scales. arabp sycho logy. com/s/ parent-adole scent-commu nicat ion-scale-pacs/). Reports assessed for eligibility (n = 28) Reports excluded: n = 20 Wrong outcome (n = 11) No standardized instrument (n = 2) Topic specific instrument (n = 2) Parent report only (n = 2) Wrong population (n = 3) Studies included in review (n = 106) Reports of included studies (n = 126)

The Parent-Child Communication Scale (PCCS McCarty)
The PCCS (McCarty, McMahon and Conduct Problems Prevention Research Group, 2003) was the second most widely used instrument. It has been translated from English into Norwegian and Khmer. Parent and child versions are different; the child version consists of 10 items (parents 20 items) in the original version and is based on a 5-point Likert scale. For quality assessment, Pek (2006) was considered as additional source. With regard to content validity, the instrument is intended to assess the parent's openness to communication (Pek, 2006), but the conceptual foundation is not specified further. In later analyses, only eight items fit the subscales parent and child communication (five and three items, respectively). This factorial structure was corroborated by confirmatory factor analyses in a Norwegian study (Orm et al., 2022). Internal consistency is between low and adequate. For the English version, norm values of a large sample exist (Pek, 2006), means and standard deviations for an at-risk sample can be found in Orm et al. (2021Orm et al. ( , 2022. There was no sufficient evidence with regard to test-retest reliability, construct, and discriminative validity, but for treatment sensitivity. The PCCS is available from the website of the Fast Track Project (https:// fastt rackp roject. org/ data-instr uments). The questionnaire is not included in the original reference publication (McCarty et al., 2003).

The Parent-Adolescent Communication Inventory (PACI Bienvenu)
The PACI (Bienvenu, 1969) is explicitly recommended for children older than 13 years. According to Carson et al. (1999) and Green and Vosler (1992), it consists of 40 items, whereas the original author claims it to be a 36-item scale.
No subscales are hypothesized nor has a factorial structure been reported. The original article (Bienvenu, 1969) was considered as an additional source for quality assessment. Psychometric data on test-retest reliability have been presented (Bienvenu, 1969). Green and Vosler (1992) reported treatment sensitivity with regard to the parent, but not the child scale. Internal consistency is less than adequate. In the present sample of studies, there was no information on content, discriminative, and construct validity for the PACI. A 21-item version of the scale is available in the original publication (Bienvenu, 1969).

Further Instruments
With regard to the remaining instruments, the PCCS (Chi, 2011) and the PCCQ consist of two and four subscales, respectively. The COMPA is recommended for ages seven to 18 and has five subscales: Parental availability to communication, children confidence/sharing, emotional support/ affective expression, meta-communication, and negative communication patterns (Portugal & Alberto, 2014). The factor structures have not been established for the PCCS (Chi), the PCCQ, or the COMPA. The means and standard deviations in larger samples (i.e., n > 400) were only reported for the FACS/MACS (Shek et al., 2006) and for the PCCS (Chi, 2011). The internal consistencies ranged from less than adequate (COMPA; Portugal & Alberto, 2014) over adequate (FCP;McLeod et al., 1972;PCCQ;& Zou, 2008;PCCS;Krohn et al., 1992;PCCS;Loeber et al., 1998) and good (PACI; Schmidt et al., 2010;PCCS;Chi, 2011) to excellent (FACS/MACS; Shek et al., 2006). Test-retest reliability, content validity, construct validity, factorial structure, discriminative validity, and treatment sensitivity were not reported in any of the included studies. To our knowledge, complete original items are only available for COMPA and PCCS (Krohn) in Portugal and Alberto (2014) and Krohn et al. (1992), respectively.

Synthesis: Psychometric Evidence of Parent-Child Communication Measures
We extracted the available psychometric data from all publications. We included additional sources on psychometric data in our assessment of the psychometric quality for four measures (i.e., PACS, PCCS (McCarty), PACI, (Bienvenu), RFCP; i.e., Barnes & Olson, 1982, 1985Bienvenu, 1969;Ritchie & Fitzpatrick, 1990;Pek, 2006). Most instruments showed at least adequate internal consistency. Means and standard deviations that can be considered as normative information (community samples N > 400 and clinical samples N > 100) were reported for almost half of the existing measures (Table 4). The conceptual and theoretical background of the instruments were often not specified clearly. Additional psychometric information is scarce, and construct and discriminative validity were not stated in an explicit way for any of the instruments. However, means (

