In this study, a pre-test post-test control group design was used to investigate the impact of eight sessions of group art therapy on the stress coping ability of international students studying in Korea (Table 1). We used the PITR drawing test to measure the students’ coping abilities.

To recruit participants for the experiment, the purpose of the study and the criteria for participation were posted on an online chat room (WeChat). Among the applicants, 30 participants who met the recruitment criteria were selected. The inclusion criteria were as follows: (1) Chinese student studying in Korea and currently residing in Korea, (2) currently completing a university or graduate degree; (3) experiencing stress. Regarding the participants’ proficiency in Korean as assessed by the Test of Proficiency in Korean (TOPIK), in the experimental group (EG), 12 students (80.0%) scored at level 2 or below, while 3 students (20.0%) scored at level 3 or above. Moreover, 10 students (66.7%) below level 2 had no recorded TOPIK grade. In the control group (CG), 10 students (66.7%) scored at level 2 or below, and 5 students (33.3%) scored at level 3 or above. Additionally, 9 students (60.0%) below level 2 had no recorded TOPIK grade. It should be noted that all participants were required to achieve at least a level 4 proficiency to graduate from university. The demographic information of the participants is shown in Table 2.

The PITR test was devised by Arnold and Abraham (Hammer, 1958) and is used to determine whether the test taker can cope with stress. Specifically, the test taker uses pencil drawings to convey information about their level of stress and their ability to cope with it. The amount, direction, and intensity of rain depicted in the picture show the participants’ subjective perception of their amount of stress, and the objects that protect the figure, such as umbrellas, raincoats, and boots; the facial expressions; the position of the figure; and the size of the figure are interpreted as stress coping resources (Jue, 2019). In this study, we used the Korean version (Lee, 2008) of Lack’s (1996) PITR-SRC interpretation method to score the participants’ drawings.

The PITR questions consist of 35 items within three domains (see Table 3). The stress domain (S) focuses on stress levels and includes 16 items, S1–S16. The resource domain (R) focuses on stress resources and includes 19 items, R1–R19. The coping domain (C) focuses on the ability to cope with stress and includes 3 items. If the coping ability score is positive (+), then there are more resources than stress and thus relatively little stress. Meanwhile, if the coping ability score is positive (+), then a negative number (−) indicates more stress than resources.

The participants completed the PITR drawing test before and after the EG participated in a group art therapy program. In order to conduct the PITR test, the researcher provided the participants with A4 paper (210 × 297 mm), a 4B pencil, and an eraser and asked them to draw a person in the rain for 15 min. Next, the researcher advised the participants as follows: “It’s raining. Draw a person in the rain. At this time, please draw the person in full form, not in the shape of a cartoon or a stick.” The researcher allowed the participants to draw freely; that is, the researcher did not advise participants on things such as the size and shape of the drawing or the position of the figure. When participants asked the research questions about how they should draw while completing the test, the researcher responded, “You can draw whatever you want.” Finally, we collected qualitative data in written form by asking participants who they drew, what the person in their drawing was doing, and how the person in their drawing felt. This was scored as supplementary data for the PITR picture analysis index. In particular, if it is not a picture of only one person, when scoring R11-R15, it was scored by referring to the answer of “Who is the main character in the picture?” (R11-Whole face; R12-Smile; R13-Centred figure; R14-Size of figure; R15-Full figure).

To ensure the reliability of the PITR scoring procedure and results, three graders, including researchers, scored the tests. Before scoring, 60 original copies and copies of the pre- and post-test PITR results from the EG and CG, scoring guidelines, and scoring tables were given to the graders. To minimize grader bias, 60 copies of the PITR test results were randomly numbered so that participants could not be identified. A Chinese interpreter was included to facilitate communication between graders while they scored the tests.

Before scoring, the researcher conducted a test to ensure the graders understood the analysis criteria for each item in the test. After a thorough discussion of the scoring system and criteria, each rater independently scored all 60 tests to ensure reliability. In addition, to ensure consistency, the graders evaluated the items when the grading results were checked against each other. A Pearson correlation analysis was performed to analyze the raters’ comparative confidence levels (see Table 4).

