Development and validation of a new parental authority instrument (PAI).
The central theoretical framework in the field of parental authority deals with describing parenting styles or types (Baumrind, 1991; Maccoby & Martin, 1983), distinguished by the extent to which the parent sets boundaries and provides guidance, explains and justifies demands and expectations, employs control and power, and also provides emotional support (Yaffe, 2014). Over the years, many measurement instruments have been developed for the self-reporting of children and parents, intended to evaluate the parent's use of specific practices, to characterize the level of parental control and acceptance, and to classify the parent into one of four parenting styles: authoritative, authoritarian, permissive, and uninvolved (e.g.: PPS: Bloom, 1985; PAQ: Buri, 1999; PCRQ: Furman & Giberson, 1995; PAC: Reitzle, Winkler Metzke & Steinhausen, 2001; PSDQ: Robinson, Mandelco, Olsen & Hart, 2001; PARQ: Rohner, 2005; CRPBI: Schafer, 1965; APQ: Shelton, Frick & Wooton, 1996). However, in the absence of a definition focused on the theoretical construct of parental authority, with its two inherent dimensions (power and legitimacy), the operationalization of the parenting styles sometimes serves to evaluate the parent's authority beyond its aim as a measure of general parental patterns (Yaffe, 2014).
One of the most common instruments in the area is Buri's Parental Authority Questionnaire (PAQ: Buri, 1991), which serves to classify parents into one of the three parenting styles described above. This typology, which appears in the context of authoritative (high demanding and high responsiveness) versus permissive (low demanding and high responsiveness), refers to the parent's control in terms of power processes (Henry, Wilson & Peterson, 1989; Olson & Cromwell, 1975). It lacks, however, the component of perceived legitimacy of the expression of parental power toward the child. The body of research dealing with conceptions of parental authority by parents and children has shown that it is essential in the definition of the parent's authority, associated consistently with obedience in practice, the tendency to reveal information to the parent, involvement in undesired activity, and a long list of outcomes among adolescents (Darling, Cumsille, Caldwell & Dowdy, 2006; Darling, Cumsille & Martinez, 2007, 2008; Darling, Cumsille & Pena-Alampay, 2005; Yau & Smetana, 2009). Furthermore, a series of studies provided by this body of research has shown that parental authority is context-dependent (related to the domain under discussion) and is not uniform among issues (Smetana, 1988, 1993; Smetana & Asquith, 1994; Tisak, 1986). It is therefore reasonable to assume that operationalization of this concept should be based on an appropriate array of stimuli, accurately representing the parent-child conflict (Yaffe, 2014). Rather, the various instruments mentioned above lack a specific behavioral and social context, and their measurement focuses on assessing the parent's general control tendencies. This is because they are aimed at measuring generalized parenting practices, behaviors, and styles in socializing children at home, existing beyond the specific context (Smetana, 1994). In the absence of a specific context, the dimension of authority they contain (parental control) is limited in constituting a sufficiently specific index of the theoretical structure of parental authority.
The current paper describes the development and validation of a new instrument to assess parental authority, aimed at measuring this specific theoretical structure and offering a complement to and substitute for the arsenal of measures in this area. The topic of parental authority has recently become popular among professionals and researchers, so it is essential to strive for more specific measurement of this construct.
Constructing the instrument was based on an operationalization of Yaffe's theoretical conceptualization (Yaffe, 2013, 2014), defining parental authority as a bi-dimensional construct of parental power and the perception of its legitimacy. Based on the two areas of power in the model of Olson and Cromwell (1975), the power dimension in the questionnaire was defined using two operational variables: parental demanding (power processes), and the ability to influence the child (power outcomes). In accordance with the original approach of the research body on conceptions of parental authority (e.g.: Smetana, 1995, 2000; Smetana & Asquith, 1994), the dimension of legitimacy is defined in the questionnaire through two operational variables: parent's right to demand and child's obligation to obey. In order to give expression to the dynamic nature of the construct (as a function of context) and the conflict at its basis (Yaffe, 2014), the instrument measures these four operational variables on the background of a range of issues from family life in parent-child relations. On the empirical level, integrating the two inherent dimensions of the original concept of authority, in order to weight them into an index of the level of parental authority in the family, constitutes a distinctive approach of parental assessment in this area of research so far.
The current paper describes the development and validation of a Hebrew version of an instrument designated to measure parental authority in the family through children's reports. In general, the process of constructing the instrument is divided into two main components: developing the questionnaire and phrasing the items, along with its preliminary checking through a pilot study, and an empirical-statistical investigation of the validity of the final instruments. These two components were implemented through the five sub-studies described below. The validated instrument was translated into English for further examination and international research purposes.
STUDY 1: DEVELOPING THE ITEMS AND CONSTRUCTING THE PRELIMINARY INSTRUMENT
The process of developing the preliminary instrument's structure and phrasing the items was based on a qualitative study in a focus group of parents and children. The strategy for developing the instrument was founded on a systematic analysis of the contents of in-depth interviews, in order to identify central themes of conflict in parent-adolescent relations. The specific issues were classified in a preliminary conceptual-logical manner into core content domains, and these were used to phrase short descriptions of conflict situations expressing a parental demand directed at the child against the background of the discussed issue.
Participants Semi-structured depth interviews were conducted with 15 parents and children from the focus group. The fifteen interviewees participating in the focus group were distributed as follows: 3 boys aged between 11.5 and 15; 5 girls aged between 12 and 15; 5 mothers to boys and girls aged between 13 and 17; 2 fathers of boys aged 13 and 16.5.
Procedure The interviews lasted about an hour on average (between 30 and 75 minutes), were usually conducted in a neutral location or in the interviewee's home, and were transcribed and content-analyzed to produce specific themes of parent-child conflict situations. The interviews were conducted by two different interviewers: the researcher himself and an additional female research assistant trained for this purpose.
At the start of each interview, the interviewee was presented with the purpose of the interview and the concept of parental authority. For child interviewees, their parents' consent for their participation in the interview was obtained in advance. Each interviewee expressed his/her consent to recording. The interview began with general questions about conflict situations in the family (For example, child: "where you and your parents disagree with each other?"; parent: "after I have tried to explain the meaning of the concept of parental authority, can you please tell me about issues of situations in your family where this is relevant?"). First the interviewees were asked to freely recall relevant cases from their home, and when these efforts were exhausted (in light of lack of response from the interviewee), efforts were made to direct the discussion to additional central content areas that had not been discussed so far.
