Validation of Bangla parenting style and dimension questionnaire

The study was conducted by complying the declaration of Helsinki 1964. Formal permission was obtained on May 09, 2016 from Clyde C. Robinson (Robinson et al., 1995) who developed the questionnaire. Before starting this study, the research protocol was approved by the Institutional Review Board (IRB) of Bangabandhu Sheikh Mujib Medical University (BSMMU) (BSMMU/2016/10524), Dhaka, which as approved on October 15, 2016 and the protocol was reviewed on 123th IRB meeting held on October 08, 2016. Different steps of the study were supervised by the Department of Psychiatry, BSMMU. Informed written consent was obtained from parents without any influences. Parents were approached through 3 selected schools with prior permission from the school committee as well as from the headmasters. Data were collected anonymously, confidentiality of the data was ensured adequately, and any unauthorized access to data was not possible.

The adaptation of PSDQ into Bangla was performed according to the state-of-the-art procedure of forward–backward translation (Beaton et al. 2000; Arafat et al. 2016; Arafat 2016). One medical graduate and 1 lay person were involved in forward translation procedure. Both of the translators were native speakers of the Bangla and fluent in English. The forward translated versions were compiled from the two forward translations to address the discrepancies between the versions, so that standard version of a forward translation can be made, which is recommended for standard translation process (Beaton et al. 2000; Arafat et al. 2016). Then the compiled version was translated back into English by a professional translator with experience in medical translation and by one medical graduate who had not been involved in forward translation steps. The back-translated versions were then compiled and compared by the researcher to resolve any discrepancy for complying the validation guideline (Beaton et al. 2000; Arafat et al. 2016). After that, all the four versions were submitted to the expert committee formulated for this validation study. The expert committee modified and finalized the items of the PSDQ Bangla, and then pretesting was conducted among 30 parents who were approached from general population. During the expert committee review, some items were discarded and some were merged into single item to ensure the cultural equivalence (Beaton et al. 2000; Arafat et al. 2016).

The original instrument (PSDQ English) has 62 items (Table 1). The authoritative domain has 27 items: 1, 3, 5, 7, 9, 12, 14, 16, 18, 21, 22, 25, 27, 29, 31, 33, 35, 39, 42, 46, 48, 51, 53, 55, 58, 60, and 62. The authoritarian domain has 20 items: 2, 6, 10, 13, 17, 19, 23, 26, 28, 32, 37, 40, 43, 44, 47, 50, 54, 56, 59, and 61. The permissive domain has 15 items: 4,8, 11, 15, 20, 24r, 30, 34, 36, 38r, 41, 45, 49, 52r, and 57 (“r” indicates the reverse codes as the items were negatively phrased). The PSDQ uses response scales on a continuum from “Always (5)” to “Never (1).” Parents received a score on all three parenting dimensions.

Item deletion was considered when items having less than 0.80 content validity index in agreement with Arafat et al. (2016, 2017). Generally, content validity can be assessed through standard back translation process, reviewing literature, and skillful panel opinion and content validity index (Arafat et al. 2016, 2017; Arafat 2016; Algin et al. 2018). Item 60 (I channel our child’s misbehavior into a more acceptable activity) was deleted from authoritative domain. No item was deleted from the authoritarian domain. Item 11 (I spoil our child), item 24 (I appear confident about parenting abilities), item 38 (I carry out discipline after our child misbehaves), item 49 (I bribe our child with rewards to bring about compliance), and item 52 (I set strict well-established rules for our child) were discarded from permissive domain (Table 1).

Item merging was performed as per the recommendation of the expert committee during the expert committee review. Item 7 (I joke and play with our child) and item 46 (I have warm and intimate times together with our child) were merged into a single item. Item 9 (I show sympathy when our child is hurt or frustrated) and item 12 (I give comfort and understanding when our child is upset) were merged. Item 22 (I allow our child to give input into family rules) and item 55 (I take into account our child’s preferences in making plans for the family) were merged into single item. Item 29 (I help our child to understand the impact of behavior by encouraging our child to talk about the consequences of his/her own actions) and item 58 (I explain the consequences of the child’s behavior) were merged. Item 48 (I encourage our child to freely express (himself) (herself) even when disagreeing with parents) and item 51 (I show respect for our child’s opinions by encouraging our child to express them) were merged (Table 2).

