Early intervention in autism spectrum disorder (ASD) is critical to improve early deficits and function in later life [Bradshaw, Steiner, Gengoux, & Koegel, 2015; Estes et al., 2015; Mundy & Crowson, 1997]. Identification of ASD at a young age allows early intervention, but early diagnosis is not always possible. According to the literature, early signs of ASD are generally detected and diagnosed as early as 12–24 months of age [Boyd, Odom, Humphreys, & Sam, 2010; Kleinman et al., 2008; Zwaigenbaum et al., 2009]. However, the average age of diagnosis is much higher in the United States and Korea [Baio et al., 2018; National Institute of Special Education, 2015]. This discrepancy stems from time-consuming evaluations, expensive care systems, long waiting times, shortage of professionals, and low awareness of paraprofessionals [Althouse & Stockman, 2006; Austin et al., 2016; Bisgaier, 2011; Fenikile, Ellerbeck, Filippi, & Daley, 2014; Kalb et al., 2012; Sunwoo, Noh, Kim, Kim, & Yoo, 2017; Wiggins, Baio, & Rice, 2006]. Additionally, in Korea, two major reasons cited for the lag between when parents first suspect symptoms and the time of diagnosis are (a) the belief that symptoms will resolve as the child grows (52.2%) and (b) doctor recommendations to delay diagnosis in young children (39.7%). This highlights the relatively limited awareness of ASD in Korean society [National Institute of Special Education, 2015]. Screening is a pre-diagnostic method to identify early manifestation of disorders, and can be the first step of the diagnostic process [Corbisiero, Hartmann-Schorro, RiecherRössler, & Stieglitz, 2017]. Typically, screening for ASD has been performed with screening instruments in the form of caregiver-rated questionnaires or simple observation by trained professionals [Towle & Patrick, 2016]. Examples of existing screening instruments are the Modified Checklist for Autism in Toddlers (M-CHAT) [Robins, Fein, & Barton, 1999; Robins, Fein, Barton, & Green, 2001], the Screening Tool for Autism in Toddlers and Young Children [Stone, Coonrod, & Ousley, 2000; Stone, Coonrod, Turner, & Pozdol, 2004; Stone, McMahon, & Henderson, 2008], the Social Communication Questionnaires (SCQs) [Rutter, Bailey, Lord, & Berument, 2003], and the Social Attention and Communication Study (SACS) [Barbaro & Dissanayake, 2010]. Of these, the M-CHAT and SCQ have been translated into Korean [Kim et al., 2015; Yoo, 2008]. Previous screening instruments for ASD have adopted either cutoff scores or at-risk behavior approaches. Both have advantages, but some studies have suggested that the identification of toddlers with ASD using cutoff scores may be ineffective [Corsello, Cook, & Levanthal, 2003; Newschaffer, Lee, David, & Lee, 2004] because scores can vary across cultures [Chiang et al., 2012; Kim et al., 2015]. Instead, at-risk responses from follow-up protocols in M-CHAT, CHAT, and SACS can identify specific characteristics of ASD and are used to collect additional information and to reduce the occurrence of false-positives [Robins et al., 1999]. Further, parental interviews or questionnaires may also result in reporter bias or subjective responses [Glascoe, 2000; Stone, Hoffman, Lewis, & Ousley, 1994]. Some screening tools are administered at a specific age, and many at-risk children may be missed due to their age [Barbaro & Dissanayake, 2010]. For example, joint attention (JA) is one of the critical factors used to distinguish children at risk of ASD from typically developing children and children with other developmental disabilities (DDs).