Content analyses of the interviews: In order to identify the relevant themes appearing in the interviews and to classify them by their apparent conceptual content area, thematic content analyses were conducted for all fifteen interviews separately. At this stage the emphasis was on locating as many subject categories as possible in each interview. The subject categories that share the same conceptual content were classified under a primary category defining their content area. The themes (sub-categories) were classified into six main subject categories, defined as content areas: 1. Routine; 2. Studies; 3. Behavior toward others at home; 4. Child's safety and welfare; 5. Social; 6. Personal. The conceptualization of the areas was based on the theoretical framework of the research body of conceptions of parental authority (e.g., Smetana, 2000), but the subject classification of the sub-categories was conducted on the basis of considerations of preliminary conceptual appearance.
Developing and verifying the items: Aiming to create a wide and profound preliminary infrastructure of items, to be tested later using statistical means, the first 41 questionnaire items were phrased based on the thematic potential appearing in the content categories. The content format of the phrasing of the items reflects two immanent theoretical principles of the concept of parental authority: demanding and disagreement. The conversion of the themes into situational items was conducted by phrasing the item in terms of a parental demand or prohibition that implies disagreement by the child (e.g., "your parents demand that you go to sleep in the evening at a time when you are watching television"). The various items serve as a background conflict situational stimulus, while the questions referring to each stimulus are derived from the operational definitions of the four aspects belonging to the construct's dimensions of power and legitimacy (e.g., "to what degree can your parents make you obey in such a case?"). The response scales in the questionnaire, intended for each specific question, were defined in parallel with phrasing the preliminary items, based on the relevant methodological considerations. The response scale for an item was defined on a 4-level Likert scale (0-3), from the lowest level in the parental authority aspect expressed on the scale (very little) up to the highest level (very much).
When the preliminary array of items was formed, an intensive clarification was conducted with two children aged 12-13 (the lower age boundary of the instrument's target population) regarding the semantic and linguistic quality of the items. Items that communicated unclearly, in terms of their linguistic phrasing or their embodiment of conflict, were rephrased or replaced. At the end of this stage, the questionnaire contained forty-one items, on the basis of four scales (i.e., questions).
STUDY 2: PILOT STUDY
The general study format planned for testing and validating the instrument was partially tried out in a preliminary pilot study, conducted among a small group of children, in order to test the efficiency and correctness of the procedure and the suitablity of the parental authority instrument and its validity indexes. We also sought to test the relevance of the questionnaire items and its questions (face validity), among its target group of respondents.
Participants and study procedure. As part of the testing of the pilot study, 40 boys and girls aged 12-14 (M = 13.19), from two eighth grade classes in a school in northern Israel, were asked to complete the following questionnaires: the parental authority instrument described above, with 41 items; a nine-item adolescent conformity questionnaire (see the measures section in Study no. 3); a ten-item social desirability questionnaire (see the measures section in Study no. 3).
The participants were requested to read and follow carefully the instructions for filling in the questionnaires, consulting the researcher and the teacher, who were present in the classroom throughout. Along with filling in the parental authority instrument, the participants were also asked to mark an X next to items they believed were completely irrelevant regarding their relations with their parents at home. This operation was used to sharpen the questionnaire's face validity by locating and screening the items that most responding children did not consider relevant to the issue.
Data from the 40 pilot participants were examined, and, when found to be filled out correctly, were entered and analyzed statistically. The results of the means and standard deviations of the variables, including correlations between the instrument's authority scales and the validity indexes, are presented in Table 1. Apart from the scale "Power 1" (demanding), the instrument's scales present the expected correlation trends, both on the internal level and in relation to the validation indexes. Since the correlations for the "Power 1" scale were not obtained as expected, its question was rephrased to reflect hypothetical parental demanding, even if the issue was not necessarily perceived as actual. The new operational definition for the scale "Power 1" was examined and confirmed with two boys aged 13-14 individually.
In addition, two items rated by the respondents as completely irrelevant, at a rate of over 50% agreement, were removed from the questionnaire (one was also highly correlated with the social desirability scale). Another item was removed due to having a very low range of responses with negligible variance. Several items were rewritten and polished carefully as necessary. At the end of the series of tests and changes, the intial draft of the parental authority instrument included 39 items derived from four different scales: demanding; potential influence; the parent's right to demand; the child's obligation to obey.
STUDY NO. 3: CONSTRUCT VALIDITY
The specific goals of this study are: examining the differences between the conceptual domains of parental authority after the instrument's items were reduced and its final structure was formulated; examining the instrument's sensitivity/vulnerability to social desirability; checking the instrument's construct validity through its ability to distinguish between different age groups of children in terms of the degree of authority attributed to their parents.
Participants The general research group constitutes the study's main axis for validating the instrument. It contained a total of 412 participants (213 boys; 199 girls), adolescents, of the age range between 12 and 17 (M=14.39; SD=1.26). For the study's purposes, this sample of adolescents was divided into five age levels, from the 7th grade to the 12th grade (Table 2). Most of the sample attended a large network of schools in the north of Israel.
Measures The Parental Authority Instrument (PAI). The instrument in the version used to examine the validity contained 28 items times 4 measurement scales (after removing 11 items as part of the factor analysis as detailed below; see Table 4). Each scale's score constitutes the mean of the responses on the item, while the general authority scale is obtained from a simple sum of the scores on these scales (on the item or scale level). The range of responses on a scale is between 0 and 3, while the range for the overall authority scale is between 0 and 12. The four scales in the instrument recorded high coefficients of reliability as internal consistency, between .90 and .92. On the basis of these 28 items, I calculated the means and standard deviations of the scores obtained for the instrument's four scales according to the five age levels, for boys and girls separately (Table 2). An overview of the distribution of the mean scores for the four scales presents means between 1.5 and 1.96 in the boys' group, and a mean range between 1.63 and 1.96 in the girls' group, with a clear and consistent trend of a decline in the authority scales among both sexes as their age increased.
