Participants
The sample included 212 Korean–American parents (59 self-reported as 1st generation immigrant, 24 reported as 1.5 generation immigrant, and 129 self-reported as 2nd generation immigrant parents) who had a child or children on the autism spectrum who transitioned to kindergarten within the past 6 years. For this study, the following definitions were used for immigrant generations: first generation Korean–American immigrant is defined as those who immigrated from South Korea to the United States as an adult and who completed elementary and secondary education (K-12) in Korea. 1.5 generation Korean–American immigrant is defined as those who immigrated from South Korea to the U.S. as a minor and who completed some or all of K-12 education in the U.S. Second generation Korean–American immigrant is defined as those who have a Korean heritage background who were born in the U.S. Second generation immigrants also completed K-12 education in the U.S. These definitions were specified verbatim on the survey form and participants were asked to choose the one that best described their immigration status.
The sample size was determined by conducting a power analysis using G*Power 3 software (Faul et al.,
2009) using the data from a pilot study conducted with 62 participants (Kim,
2019) addressing the same research questions. Results from the power analysis suggested that 200 participants would comprise an adequate sample size for this study.
Inclusion criteria for participating in the study were (1) Korean–American parents residing in the U.S., (2) respondent self-reported as having a child diagnosed with ASD, and (3) child with ASD transitioned to kindergarten within the last 6 years. All participants met the inclusion criteria based on their self-reported responses.
Participants were recruited via online platforms that are specifically targeted for Korean–American parents or Korean–American parents with a child with ASD. Examples include a closed Facebook group for Korean–American parents and a membership-based online discussion forum for parents with a child with ASD within a portal site for Korean–American mothers. In addition, a recruitment flyer was emailed to a community-based Korean–American organization called Korean Community Center, a non-profit disability service organization called Easterseals Southern California, and a Regional Center in Los Angeles, California. Personnel in these organizations shared the flyers with parents who may have a child with ASD. These recruitment modalities were carefully selected as a way to account for the anonymous nature of the survey and for the self-reported inclusion criteria (e.g., diagnosis of ASD).
Procedures
All procedures were approved by the Institutional Review Board of the University of California, Los Angeles (IRB # 20-00436). The survey instrument was created on Qualtrics, and it was disseminated via anonymous online link. The recruitment flyer, informed consent, and the survey were available in both Korean and English. They were first written in English, then translated by a professional translator. The translated version was reviewed and edited by the author, who is bilingual in English and Korean, to ensure accuracy.
Upon receiving the survey link, interested parents had an opportunity to review the information about the study and the informed consent document. The first three questions of the survey were used for validation purposes. The first question asked for the respondents’ consent to participate. The second question asked if the participant considered themselves as 1st, 1.5 or 2nd generation Korean–American immigrants, or if they do not consider themselves as Korean–American. If they chose the last choice, the survey was discontinued per the exclusion criteria. The third question asked if they had a child on the autism spectrum who transitioned to kindergarten within the last 6 years. If the answer was no, the survey was discontinued. At the end of the survey, participants had the option to enter their email address to receive a $10 Amazon e-gift card. Their email addresses were used for this purpose only, and they were deleted from the database afterwards. The median length of time for survey completion was approximately 15 min.
Measures
The survey instrument was adapted from The Family Experiences and Involvement in Transition (McIntyre et al.,
2007) and Elements for Transition to Kindergarten (Forest et al.,
2004). The adaptation process included adding culturally-relevant survey items such as cultural values, stigmas, linguistic barriers, and discrimination. Additionally, common characteristics of ASD (e.g., challenges in social communication, self-care skills, behavioral readiness, sensory needs, etc.) were taken into consideration when creating the survey.