Item 6 (I spank when our child is disobedient), item 19 (I grab our child when he/she is being disobedient), item 37 (I use physical punishment as a way of disciplining our child), item 43 (I slap our child when the child misbehaves), item 61 (I shove our child when the child is disobedient) were merged into single item and 4 items were reduced. Item 10 (I punish by taking privileges away from our child with little if any explanations) and item 28 (I punish by putting our child off somewhere alone with little if any explanations) were merged. Item 13 (I yell or shout when our child misbehaves) and item 32 (I explode in anger towards our child) were merged. Item 17 (I scold and criticize to make our child improve) and item 50 (I scold or criticize when our child’s behavior doesn’t meet our expectations) were merged. Item 23 (I argue with our child) and item 44 (I disagree with our child) were merged. Item 40 (I tell our child what to do) and item 59 (I demand that our child does/do things) were merged (Table 2).

Item 4 (I find it difficult to discipline our child) and item 57 (I appear unsure on how to solve our child’s misbehavior) were merged. Item 15 (I allow our child to annoy someone else) and item 45 (I allow our child to interrupt others) were merged. Item 20 (I state punishments to our child and does not actually do them) and item 34 (I threaten our child with punishment more often than actually giving it) were merged (Table 2).

Table 2

Merging of items.

Although the original PSDQ pertains 62 items after content validity index, the expert committee meeting the prefinal PSDQ Bangla pertained 39 items. The distribution of the items as per domain is given as follows: authoritative items (21 items) 1, 3, 5, 8, 11, 12, 14, 16, 17, 19, 21, 22, 23, 25, 27, 30, 32, 33, 35, 37, 39; authoritarian (11 items) 2, 6, 9, 10, 13, 18, 28, 31, 34, 36, 38; and permissive (7 items) 4, 7, 15, 20, 24, 26, 29.

This validation study was conducted from January 2017 to June 2018 in three schools of Dhaka city with the self-reporting semi-structured Bangla version of PSDQ. The purposively selected schools were the Engineering University School & College and Nabakumar Institute & Dr. Shahidullah College, and Academia. Engineering University School & College and Nabakumar Institute & Dr. Shahidullah College are Bangla medium school and Academia is an English medium school. Nabakumar Institute & Dr. Shahidullah College was considered as lower middle social class, Engineering University School & College was considered as middle-class background, and Academia was considered as representation of upper class in socioeconomic background. Parents having children aged 6–16 years were approached through schools to participate in the study. Sample size was estimated based on 5:1 ratio of participants to item as different recommendations are available regarding the ratio with consideration of rules of thumb as participants to variables 2:1 to 20:1 (Arafat et al. 2016; Arafat 2016; Anthoine et al. 2014). With proper permission and assistance, author visited the schools. Then the author visited the classrooms to convince the children to deliver the questionnaire to their parents and to get back it on a fixed date. Author revisited on the assigned date and collected the responded questionnaire. A total 502 questionnaire were distributed, and data were collected from 349 parents. After collection, data were analyzed using Statistical Package for the Social Science (SPSS) version 16.0 software.

Sociodemographic variables were analyzed and presented as frequency distribution. The response distribution of PSDQ Bangla was in 5-point Likert scale ranging from 1 (lowest) to 5 (highest). Thus, the item mean was calculated on the same distribution. Psychometric properties of PSDQ Bangla were assessed as per standard recommendations (Arafat et al. 2016, 2017; Arafat 2016; Algin et al. 2018). Reliability was assessed in internal consistency form by Cronbach’s alpha coefficient and a cut-off of ≥0.70 was considered acceptable, which is suggested by standard recommendations (Arafat et al., 2016). Face validity was assessed by following standard translation, back translation process, and expert committee review, which is supported by recommendations (Arafat et al. 2016, 2017; Beaton et al. 2000; Arafat 2016; Algin et al. 2018; Parsian and Dunning 2009). Content validity was also assessed by following standard translation process and content validity index, which is supported by existing recommendations (Arafat et al. 2016, 2017; Beaton et al. 2000; Arafat 2016; Algin et al. 2018; Parsian and Dunning 2009). Construct validity was assessed using exploratory factor analysis with principal component analysis (PCA) with varimax rotation. Factor rotation was performed to ascertain the factors of the construct that is preloaded in the SPSS software. Criterion validity could not be assessed because of the lack of culturally appropriate instrument.