TABLE 2 Description of the Means & SDs of the Instrument's Scales by Child's Age and Sex
(Table 2 is available upon request from the author)
A general examination of the correlations between the four authority scales composing the instrument (Table 3) indicates an internal correlation between them among boys (n=211; r= .62 to .79, p< .001) and girls (n=199; r= .58 to .79, p< .001), which accords with the theoretical link expected between these aspects of parental authority. In both groups medium to high correlations were obtained between the pairs of scales operationally defining each of the two theoretical dimensions of parental authority.
TABLE 3 Correlation Matrix between the Instrument's Scales & the Age Variable, Reporting Separately for Boys and Girls
(Table 3 is available upon request from the author)
Social Desirability Questionnaire. Social desirability questionnaires are used in studies that involve self-reporting that is suspected of having this sort of bias. From this point of view, a negligible connection between parental authority and social desirability could strengthen the instrument's construct validity. The current study used the ten-item version (M-C 1: Strahan & Gerbasi, 1972), selected as the preferred shortened format of the Social Desirability Questionnaire by Crowne and Marlowe (1960) (see: Fischer & Fick, 1993). The tool measures the individual's need for social affirmation as a personality trait through a list of items the respondent is requested to affirm or deny. The mean level of social desirability observed in the sample was 4.46 and the standard deviation was 1.90, without any significant effect of the respondent's sex on social desirability.
Procedure First, we made contact with the school central management, and were given permission to conduct the study. Prior to entering the schools, requests to the children's parents were handed out in all the target classes so that the parents could permit their children to participate in the study.
At the time of administering the questionnaires in the classrooms, the researcher, or a representative on his behalf, and a supervising teacher from the school were present throughout. The time given for filling in the questionnaires was between 45 and 60 minutes at most, as required. In a procedure that was successfully implemented in the pilot study, the children were given a series of four anonymous questionnaires (in the following order: personal details questionnaire, original Parental Authority Instrument, Conformity to Parent questionnaire, Social Desirability questionnaire). The Parental Authority Instrument and the conformity questionnaire were numbered, so that it would be possible to connect the child's questionnaire with that of his/her parents in the follow-up study, according to the class name list.
A series of confirmatory factor analyses was conducted in order to affirm the assumption of the differential factor structure of the parental authority domains, while reducing the large quota of items to achieve an optimal factor structure. In order to enable the existence of a stable factor analysis that would constitute a basis for a coherent examination of the instrument's validity later on, the scores of the instrument's four scales were averaged into a single uniform score. The justification for unifying the scores of the four scales in relation to each item relies on their statistical and conceptual matching (see Table 3). An identical approach regarding the unifying of the scales for purposes of identifying the general domain contents was also adopted in a study by Smetana and Daddis (2002). The factor analyses were conducted using Varimax type rotations on the 39 general authority scale items, according to a division into 5 predefined factors.
Three main considerations guided the decision-making process of selecting the items: 1) First, we located and removed items whose maximum loading level on the factors was below 0.4 and items that had similar loadings on three or more factors, where the conceptual distinction between them was important in principle. 2) Regarding each pair (or more) of items that were identical in terms of content, one of the items belonging to that factor was omitted, the one whose loading level on the lower factor was the lowest. Whenever the loading level on the common factor was identical among the pair of items, conceptual considerations were employed to select them. 3) Among all the remaining items, a few items were scanned and selected for omission because their commonality level was the lowest, in an attempt to obtain a higher percentage of common variance explained by the five-factor model. Following these rounds of factor analysis, 11 items were omitted from the original instrument, on the basis of this series of statistical considerations, as accepted in data reduction operation of factor analysis (see Floyd & Widaman, 1995).
The confirmatory factor analysis to confirm the five domains of parental authority refers to the 28 items of the general authority scale that remained after the item reduction. Table 4 presents the list of these items' topics according to their factor composition, along with data
regarding the loadings of the factors, the eigenvalue, and the percentage of variance explained by each factor. Separate analyses by sex provided consistent solutions, with a nearly identical factor composition, and therefore the groups were unified.
The factor analysis showed that the five-factor solution explains about 55.5% of the general variance in the PAI scores. This factor composition proivdes a sufficient solution in terms of the rate of general variance it explains, being significantly higher than the 50% threshold proposed by Streiner (1994). All items are attributed to a single factor by maximal loading (over 0.4), apart from item 15, whose loading on two factors enables its classification on a conceptual basis. The composition of the proposed distribution of items creates sufficient internal consistency data (.70-.86), considering the relatively small number of items included in most factors.
TABLE 4 Confirmatory Factor Analysis of Items' Domains: Loadings, Factor Eigenvalue, Percentage of Variance Explained by Factors, & Internal Consistency (Alphas)
(Table 4 is available upon request from the author)
The five factors obtained are characterized as follows: The first factor contains issues of a personal nature, explains 15.98% of the variance, and contains 10 items on the issues: viewing contents, providing information to parent, appearance (hair style and clothing), food consumption, social relations, Internet contents, privacy in room, managing pocket money and leisure activities. Factor 2 explains 11.66% of the overall variance and contains 4 items from the household routine domain: doing homework, activity at home after school, academic functioning, sleep times. Factor 3 contains family behavioral conventions, and contains 6 items (raising voice at parent, joining family visit, violence between siblings, babysitting sibling, room tidiness and cleanliness, help with housework), explaining 10.35% of the overall variance. The fourth factor is characterized by a strong link to the child's personal safety and welfare, and includes 4 items: smoking, piercing, alcohol use and truancy from school (explains 9.5% of the overall variance). The fifth factor involves the social area, and explains 7.9% of the instrument scores' overall variance, based on four subjects involving the child's leaving home (time of return from going out, telling parents where I'm going, going to clubs, social problems).
Differences between domains In the stage that followed the classification of the items into five domains using factor analysis, a multivariate analysis of variance (sex x domain: 2 x 5) was conducted, with repeated measures ANOVA, for the differences between the means of the items in each factor (domain) by sex (In the absence of the respondent's birth order effect on parental authority scores, the analyses were conducted on all the respondents. Other demographic differences were not taken into account due to their negligible quantities in the sample). The aim was to confirm the assumption of differentiation between the domains, as a support to the instrument's validity regarding its structural loyalty to the rationale of the theoretical construct (Messick, 1995).