The initial prototype was field-tested by eight professional and community members, six of whom were Korean–American bilingual mothers. Average duration was measured, and feedback regarding clarity of the survey questions was sought. Then the survey instrument was revised based on their feedback. Feedback regarding wording, content validity and relevance for the Korean–American population was also provided by the professional translator after the revision process. This version of the survey instrument was used for the pilot study with a smaller sample (N = 62). After the data from the pilot study was analyzed, additional demographic variables such as annual income, parents’ educational level, child’s services received were added to be used for the current study. Cronbach’s alpha and mean inter-item correlations revealed adequate internal consistency for the three subdomains (Table
1).
Table 1
Reliability analysis of the survey instrument
Priorities in school readiness skills | 0.827 | 0.370 |
Priorities in parent support | 0.707 | 0.188 |
Challenges and barriers | 0.868 | 0.410 |
The survey measured the following three domains: (1) parent demographics including cultural and linguistic characteristics (e.g., generation of immigration, English fluency, etc.) and child characteristics, (2) parents’ concepts of successful TTK (i.e., priorities in school readiness skills and parent support), and (3) challenges and barriers during the transition process.
Demographic information of the participants included generation of immigration, years of residence in the U.S., language spoken at home, family annual income, employment status, educational level, and their relationship to the child. Responses in the parent demographics domain were measured using multiple choice questions. Moreover, child characteristic information included their current school placement, child’s current grade, child’s diagnosis other than ASD, sex, type of preschool attended, child’s main mode of communication, and services that they are currently receiving.
Parents’ Perspectives on Successful TTK
Parent perspectives on successful TTK were measured by the following two subdomains: (1) parents’ priorities for their child’s school readiness skills, and (2) importance of parent support in order to help their child to have a successful TTK.
In the first domain, the participants were asked, “In your opinion, how important is it to have the following skills in order to have a successful transition to kindergarten? Please rank them (1 = most important, 8 = least important). This subdomain (i.e., priorities for their child’s school readiness skill) had eight items (‘Academic readiness,’ ‘Behavioral readiness,’ ‘Social skills,’ ‘Self-help skills,’ ‘Cooperation with teachers,’ ‘Making needs known to others,’ ‘Adjusting to new environment and routine,’ ‘Separate well from parents in the morning’).
In the second domain, the participants were asked, “In your opinion, how important is it for parents to have the following skills in order to help their children to have a successful transition to kindergarten? (1 = most important, 10 = least important). This subdomain (i.e., importance of parent support) had 10 items (‘Parents’ knowledge on academic expectations, ‘Parents’ knowledge on behavioral expectations,’ ‘Building positive relationship with educators,’ ‘Having open communication with teachers when concerns arise, ‘Parents’ knowledge on U.S. school system,’ Parents’ knowledge on how to access appropriate services (e.g., speech therapy, occupational therapy, etc.),’ ‘Providing academic support at home,’ ‘Having strong relationship with other parents at school,’ ‘Being involved in school events,’ and ‘Being involved in community resources related to ASD’).
Challenges Experienced During TTK
In the third domain, respondents were asked, “To what extent did the following items serve as barriers in helping your child to have a successful transition to kindergarten?” There were 11 items as follows: ‘Linguistic barriers,’ ‘My lack of understanding on ASD,’ ‘My lack of knowledge about US school system,’ ‘My family’s lack of understanding on ASD,’ ‘How others may view my child’s disability/stigma,’ ‘Lack of support from my spouse,’ ‘Difficulty in building authentic relationship with educators,’ ‘Being uncomfortable discussing my child’s disability with others,’ ‘Lack of time/scheduling issues,’ ‘Difficulties in coming to an agreement in regards to placement, support, or services,’ and ‘Other life stressors.’ These items were measured using a 4-point Likert-style scale, and each point on the scale was be labeled as either “Not At All,” “Somewhat,” “Quite,” or “Very Much.” At the end of the section, participants were asked to write what other life stressors they experienced if they were not mentioned in the survey.
Analyses
To determine demographic information of the respondents and their child characteristics, frequency counts and percentages were calculated using SPSS version 27(IBM Corp. Released 2020).