Discussion
The aim of this review was to identify instruments that measure parent-child communication from the child's perspective as well as which samples they have been applied to and to assess the psychometric quality of these instruments. We identified twelve instruments across 106 studies. The PACS had been used in most studies (k = 85) by far. This means there are relatively few multiple-item and general (i.e., not topic-specific) child-rated parent-child communication instruments, given that we included > 100 studies in this review. On the one hand, this indicates some unification in the field, with a strong PACS dominance. Given that the other instruments were used in four studies each or less, with five instruments having only been used in single studies, a potential conclusion could be that the PACS should be considered the "gold standard" for child-reported parent-child communication quality. However, frequency does not ensure quality, which will be elaborated on below. Before addressing samples and psychometric quality, however, it is important to consider that the existing instruments measure quite different aspects of parent-child communication, ranging from communication problems, openness or conformity and conversation orientation over problem-solving oriented, or avoidant communication to meta-communication. This variety reflects the complexity of communication and points to the fact that doing research on communication also should entail conceptual consideration. In addition to using measures with adequate psychometric properties, researchers and/or clinicians should be aware that they can aim at measuring quite different aspects of parent-child communication.
In terms of theoretical foundation, the information on most instruments was scarce. Considering that multiple studies were excluded due to using single-item measures (e.g., how would you rate the communication between you and your father/mother) or topic-specific measures (e.g., about alcohol, drugs, sex), there seems to be relatively little theory-based consensus in the field concerning how to conceptualize "parent-child communication" in a child-rated instrument. The two most dominating dimensions concern how open and/or problematic the communication is considered. However, the range of existing subscales indicates that there may be additional relevant dimensions of parent-child communication. In addition, the theoretical background of openness and communication problems also needs further clarification, since, for example, openness can be considered as a matter of openness for different perspectives or 1 3 as openness for expressing emotions. Hence, a synthesis of previous models of parent-child communication, as has been attempted for the variable family communication (Murphy et al., 2017), would be desirable. Up to now, the field lacks standards for measuring other dimensions of parent-child communication.
The samples used ranged from general community samples via at-risk samples to clinical samples, with a dominance for community samples. More than half the studies were conducted in the North Americas (USA and Canada), with an even number of European and Asian studies in "second place," followed by only a handful of studies from Middle and South America or Australia, and only one study from Africa. Although the USA samples included a mix of White, Black, Asian, and Hispanic Americans, there is clearly a need to examine and validate parent-child communication measures cross-culturally. Only the PACS has been used in clinical, at-risk, and community samples. The COMPA, PACI, RFCP, PCCS (Loeber), and PCCS (McCarty) instruments have been used in at-risk and community samples. The PACS and PCCS (McCarty) instruments can detect the differences in the quality of parent-child communication between these samples.
In terms of psychometric quality, the evidence backing most of the scales was insufficient. Our evaluation of this was based both on original publications of the instruments, as well as the synthesized data across studies that have used the instruments. Generally, little is known about most instruments, and even the English version of the PACS reached only 11 of 27 possible points in our quality assessment based on information from 51 publications. Adequate evidence for psychometric core aspects such as test-retest reliability or factorial structure was found for less than a handful of instruments. Across the studies, construct and discriminative validity have hardly been evaluated in the included studies. In addition, little information was available regarding convergent validity with regard to other measures such as observer-based (Hadley et al., 2013) or parent-based reports (Hartos & Power, 2000a, 2000b. However, it is important to note that while our main overall goal was to assess the psychometric properties of parent-child communication measures, this was not the goal of the reviewed studies. Rather, most of the included studies considered various research questions, of which parent-child communication was one of the several. At best, the results of some of the individual studies can be interpreted as preliminary/emerging evidence for construct and discriminative validity for one measure in one sample. However, when considered combined in this review, overall systematic evidence for the psychometric properties of even widely used scales (e.g., the PACS) must be said to be severely lacking.
We also aimed to consider availability of the instruments. Various abbreviations for the different instruments were often used inconsistently in the literature. Further confusion is added since some of the instruments are based on each other: The PACS has been used as one of several sources for the PCCS (Loeber), which, in turn, was used for the PCCS (McCarty) and the PCCS (Krohn). It is also noteworthy that the original PCCS (McCarty) instrument (child version) contains six items from the PACS almost literally, only the grammar has been changed from first to second person. The parent communication subscale of the PCCS (McCarty) consists of five items, with four from the PACS (three from the open, one from the problem subscale). The child communication subscale comprises three items with only one from the PACS. However, the two remaining items are fairly close to other PACS items. This leads to confusion about the measures and underlines the importance of indicating correct sources, number of items, scale range, and example items. In terms of availability, the instruments had been translated and used in a total of twelve languages.

Limitations
To our knowledge, this review is the first to give a comprehensive and systematic summary of parent-child communication measures and their psychometric properties. However, there are some limitations. The review did not include observational or parent report measures. The selection criteria limited the search down to published peer-review articles but did not consider other publications, unless they were cited as a source for psychometric information in the included papers. In that case, they were considered in the quality assessment. If publications such as dissertations would have been included, other information regarding conceptual considerations and psychometric evidence might have been found. In terms of age, our search was limited to ages 8-21 years. Hence, the psychometric evidence for single instruments may differ with regard to emerging adults. Even though we searched for instruments in any language, we only considered publications in English, leading to a certain bias between English and non-English instruments. For that reason, the quality assessment does not allow for an exact comparison between instruments of different languages. Last but not least, since genuine psychometric studies were hardly found, we decided to not use the EMPRO tool (Valderas et al., 2008) to conduct the quality assessment as initially planned. Instead, we decided to gather psychometric information with an adapted tool that was more apt to also consider information found in a huge number of non-psychometric study reports.

Implications and Conclusions
This review has implications for practice and research. For researchers and practitioners interested in examining parent-child communication from the child's perspective, there are several child-rated instruments to choose from in English, Chinese, and some European languages. We have provided availability information for the four mostused child-rated parent-child communication scales in the methods section. The most widely used scale is the PACS (Barnes & Olson, 1982), which considers the degree of openness and problems in communication, and with some evidence of factor structure and other psychometric properties. However, the PACS should not necessarily be the default choice due to frequency alone. Upon choosing which instrument to apply, practitioners and researchers should take active and informed choices about which aspect of parent-child communication they aim to assess, as the instruments focus on different dimensions of parent-child communication. For example, whereas the PACS concerns openness and communication problems, the RFCP measures conformity and conversation orientation.
In terms of research implications, there is a need for more studies assessing the psychometric properties of parent-child communication scales, and when choosing an instrument, also conceptual and cross-cultural aspects should be carefully considered. To avoid confusion, instruments should be reported with correct sources, number of items, scale range, and example items. Future research should also consider triangulating child and parent reports as well as observational measures.
Funding Open Access funding enabled and organized by Projekt DEAL.

Data Availability
No new data were created during the study.

Conflict of interest
The authors have no financial or proprietary interests in any material discussed in this article.
Ethical approval This work does not utilize human subject data and no ethical approval is required.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.