An individual would generate and construct mental representations of future events to make appropriate behavioral decisions and guide their behavior (Rochat et al., 2019). Furthermore, this process enables individuals to elaborate cognitively on the acquisition of the desired object (May et al., 2008). Such cognitive process, in the research field of craving, is termed desire thinking, and it is defined as a conscious and voluntary process that produces and elaborates desire-related content around an appetitive target (Caselli & Spada, 2015).

According to the elaborated intrusion theory of desire (EIT; Kavanagh et al., 2005, Kavanagh et al., 2009), some automatic associations activated by external or internal triggers will intrude into the consciousness level (e.g., desire target could make me happy). Individuals will process these associations cognitively (e.g., desire thinking). A feeling of pleasure or relief is evoked by engaging in desire thinking, which is considered to result in the motivational component of the desire experience (i.e., craving). However, the desire target is still not achieved or obtained but the craving is continued, which often results in engaging in the desired activity (e.g., addictive behaviors) to attain relief. It has also been demonstrated that desire thinking is positively related to craving (Caselli et al., 2013, Caselli et al., 2015, Caselli et al., 2017, Martino et al., 2017), as well as addictive behaviors (Albery and Spada, 2021, Caselli et al., 2012, Caselli et al., 2012). Recently, the interaction of person-affect-cognition-execution (I-PACE; Brandtner et al., 2021) model has also included desire thinking in order to explain the development and maintenance of the internet-use disorder. Specifically, internal and external triggers activate desire thinking through two pathways: A pleasure-oriented pathway (corresponding to gratifying expectations of desire thinking) and a relief-oriented pathway (corresponding to compensating expectations of desire thinking). The outcomes of desire thinking (i.e., imagery and verbal thoughts of seeking gratification and compensating negative feelings) fuel the strength of the craving experience, which may be mitigated by engaging in the specific behavior (e.g., internet using behavior). Over time, these associations between desire thinking, craving and engaging in the specific behavior may become stronger, and thus control over the behaviors by general inhibitory mechanisms may become more difficult and lead to addictive behavior (Brand et al., 2019). Thus, this cognitive process (i.e., desire thinking) plays a crucial role in the development and maintenance process of addiction, and it is, therefore, necessary to study it further.

To measure desire thinking, Caselli and Spada (2011) developed the Desire Thinking Questionnaire (DTQ), which is a self-report measurement and includes verbal perseveration (VP) and imaginal prefiguration (IP) as two main dimensions. VP refers to repeated conversations with oneself about the merits of engaging in appetitive target-related activities and achieving them. The IP refers to the allocation of attentional resources to the appetitive target-related information, as well as the multi-sensory elaboration of positive mental images or memories relating to the appetitive target. The two dimensions contain five items respectively. It was shown that the original DTQ had a two-factor structure, good internal consistency, and test–retest reliability, as well as good divergent and predictive validity (Caselli & Spada, 2011).

Recently, the DTQ has also been revised in various languages (e.g., Italian, French, German, Persian, Dutch, Turkish, Norwegian, Polish, and Hebrew; Aydın et al., 2022, Brandtner et al., 2021, Caselli and Spada, 2011, Chakroun-Baggioni et al., 2017, Dragan and Grajewski, 2021, Efrati et al., 2020, Khosravani et al., 2022, Markus et al., 2019, Solem et al., 2021) among various samples. The main contents of these studies are presented in Appendix A. Here we mainly summarized the following findings from the above studies. First, in various versions of DTQ, the two-factor structure is relatively stable and almost all items included in each dimension are consistent with the original DTQ (i.e., VP included items 4, 5, 6, 7, 10, and IP included items 1, 2, 3, 9) except item 8. Second, the total number of items in the various revisions of DTQ is different. The Norwegian version excluded item 8 which loaded on verbal perseveration (Solem et al., 2021), the Hebrew version excluded items 5, 6, 8, and 9 due to cross-loadings on two dimensions (Efrati et al., 2020), and the Turkish version excluded item 9 due to below 0.4 cutoff on both dimensions (Aydın et al., 2022). Third, the age stage of the sample group in the various revised DTQs is mainly adults. Except for the Hebrew and Turkish version of the DTQ which sampled adolescents (Efrati et al., 2020), other language versions of the DTQ mainly sampled adults. More importantly, the scores of adolescent desire thinking seem to be lower than that of adult desire thinking, and the scores in the two dimensions of adolescent are unstable. Specifically, in the study of Aydın et al. (2022), the average score of adolescents on VP is 7.86 (SD = 3.56), while that on IP is 4.73 (SD = 2.15). However, in the study of Efrati et al. (2020), the average score of adolescents on VP is 6.12 (SD = 2.63), while that on IP is 8.25 (SD = 2.72). However, in the study of non-clinical adults as the main sample, the average scores on VP and IP are basically more than 8 (Dragan and Grajewski, 2021, Solem et al., 2021). It is worth noting that this is only an intuitive and descriptive statistical result. Fourth, in general, the sample type of various revised DTQs is primarily non-clinical (e.g., game players and social media users), with only three studies using clinical samples. Specifically, the Italian version and the Dutch version explored the validity in a sample of alcohol abusers (Caselli and Spada, 2011, Markus et al., 2019). Additionally, the Persian version has been validated in three groups of individuals suffering from alcohol dependence, nicotine addiction, and problematic social media use (Khosravani et al., 2022). Last, unstable in scores of desire thinking was from various cultural backgrounds. Caselli and Spada, 2011, Markus et al., 2019, and Khosravani et al. (2022) all sampled alcohol abusers under various cultural backgrounds (i.e., Italian, Dutch and Persian). In the study of Markus et al. (2019), the alcohol use disorder outpatients and inpatient’s mean scores of desire thinking were 17.1 (SD = 5.9) and 17.9 (SD = 7.7), which was lower than the means scores of the other two studies intuitively (M = 25.03, SD = 5.83 and M = 24.8, SD = 6.8).