Priorities for Successful Transition to Kindergarten
To answer the first research question in determining how Korean–American parents conceptualize successful transition to kindergarten, the ranking scores from the two subdomains were reverse-coded using SPSS version 27 (IBM Corp. Released 2020). The pilot study (Kim,
2019) utilized a four-point Likert-type scale for these two subdomains (i.e., Priorities in school readiness skills, Priorities in parent support) but demonstrated insufficient variances with generally high means. Moreover, the differences in mean were mostly statistically insignificant. In order to address this limitation, a decision was made to use ranking scores instead for these two subdomains in the current study.
Listwise deletion was used to handle missing data. Mean and standard deviation were calculated, and one-sample t-test was conducted to estimate 95% confidence interval for each survey item within the two subdomains.
Challenges and Barriers
Multiple imputation was first conducted for the variables in the Challenges and Barriers domain to address the missing data using R (R Studio Team,
2020). Missingness ranged from 0.14 to 2.36% prior to conducting missing imputation. Sources of missingness was unknown. However, a Test of Missing at Completely Random (Little,
1988) was conducted and the result suggested that the data was missing completely at random.
To answer the second research question, descriptive statistics were conducted to determine the extent to which Korean–American parents reported to experience various challenges during their children’s TTK.
Predictors for Challenges and Barriers
To explore possible predictors for Korean–American parents’ reported challenges during their child’s TTK, multiple regression was conducted. First, a new variable was created that reflected the sum of all the variables within the Challenges and Barriers domain. This variable was used as the DV for the model. The following variables entered the model as predictor variables in the following order hierarchically: (1) immigration generation, (2) responding parent’s primary language, (3) family’s annual income, (4) child’s mode of communication, (5) child’s school placement, (6) preschool types, (7) responding parent’s employment types, (8) co-occurring diagnosis other than ASD, (9) child’s current grade, and (10) parent’s education level. Dummy variables were created for the following predictors due to the categorical nature of the variables: immigrant generation, responding parent’s primary language, child’s mode of communication, child’s school placement, preschool types, and responding parent’s employment status. Family’s annual income was coded as follows: 1 = $15,000 or less; 2 = $15,001–$35,000; 3 = $35,001–$65,000; 4 = $65,001–$100,000; 5 = $100,000–$150,000; 6 = $150,001–$200,000; 7 = $200,001 or more. Child’s current grade entered the model in order to control for possible confound due to their varying age—Participants’ TTK experiences may be very recent for those with a kindergartener, but it may be distant for those with a 5th grader.
Assumptions for multiple regression were tested for the outcome variable and the possible predictors that would enter the model. To test the assumption of linear relationship between the IVs and the DV, scatter plots were generated between the IVs and the DV. Upon visual analysis, no violation of assumption was detected. In order to test the assumption that there is no multicollinearity in the data, analysis of collinearity statistics was conducted. VIF scores were well below 10 except for the two dummy variables (‘Preschool type: Public special education preschool’ and ‘Private preschool with English speaking staff’) with the VIF scores of 9.236 and 8.320 respectively. Further, the tolerance scores were all above 0.2, except for the aforementioned two variables with tolerances scores of 0.108 and 0.120 respectively. Since the variable ‘Preschool type: Public special education preschool’ was a significant predictor with p < 0.05 in this model, a decision was made to remove the variable ‘Private preschool with English speaking staff’ for the final model. The final model met the assumptions of multicollinearity with the VIF scores ranging from 1.084 to 1.579, and the tolerance scores ranging from 0.633 to 0.911. To test the assumption of independent residuals, the Durbin–Watson statistic was conducted. The Durbin–Watson statistic for the final model was 1.446. To test the assumption that the residuals are normally distributed, the P–P plot for the model was generated. The P–P plot for the final model suggested that the assumption of the normality has been met. Lastly, to test the assumption that there are no influential cases biasing the final model, Cook’s Distance was calculated. Cook’s Distance values were all under 1, suggesting individual cases were not unduly influencing the model. In summary, no violations to these assumptions were found in the final model. The above analyses were conducted using SPSS version 27 (IBM Corp. Released 2020).