Comparing the levels of authority between the domains presented in Table 5 shows significant differences between issues for two sexes in these five measurements (F (4,1636) = 426.03, p <.001; [[eta].sup.2] = .51), while the subject's sex had no main effect on the authority score. There is a within subject effect of a significant size for the domain, indicating general differences in the level of parental authority attributed to the issues of different context.
This supports the hypothesis of differences between the domains of parental authority and strengthens the instrument's construct validity. A careful examination of the specific differences between the five domains reveals significant differences between almost all their possible combinations, and in particular between the extreme domains. This is true also regarding each sex separately, beyond the overall group findings, when both among the boys' group (F(4, 844) = 205.66, p < .001; [[eta].sup.2] = .49) and among the girls' group (F(4, 792) = 232.19, p < .001; [[eta].sup.2] = .54) significant within-subject domain differences were recorded.
Age differentiation The central theoretical premise of the construct of parental authority, derived directly from its principle of dynamics, defines it such that its overall level depends on the child's age. The current instrument's ability to reflect significant decrease in parental authority as a function of age constitutes an essential condition for proving its construct validity. To examine the current hypothesis we calculated the general correlations between these variables. The results in Table 6 present a significant negative correlation, of a medium size, between the instrument's overall authority score and the age variable (r= -.46, p < .001) in the general research group. An identical trend is also apparent among both the boys (r= -.49, p < .001) and the girls (r= -.43, p < .001), as can be seen in Table 3. These findings support the age differentiation hypothesis, providing support for the current layer of the instrument's construct validity.
Discernment from social desirability The findings in Table 6 show a weak but significant correlation between the overall authority scale and the social desirability scores (r = .17, p < .05) in the general research group. The correlations with the three authority scales composing the overall scale indicate an identical negligible bias related to social desirability, while the connection with the Power 2 scale is not significant. An attempt to trace the origin of this connection among the sex groups leads to the girls' group, in which a significant correlation of a moderate size was recorded between the overall authority scale and the social desirability scores (r = .28, p < .05). The correlations among the boys, in contrast, were not significant. Subject to this qualification, it appears that the instrument provides sufficient discrimination from social desirability.
STUDY 4: CONCURRENT VALIDITY
The primary aims of the current study are examining the validity of the parental authority instrument versus the concurrent criterion of the child's conformity to the parents, obtained separately from the children and their parents using two different reporting methods.
Participants and procedure To the general research group described in Study no. 3 we added for the current study a group of parents of the children who participated in the previous study. To collect the parents' reports on the conformity behavior of the children from the general group, self-reporting research questionnaires were filled in by 130 parents, including 31 fathers and 87 mothers (12 did not provide sex information, but the parents' sex was not found to affect the reports). The data collection procedure for the general research group of children was described in the methods section of the previous study. The data collection among the parents took place at two different times, when parent-teacher events were held at the different schools. On these occasions a representative of the researcher arrived at school and approached the parents waiting for their meeting with the teacher with a request to fill in a short report about their child. Parents were informed about the ethics measures taken in order to guarantee their child's anonymity.
Measures Parental Authority Instrument (PAI). As described in Study no. 3.
Conformity to Parent Criterion. The concurrent validity of the Parental Authority Instrument was tested against conformity scores. The child's conformity index serves in studies to estimate the child's actual responsiveness and obedience to the parent's expectations and demands. The collecting of data regarding this criterion in the current study took place in two separate evaluations, from two different sources. Parents' reports about their child's conformity were obtained using the obedience versus resistance scale from the Parent Report of Child Behavior Toward Parent questionnaire (Schafer, 1975). This assessment was intended to provide an index regarding the child's degree of conformity to the parent's demands, based on the latter's subjective impression of the child's obedience. Two scales, obedience and resistance to control, containing 10 items, were taken from the general questionnaire. In the current study, reliability data were recorded as sufficient internal consistency for these two scales, ranging between .74 and .69. The mean conformity score obtained here on the basis of a 1 to 4 response scale was 3.05, with a standard deviation of .49, with no effect of the child's sex or the parent's sex on the measured variable. The children's report regarding this criterion was obtained using the conformity scale of Thomas, Gecas, Weigert & Roony, 1974. The scale includes 9 statements reflecting hypothetical situations in parent-child relations, phrased in the first person, to which the child is requested to respond by agreeing (obeying) or disagreeing (disobeying) with his parents, on a four-level response scale (with the higher score indicating stronger conformity). This scale has been used in many studies to measure conformity to parents in different empirical contexts (e.g.: Ghazrian, Supple & Plankett, 2008; Henry et al., 1989; Peterson, Rollins & Tomas, 1985), and served as a criterion to test the construct validity of a questionnaire of parental behavior in different cultures (Supple & Peterson, 2004). The scale in its Hebrew version recorded in the current study a sufficient internal consistency for its nine items ([alpha] =.72), with a mean conformity score of 2.19 and a standard deviation of .44, without any effect of the respondent's sex on the score. The translation of the two tools into Hebrew was conducted using the back-translation method, i.e., first translating into Hebrew, then back into English, and finally back into Hebrew.
It was hypothesized that the instrument under examination would predict significantly the scores of a concurrent criterion of a child's conforming behavior to the parent, obtained from two related sources. The relevant findings are discussed below according to the reporting source regarding the criterion:
A. Children's reports: Regarding the overall study group of children (N = 412), the general correlations (zero order) were calculated between the conformity scale and the scores of the scales obtained from the parental authority instrument (Table 7).
The overall authority scale scores have medium positive and significant correlations with the children's conformity reports, which affirms the validity of the instrument. This general trend is also apparent in the four scales. The multiple regression analysis affirms that each of the instrument's scales provides a significant unique contribution to explaining the criterion, which indicates the unique importance of each of them regarding the relevant variables. Together the tool's four scales explain about 38% of the variance in the conformity scores (F(4, 402) = 63.17, p < .001). The equivalent findings in the sex groups match these trends, presenting significant connections between the overall authority scale and the conformity criterion in both the girls' group (r = .63, p < .001) and the boys' group (r = .54, p < .001).