The aim of this study was to assess the psychometric properties of the PSDQ Bangla. Internal consistency of PSDQ Bangla was 0.84 for the total score and subscales ranged from 0.78 to 0.95 (Table 3), which is significant for the construct as the Cronbach’s alpha ≥0.70 is considered as acceptable (Arafat et al. 2016; Arafat 2016). The original PSDQ with 62 questions revealed Cronbach’s alpha of 0.91 for the authoritative domain, 0.86 for the authoritarian domain, and 0.75 for the permissive domain (Robinson et al. 1995). So, the PSDQ Bangla revealed a nearly similar internal consistency of the original scale. The Turkish validation study revealed that the internal consistency of PSDQ was 0.63, the authoritarian parenting style was 0.71, the authoritative parenting style was 0.84, and the permissive parenting style was 0.38 (Önder and Gülay 2009). The Italian validation study also revealed acceptable reliability in all three domains of the scale, although these were measured by confirmatory factor analysis (Tagliabue et al. 2014). The Chinese validation study revealed that the Cronbach’s alpha of each subscale was between 0.62 and 0.86 (Fu et al. 2013). The Lithuanian 32-item PSDQ revealed the Cronbach’s alpha for the authoritative domain was 0.85, the authoritarian domain was 0.76, and the permissive domain was 0.58 (Kern and Jonyniene 2012). The Persian study found that the Cronbach’s alpha for the authoritative and authoritarian domains was 0.86 and the permissive domain was 0.41 (Morowatisharifabad et al. 2016). The Albanian study found, Cronbach’s alpha for the authoritative domain was 0.82, the authoritarian domain was 0.67, and the permissive domain was 0.52 (Ismaili 2015). A recent review found that most of the selected articles (96.23%) showed adequate values for both the authoritative (0.71–0.97) and authoritarian (0.62–0.95) domains with variations in the permissive domain, which showed lower Cronbach’s alpha values (0.38–0.89) (Oliveri et al. 2013).

Face validity and content validity were systematically assessed and maintained during the development of the research instrument (PSDQ Bangla), which is supported by previous studies (Parsian and Dunning 2009; Arafat et al. 2016, 2017; Arafat 2016; Algin et al. 2018). Expert committee assessed every item of PSDQ Bangla by comparing the translations and back translations and ensured the required equivalences (Beaton et al. 2000; Arafat et al. 2016). Construct validity was assessed by the EFA of the principal component with varimax rotation. The EFA with the principal component with varimax rotation was used to detect the factorial structure in observed measurements. The KMO and Barlett’s test of sphericity was applied to the fitness of data for factor analysis based on the sample size, which ensured adequate sample size in the current study statistically. Similar steps of construct validity assessment were found in the Turkish validation study (Önder and Gülay 2009). However, criterion validity of PSDQ Bangla cannot be assessed because of the lack of culturally adapted instruments, which is supported by previous studies (Fu et al. 2013; Arafat et al. 2016, 2017; Arafat 2016). The number of factors of the PSDQ Bangla based on eigenvalues obtained from PCA and a total of three factors were found, which is similar with the original scale (Robinson et al. 1995). Other validation studies also revealed similar three factorial structures, that is, authoritative, authoritarian, and permissive domain, which is also supported by Baumrind’s ground work (Robinson et al. 1995; Önder and Gülay 2009; Fu et al. 2013; Olivari et al. 2013; Kern and Jonyniene 2012; Morowatisharifabad et al. 2016)