Different versions of the DTQ have been found to have a stable two-factor structure as well as good reliability and validity in various samples. There is, however, neither the Chinese (the most widely used language) version of the DTQ (DTQ-C) nor research has been conducted to assess the psychometric validity of the DTQ in adolescent mobile phone (the most popular and influential medium for accessing the internet) users.

Problematic mobile phone use (PMPU), also known as smartphone addiction, refers to problematic behaviors in individuals who are unable to self-regulate their mobile phone usage and ultimately cause negative consequences in their daily lives (Billieux, 2012). A survey by the Pew Research Center showed that 95 % of adolescents own a mobile phone (Pew Research Center, 2018). According to a survey conducted in China, 65 % of minors own a mobile phone (China Internet Network Information Center, 2020). There is an increase in mobile phone use and addictive trends, which contributes to PMPU and several physical (e.g., myopia), psychological (e.g., depression and anxiety), and interpersonal issues (e.g., interpersonal alienation; Busch & McCarthy, 2021). Compared to other age groups, the prevalence of PMPU among adolescents is relatively high (Horwood et al., 2021). Recent meta-analyses showed that the prevalence of PMPU among children and adolescents was 23.3 % globally (Sohn et al., 2019) and 23.9 % in China (Ma et al., 2021). Considering the rapid increase in adolescent mobile phone users, it appears necessary to develop tools that can both identify those at risk of PMPU and examine the assumptions that underlie theoretical models of internet addiction.

Several studies have shown that a high level of desire thinking was a significant characteristic of problematic internet users and was also a risky factor of internet addiction. Specifically, Spada et al. (2014) found that problematic internet users had higher levels of desire thinking than nonproblematic internet users. It has also been suggested that desire thinking may play an important role in problematic internet usage. Faghani et al. (2020) have found that desire thinking played a fully mediating role in the relationship between difficulties in emotion regulation and problematic internet use. For specific subtypes of problematic internet use, high-risk engagement gamers scored higher desire thinking than low-risk engagement and regular gamers (Dragan & Grajewski, 2021). In addition, desire thinking has also been found to be cross-sectionally related to the level of problematic social media use (Khosravani et al., 2022, Marino et al., 2019, Solem et al., 2021), problematic internet pornography use (Allen et al., 2017, Marino et al., 2023), internet gaming disorder (Aydın et al., 2022). A meta-analysis, which included ten studies to assess the relationship between desire thinking and addictive behaviors, showed that two factors of desire thinking were positively associated with addictive behaviors (including problematic Internet use) in both clinical and community samples (Mansueto et al., 2019). Thus, desire thinking could be considered as a crucial role in the initiation and persistence of problematic internet use. However, to date, PMPU as a subtype of problematic internet use has not been examined in its relationship with desire thinking.

This study validated the psychometric properties (e.g., the factor structure, reliability, and validity) of the DTQ-C in Chinese adolescent mobile phone users. Based on the existing empirical basis, we mainly included personality, cognitive, emotional, and addictive variables, which were frequently used in previous studies, to test the validity of the DTQ-C. Specifically, big five personality traits, negative affect, impulsivity, and PMPU have been frequently used to analyze concurrent and incremental validity in previous studies of psychometric validation of the DTQ (e.g., Efrati et al., 2020, Solem et al., 2021, Aydın et al., 2022, Khosravani et al., 2022). In addition, impulsivity is negatively related to self-control both in concepts and empirical results (Nigg, 2017). Therefore, we included self-control instead of impulsivity in concurrent and incremental validity analysis. Previous studies have shown that desire thinking has good discrimination validity from craving and rumination (Caselli & Spada, 2011), so we also examined the divergent validity under associations among the two subscales of the DTQ-C and brooding (matching the traditional definition of rumination as extended thinking; Castro-Calvo et al., 2022), craving.

We posed the following hypotheses: (1) DTQ-C would be the two-factor structure; (2) desire thinking would be positively correlated with neuroticism, brooding, depression, anxiety, and stress and was negatively correlated with self-control, extroversion, agreeableness, and conscientiousness; (3) craving and brooding would belong to different dimensions from desire thinking; (4) the VP and IP would be positively associated with the severity of PMPU, after controlling for demographic characteristics, big five personality traits, negative affect, and self-control.