B. Parents' reports. Due to the risk of inflation involved in measuring correlations relying on reports from an identical source (Campbell & Fiske, 1959), we sought to examine the instrument's predictive ability with an equivalent behavioral criterion obtained from independent reports by the children's parents. Since the parent's reported sex had no effect on the scores of the current index, all the relevant findings are presented as one group, without distinguishing between mothers and fathers. The relevant data in Table 7 show significant correlations with the criterion scores for all the parental authority instruments' scales, as well as for the overall authority scale. A subsequent multiple regression analysis showed that the unique contribution of the instrument's scales to explaining the scores of the criterion is not significant. However, in total the general regression equation for predicting the criterion by the four scales of the parental authority instrument explains about 14% of the variance in the scores of this index (F(4, 124) = 5.00, p < .001). Consistent with the concurrent validity hypothesis, significant correlations were found between the overall parental authority scores and the conformity index for the two sex groups. However, whereas only a small correlation obtained in the girls' group (r = .25, p < .05), a medium size correlation was found in the boys' group (r = .42, p < .001).
STUDY NO. 5: EXAMINING THE DISTINCTION BETWEEN CONTRASTING GROUPS
As part of the examination of the instrument's construct validity, a criterion of comparison between contrast groups in relation to parental authority was determined. The aim of the current study was to compare a group of parents of normative children and a group of parents of children with a delinquent background (hereafter: the clinical group), who exhibit a lack of parental authority. This is to prove the instrument's ability to distinguish between the groups through the children's reporting.
Measures Parental Authority Instrument (PAI). As described in Study no. 3. The current study used 22 out of the questionnaire's 28 items in its final format (without items 4, 9, 10, 11, 12, 13), relevant in content to the demographic and situational features of the clinical group respondents (for example, items from the household routine group are irrelevant to children residing in an institution).
Participants and procedure To compare the scores of the Parental Authority Instrument obtained from the general study group described in Study 3, an additional group of adolescents was sampled, including 30 boys and girls, from two different centers for treating youth at risk. Treatment center A is an educational-therapeutic home for adolescents at risk with a background of delinquency, in which educational and therapeutic activities are held in a boarding school setting. From this institution, 14 resident adolescents were sampled, including 5 girls, aged between 15 and 17 (M=16.5, SD=0.65), with a varied delinquent background involving a criminal record. In educational terms, this group belongs to three age levels: ninth grade (n=6), tenth grade (n=3), and eleventh grade (n=5). The second treatment center is part of a project for advancing and assisting adolescents at risk operated by the municipal welfare office of a town in northern Israel. The local social worker individually supervised the distribution of the questionnaires, and was guided to include adolescents whose behavioral background includes an explicitly delinquent nature. Sixteen adolescents filled in the questionnaires (14 boys, 2 girls), in an age range between 15 and 17 years (M=15.9, SD=0.77), from three age levels: eighth-ninth grade (n=3), tenth grade (n=9), and eleventh grade (n=4).
From the general research group a control group was selected as a comparison group contrasting to the clinical group only from participants from the two higher age levels in the general sample, whose age range (15-17) and mean age (16.15) were identical to the mean and age range in the clinical group, based on an identical age distribution between the two groups (F = 3.23, p > .05).
According to the current research hypothesis, it was expected that the parental authority scores obtained from the instrument would be significantly lower in the clinical group. For purposes of the test, a t test for independent samples was used to compare the mean authority scores of the clinical group and the scores of the two upper age levels from among the general group (tenth and eleventh grades).
Consistent with the hypothesis, the findings of Table 8 show that there are clear and significant differences between the two groups on the overall authority scale scores (t = 3.42, p < .005, d = .77), whereby adolescents with a delinquent background tend to attribute to their parents less authority relating to the issues appearing in the questionnaire. Differences in this direction are also significantly apparent in the instrument's other authority scales, apart from the 'legitimacy 2' (right to demand) scale, in which identical but not statistically significant trends were recorded.
The current work reflects a research-methodological attempt to develop a means of measuring the theoretical structure of parental authority, based on the bi-dimensional model of the construct (Yaffe, 2014). On the basis of the conceptual differentiation from the classical concept of the "authoritative parent," the current measurement approach can be seen as an attempt to extricate from the theoretical structure of "parenting style" the feature of "parental authority," and to create an operational distinction between them. Accordingly, the purpose of the current study was to test and affirm the validity of the proposed questionnaire as an instrument that measures parental authority through children's reports, on the basis of a list of criteria accepted in measurement evaluation.
The accumulated evidence shows that the current Parental Authority Instrument meets the minimum requirements for the validity of a measuring instrument and sufficiently meets the specific criteria set out for this purpose. In this context, the current instrument presents: data regarding the internal structure as required; distinction between the domains of parental authority; sufficient ability to predict the conformity criteria determined; a sufficient level of differentiation from the social desirability variable; scores in an inverse relation to the respondent's age; and the ability to distinguish between contrasting groups in accordance with the theoretical prescription.
Regarding its internal structure, most of the items on the questionnaire's four scales that compose the overall authority score produced mean scores around the middle of the range of the response scale, with item-scale correlations of a medium power, indicating a good distinguishing ability of the items. I also obtained high internal reliability coefficients for each scale. The instrument's four scales defining the two theoretical dimensions of the construct of parental authority--legitimacy and power--presented a medium size and above correlative match in each dimension, in a manner agreeing with the apriori premise regarding the required linkage between them. These authority scales actually constitute distinct operational aspects of each theoretical dimension, which are supposed to bond due to their very conceptual and logical connection. Regarding the legitimate authority dimension, the relatively high intra-dimensional connection obtained from the two gender groups between the "right to demand" and "obligation to obey" scales reflects the conclusion of Darling et al. (2008), that these are indeed two different aspects of parental authority. The power dimension in the questionnaire, operationally defined by the scales "demanding" and "potential influence," recorded a medium power and significant intra-dimensional connection. This agrees with the premise whereby parental demanding is essential for defining parental authority, since it provides meaning to the other aspects that define this theoretical structure. The absence of clear rules empties the necessity of parental authority, in light of the absence of parent-child conflict. The consistent correlations found between parental demanding and the other authority scales of the instrument support this idea.
Regarding the domains of parental authority (according to the overall authority scale), based on the findings of the research body about conceptions of parental authority, we created a distinction between the conflict areas of content in the definition of this construct, which found empirical support through factor analysis. In general, the results of the analysis according to five factors led to a stable and balanced solution, consistent with the preliminary theoretical rationale of issues attribution according to domains. Of the 28 items included in the analysis, ten were allocated to the personal realm, and the rest were almost evenly divided between the four remaining content areas: routine, child's safety and welfare, family behavioral conventions, and the social domain. Subject to a number of semantic nuances, these content areas, with the issues they include, are essentially similar to the domains found in the studies of Smetana and her colleagues (Smetana, 1988, 1993, 1995, 2000; Smetana & Asquith, 1994; Smetana, Crean & Campione-Barr, 2005; Smetana & Daddis, 2002). However, the current questionnaire items did not include explicit "moral" issues (e.g.: stealing, cheating), since the willingness to obey rules that results from them is dependent on their subject matter rather than on parental demanding per se (Braine, Pomerantz, Lorber & Krants, 1991; Turiel, 2002). Hence, moral conflicts were not included in the current instrument, whose purpose is to measure extent of the child's responsiveness to parental authority. This is in contrast with the traditional measurement approach, which addressed the conceptions of parental authority in principle rather than in the individual measuring of this variable.
Another innovation proposed in the PAI is the portrayal of the child-parent disagreement in the situation described at the item itself rather than referring this element literally in only one operational context (e.g., in the question: "assuming you disagree, are you obligated to obey?"). In our opinion, such a construction authentically and non-artificially reflects the conflict potential derived from the very situation presented rather than from an imposed apriori premise.
The domain effect on the level of the overall parental authority measured by the instrument among the research group is generally consistent with the theoretical dynamic characteristic of the construct. The differences in authority scores found between the questionnaire's domains reflect the premise that parental authority is not equivalent in relation to different contexts, and that the level of parental authority is reduced as more personal issues arise. Thus the instrument proved its ability to reflect a basic theoretical trait of the construct of parental authority as affirmed by a long research chain in the area (Darling et al., 2005; Smetana, 1988, 1993, 2000; Smetana & Asquith, 1994; Tisak, 1986).
Evidence regarding the concurrent validity in the form of moderate to medium size correlations between the parental authority scales and two indexes of the conformity to the parent criterion from independent reporting sources were obtained in the general group and in the gender groups separately. More limited support for the ability to predict this criterion regarding the parent's index was obtained in the girls' group. However, the instrument's authority scales were found to have a unique significant contribution to explaining the criterion scores according to the children's index. The instrument's combined scales (i.e., the overall parental authority scale) predict the criterion's score more successfully beyond any individual scale, and any combination between the scales that does not include all four of them. This strengthens the assumption regarding the functional distinction of the four scales in terms of the importance of each one of them in respect to related variables. Against the background of the strong statistical relation between them, the authority scales' ability to provide a unique contribution to predicting the criterion supports their nature as essential and distinctive aspects of the construct.
The PAI is only minimally contaminated by social desirability, with the exception of a very moderate but significant connection found in the girls' group. I recommend monitoring the influence of social desirability only among the girls' group when using the instrument.
Further evidence for the validity of the PAI relies on the inverse connection obtained between the questionnaire's authority scales, headed by the overall authority scale, and the respondents' age among the groups of boys and girls in the sample. Consistency with the age variable is an essential, though insufficient, condition for proving the constrcut validity of instruments measuring this type of variables (Anastasi, 1990). A dual decrease in the parents' authority to influence the child (power) and the child's willingness to accept this authority (legitimacy) is apparent here. The instrument's scales reflect the specific aspects of parental authority that are influenced by developmental processes involved in the child's growth, against the background of progressive changes in moral thinking and the constant striving for autonomy. These results indicate the instrument's ability to reflect these changes that are described in the psychological approaches of moral development and in the attitude toward authority among children (Piaget, 1932/1965; Youniss & Smollar, 1985).
Comparing the overall level of parental authority between the upper age sample from the general study group and the "clinical" study group provided significant differences, in the expected direction, that were supported by a strong effect size. The findings strengthen the instrument's validity, presenting a significantly lower level of parental authority among a group of adolescents with a delinquent background and behavioral problems, in accordance with the theoretical outline dictated by the current construct. This affirms the basic assumption that the current instrument measures a theoretical construct (parental authority), in which empirical evidence links the lack of its presence with the outcomes of externalized emotional and behavioral problems and anti-social behavior among adolescents (see: Dekovic, Janssens & Van As, 2003; Dision & McMahon, 1998; Parket & Benson, 2004).
Apart from the limited sampling (method and size) that might compromise the instrument's validity, there are several important limitations to consider. First, the reasonable possibility that parental authority depends on the parent's gender (e.g. Laible & Carlo, 2004; Mckinney & Renk, 2008), was not considered in this study. All the questionnaires were phrased using uniform language to refer to both parents for reasons of convenience and on the assumption that the respondents attributed their answers on all the indexes to the relevant parent. However, one should take into account that the relevant parent differs among children and also among areas. Second, the method of unifying the questionnaire's authority scales employed in this study was basic, and did not reflect the possible differential weight of the dimensions and aspects of the theoretical construct. We should aim later on to develop a more complex weighting formula that would express relatively and functionally the differential weight of the construct's dimensions. This sort of formula must be based on a solid theoretical rationale.
The psychometric features of the PAI, the wide situational substrate it contains, and its foundational scoring method make the instrument suitable for both research and diagnostic purposes and enable a collective and individual assessment of parental authority among a wide range of adolescents.
The English version of the instrument (see note following References) requires further examination to reinforce the construct validity of the translated form. It would be essential to gradually replicate some of the core methodological operations described in the current study (e.g., testing age and domain differentiation) within groups of English speaking teenagers. This may also take place as part of a broader study using the instrument in diverse research contexts.
Ohalo Academic College, Katzrin, Israel
Tel-Hai Academic College, Israel
Author info: Correspondence should be sent to: Dr. Yosi Yaffe, Ohalo Academic College, P.O.B. 222, Katzrin 129000, Israel. email@example.com
Anastasi, A. (1988). Psychological testing. New York: Macmillan.
Baumrind, D. (1991). The influence of parenting style on adolescent competence and substance use. Journal of Early Adolescence, 11, 56-95.
Bloom, B. L. (1985). A factor analysis of self-report measures of family functioning. Family Process, 24, 225-239.
Braine, L.G., Pomerantz, E., Lorber, D. & Krantz, D.H. (1991). Conflicts with authority: Children's feelings, actions and justification. Developmental Psychology, 27(5), 829-840.
Buri, J. R. (1991). Parental Authority Questionnaire. Journal of Personality Assessment, 57(1), 110-119.
Campbell, D. T. & Fiske, D. W. (1959). Convergent and Discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105
Crowne D. P. & Marlowe D.A. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.
Darling, N., Cumsille, P., Caldwell, L.L., & Dowdy, B. (2006). Predictors of adolescents' disclosure strategies and perceptions of parental knowledge. Journal of Youth and Adolescence, 35, 667-678.
Darling, N., Cumsille, P., & Martinez, M.L. (2007). Adolescents' as active agents in the socialization process: Legitimacy of parental authority and obligation to obey as predictors of obedience. Journal of Adolescence, 30(2), 297-311.
Darling, N., Cumsille, P., & Martinez, M. L. (2008). Individual differences in adolescents' beliefs about the legitimacy of parental authority and their own obligation to obey: A longitudinal investigation. Child Development, 79(4), 1103-1118.
Darling, N., Cumsille, P. & Pena-Alampay, L. (2005). Rules, legitimacy of parental authority and obligation to obey in Chile, the Philippines and the United States. New Direction for Child and Adolescent Development, 108, 47-60.
Dekovic, M., Janssens, J., & Van As, N. (2003). Family Predictions of Antisocial behavior in adolescence. Family Process, 42(2), 223-235.
Dishion, T. J., & McMahon, R. J. (1998). Parental monitoring and the prevention of child and adolescent problem behavior: A conceptual and empirical formulation. Clinical Child and Family Psychology Review, 1, 61-75.
Fischer, D. G. & Fick, C. (1993). Measuring social desirability: Short forms of the Marlowe-Crowne social desirability scale. Educational and Psychological Measurement, 53, 417-424.
Floyd, F. J. & Widaman, K. F. (1995). Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7(3), 286-299.
Furman, W. & Giberson, R. S. (1995). Identifying the links between parents and their children's sibling relationships. In S. Shulman (Ed.), Close relationships and socioemotional development: Human development (Vol. 7, pp. 95-108). Westport, CT: Ablex Publishing.
Ghazarian, S. R., Supple, A. J., & Plunkett, S. W. (2008). Familism as a predictor of parent-adolescent relationships and developmental outcomes for adolescents in Armenian American immigrant families. Journal of Child Family Study, 17, 599-613.
Henry, C. L., Wilson, L. M., & Peterson, G. W. (1989). Parental power bases and processes as predictors of adolescent conformity. Journal of Adolescent Research, 4, 15-32.
Laible, D. J., & Carlo, G. (2004). The differential relations of maternal and paternal support and control to adolescent social competence, self-worth, and sympathy. Journal of Adolescent Research, 19, 759-782.
Larson, M. R. (2000). Social desirability and self-reported weight and height. International Journal of Obesity, 24, 663-665.
Maccoby, E. E. & Martin, J. A. (1983). Socialization in the context of the family: Parent-child interaction. In P. H. Mussen (Series Ed.) & E. M. Hetherington (Vol. Ed.), Handbook of child psychology: Vol. 4. Socialization, personality, and social development (pp. 1-101). New York: Wiley.
McKinney, C. & Renk, K. (2008). Differential parenting between mothers and fathers: Implications for late adolescents. Journal of Family Issues, 29(6), 806-827.
Messick, S. (1995). Validity of psychological assessment: Validation of inferences from persons' responses and performances as scientific inquiry into score meaning. American Psychologist, 50, 741-749.
Olson, D. H. & Cromwell, R. (1975). Power in families. In R. Cromwell & D. H. Olson (Eds). Power in families (pp. 3-11). New York: Halsted Press.
Parker, J. S. & Benson, M. G. (2004). Parent-adolescent relations and adolescent functioning: self-esteem, substance abuse and delinquency. Adolescence, 39(155), 519-530.
Peterson, G. W., Rollins, B. C., & Thomas, D. L. (1985). Parental influence and adolescent conformity--compliance and internalization. Youth & Society, 16(4), 397-420.
Piaget, J. (1932). The moral judgment of the child. London: Routledge and Kegan Paul; New York: Free Press.
Piaget, J. (1965). The moral judgment of the child. New York: Free Press Paperback.
Reitzle, M., Winkler Metzke, C., & Steinhausen, H. C. (2001). Parents and children: The Zurich brief questionnaire for the assessment of parental behaviors. Diagnostica, 47, 196-207.
Robinson, C. C., Mandleco, B., Frost Olsen, S., & Hart, C. H. (2001). The parenting styles and dimensions questionnaire (PSDQ). In B.F. Perlmutter, J. Touliatos, & G.W. Holden (Eds.), Handbook of family measurement techniques. Vol. 2: Instruments and index (p. 190). Thousand Oaks, CA: Sage.
Rohner, R. P. (2005). Parental Acceptance-Rejection/Control Questionnaire (PARQ/Control): Test manual. In R. P. Rohner & A. Khaleque (Eds.), Handbook for the study of parental acceptance and rejection (pp. 137-186). Storrs, CT: Rohner Research Publications.
Schafer, E. S. (1965). Children's Reports of Parental Behavior: An Inventory. Child Development, 36 (2), 413-424.
Schafer, E.S. (1975). Parent report of child behavior toward parent. Princeton, NJ: ETS.
Shelton, K. K., Frick, P. J., & Wootton, J. (1996). Assessment of parenting practices in families of elementary school-age children. Journal of Clinical Child Psychology, 25, 317-329. (Alabama Parenting Questionnaire)
Smetana, J. G. (1988). Adolescents' and parents' conceptions of parental authority. Child Development, 59, 321-335.
Smetana, J. G. (1993). Conception of parental authority in divorced and married mothers and their adolescent. Journal of research on adolescence, 3(1), 19-39.
Smetana, J. G. (1994). Beliefs about parenting: Origins and developmental implications. San Francisco: Jossey-Bass Publishers
Smetana, J. G. (1995). Parenting styles and conception of parental authority during adolescence. Child Development, 66, 299-316.
Smetana, J. G., (2000). Middle-class African American adolescents' and parents' conception of parental authority and parenting practices: A longitudinal investigation. Child Development, 71(6), 1672-1686.
Smetana, J. G. & Asquith, P. (1994). Adolescents' and parents' conceptions of parental authority and adolescent autonomy. Child Development, 65, 1147-1162.
Smetana, J. G., Crean, H. F., & Campione-Barr, N. (2005). Adolescents' and parents changing conceptions of parental authority. New Directions for Child and Adolescent Development, 108, 31-46.
Smetana, J. G. & Daddis, C. (2002). Domain-specific antecedents of parental psychological control and monitoring: The role of parenting beliefs and practices. Child Development, 73, 563-580.
Strahan, R. and Gerbasi, K. C. (1972). Short, homogeneous versions of the Marlowe-Crowne social desirability scale. Journal of Clinical Psychology, 28, 191-193
Streiner, D. L. (1994). Figuring out factors: The use and misuse of factor analysis. Canadian Journal of Psychiatry, 39, 135-140.
Supple, A. J., Peterson, G. W., & Bush, K. R. (2004). Assessing the validity of parenting measures in a sample of Chinese adolescents. Journal of Family Psychology, 18, 539-544.
Theriault, S. & Holmberg, D. (1998). The new old-fashioned girl: Effects of gender and social desirability on reported gender-role ideology. Sex Roles, 39 (1/2), 97-112
Thomas, D. L., Gecas, V., Weigert, A., & Roony, E. (1974). Family socialization and the adolescent: Determinants of self-concept, conformity, religiosity and counterculture values. Lexington: Lexington Book.
Tilgner, L., Wertheim, E. H., & Paxton, S.J. (2003). Effect of social desirability on adolescent girls' responses to an eating disorders prevention program. Social Desirability Effect, 211-216. Published online in Wiley InterScience.
Tisak, M. S. (1986). Children's conceptions of parental authority. Child Development, 57, 166-176.
Turiel, E. (2002). The culture of morality: Social development, context and conflict. Cambridge: Cambridge University Press.
Weber, M. (1968). Economy and society: An outline of interpretive sociology. (Guenther Roth and Claus Wittich, Eds.). New York: Bedminster Press.
Yaffe, Y. (2013). "Parental authority": What do we know about the construct? International Journal of Educational Research and Development, 2(9), 211-219.
Yaffe, Y. (2014). "Parental authority": Theoretical conceptualization: The bi-dimensional model. Dapim, 58, 106-127 [Hebrew].
Yau, J. Y., Tasopoulos-Chan, M., & Smetana, J. G. (2009). Disclosure to parents about everyday activities among American adolescents from Mexican, Chinese, and European backgrounds. Child Development, 80, 1481-1498.
Youniss, J. & Smollar, J. (1985). Adolescents' relations with mothers, fathers and friends. Chicago: University of Chicago Press.
Note: The PAI in English or Hebrew is available at no cost at https://www.researchgate.net/publication7311676894_Development_and _validation_of_a_new_Paretal-Authority_Instrument_PAI
TABLE 1 Means, SDs, & Correlation Matrices of Parental Authority Scales & Instrument Validation Dimensions Scale/Index (1) (2) (3) (4) (5) (6) 1.Conformity -- to parent 2.Power 1 -.06 -- (demanding) 3.Power 2 .48 ** .04 * -- (pot.influence) 4.Legitimate 46 ** .45 ** 84 *** -- authority 1 (obligation) 5.Legitimate .57 ** .30 .78 ** .76 ** -- authority 2 (right) 6. Social .41 ** .37 * .48 ** .46 ** .57 ** -- desirability Scale average 2.33 1.46 1.83 1.77 1.81 4.96 Standard deviation .48 .56 .60 .60 .62 1.97 Note:N=40 * p< .05, ** p< .01, *** p< .001 TABLE 5 Means, SDs, & ANOVA Results for Differences Between Domains by Sex Boys Girls Combined Domain (#=212/213) (N=199) (N=411/412) (factor) Mean SD Mean SD Mean SD Personal 1.23 .60 1.16 .55 1.19 .58 Routine 1.74 .57 1.62 .58 1.69 .58 Family 1.77 .49 1.73 .48 1.75 .49 Conventions Social 1.67 .65 1.88 .57 1.77 .62 Child's 2.25 .58 2.23 .54 2.24 .56 Welfare Total 1.73 .48 1.72 .44 1.73 .46 Within/between Subject Effect F sex = .112 F domain = 426.03 *** *** p <.001 TABLE 6 Pearson Coefficients Between PAI scales & the Child's Age & the Social Desirability Scores scale Age Social desirability (child's reporting) Power 1 (demanding) -.37 *** .13 * Power 2 (potential -.42 *** .12 influence) Legitimate authority 1 (obligation -.43 *** .17 * to obey) Legitimate -.37 *** .14 * authority 2 (right to demand) Overall -.46 *** .17 * authority scale * p < .05, *** p < .001 TABLE 7 Pearson Coefficients between PAI Scales and the Criterion Indexes Conformity to Obedience/resistance to parent parent (child's report) (parent's report) Power 1 .30 *** .28 *** (demanding) Power 2 .55 *** .28 *** (potential influence) Legitimate .59 *** .35 *** authority 1 (obligation to obey) Legitimate .56 *** .30 *** authority 2 (right to demand) Overall authority .58 *** .36 *** scale * p< .05, *** p<.001 TABLE 8 Means, SDs, & t Scores for the Difference Between the Clinical & General Group on the Four Authority Scales & the Overall PAI Scale General Clinical t Scale M SD M SD Power 1 1.30 .42 .85 .62 3.78 ** (demanding) Power 2 1.36 .54 .83 .65 4 70 *** (potential influence) Legitimate authority 1 1.40 .50 .91 .69 3.61 ** (obligation to obey) Legitimate authority 2 1.53 .54 1.38 .74 1.01 (right to demand) Overall authority scale 5.59 1.68 3.97 2.45 3.42 ** Note: N between 30 (clinical) and 121 (general) * p<.05, ** p<.01, *** p<.001
|Printer friendly Cite/link Email Feedback|
|Publication:||North American Journal of Psychology|
|Date:||Mar 1, 2017|
|Previous Article:||The effects of making reading inferences on mastery goals, self-esteem and TOEIC reading comprehension.|
|Next Article:||The role of alcohol consumption and romantic attachment insecurity as risk factors for disrupted sleep and emotion regulation among underage and...|