Raters A and B had a high “stress” confidence correlation coefficient of 0.782, a “stress resources” confidence correlation coefficient of 0.843, and a “coping ability” confidence correlation coefficient of 0.810. Meanwhile, Raters B and C had a “stress” confidence correlation coefficient of 0.936, a “stress resource” confidence correlation coefficient of 0.982, and a “coping ability” confidence correlation coefficient of 0.952. Lastly, Raters A and C had a “stress” confidence correlation coefficient of 0.814, “stress resources” confidence correlation coefficient of 0.823, and a “coping ability” confidence correlation coefficient of 0.811. These results suggest that the three graders were fully familiar with the scoring criteria and independently graded according to these clear criteria (see Figures 1–4).

The 30 Chinese international student participants living in Korea were randomly divided into the EG (n = 15) and CG (n = 15). As established above, the EG completed an eight-session group art therapy program facilitated by an experienced Korean art therapist and the CG did not complete this program. Each session of the therapy program was 120 min long (this length was based on interpretation time and group size) and comprised three phases: the introductory phase (20 min), the task and activity phase (60 min), and the organizing and verbal sharing phase (40 min). The EG’s program consisted of themes related to stress, such as emotional suppression, barriers, cultural adaptation stress, academic stress, self-acceptance, positive images, and future vision after reviewing literature and consulting with a therapist experienced with diverse cultures. Table 5 shows intervention.

The data were analyzed as follows: SPSS 26.0 was used for statistical processing. A frequency analysis was performed to assess the demographic characteristics of the participants. Mann–Whitney U tests were performed to test the homogeneity of the EG and CG before the program was implemented. A comparative analysis was performed based on the PITR-SRC (Lack, 1996) scoring method developed by Lee (2008) and reliability tests were performed to determine if the raters’ scoring results matched. A Wilcoxon signed-rank test was performed to check for differences in scores between the groups. Finally, participants’ nonverbal expressions, verbal expressions, and changes were analyzed in the context of the PITR.

This study is part of the author’s doctoral dissertation (Yin, 2023) and its description has been modified to align with the journal format. This study was approved by the Institutional Review Board (IRB) of the Jeonju university (JJIRB-220421-HR-2022-0407). The researcher advised the participants on the purpose of the study, their right to confidentiality, and their right to stop participating in the study without any penalty. All data were gathered after IRB approval and participant consent. Since a Korean art therapist facilitated the program, a Chinese interpreter was present during the sessions to facilitate accurate communication. All data were stored on the researcher’s personal computer with a password to protect privacy.

The Mann–Whitney U test was conducted to determine the homogeneity of the EG and CG. As a result of the homogeneity test of the means of the EG and CG, the average score for the “stress” domain in the PITR test among the EG was 14.67 points; while this is higher than the CG’s score of 11.87 points, this difference was not statistically significant (p > 0.05). Meanwhile, the mean score for “stress resources” was higher in the EG (2.13) than in the CG (0.33); however, these results were also not statistically significant (p > 0.05). Regarding “coping ability,” the EG demonstrated a lower mean score (−12.53) than the CG (−11.53); again, these results were not statistically significant (p > 0.05). Therefore, the homogeneity of the PITR test between the two groups was confirmed (see Table 6).

In the post-test conducted after the group art therapy, a Mann–Whitney U analysis was performed to identify the differences in the post-test results of the EG and CG. The results are shown in Table 7. Specifically, the comparison indicated statistically significant differences across all areas. In the domain of “stress,” the EG’s mean score (5.27) was lower than the CG’s mean score (15.73) (p < 0.001). Regarding “stress resources,” the EG’s mean score (6.93) was higher than the CG’s main score (−0.33) (p < 0.001). In terms of “coping ability,” the EG’s mean score (1.67) was statistically significantly higher than the CG’s mean score (−16.07) (p < 0.001). Therefore, the pre-post test results reveal that the group art therapy program significantly improved the EG’s stress-related capacities compared to the CG.

To analyze the effect of the group art therapy program on Chinese international students in Korea, we compared the pre-post and post-test PITR results of the two groups. A Wilcoxon signed-rank test was conducted to determine whether the pre- and post-test changes were statistically significant (see Table 8).

The pictures in the EG for the post-test usually appeared to have no rain or a dot shape, and the protagonist in the picture had no contact with the rain. There was not enough wind, puddles, or lightning. Most of the pictures were cloudless or appeared to have friendly clouds. There were also many pictures in which there was appropriate protection, and one or more of the protections were depicted. The impressions of the picture appeared to be positive, with low stress scores and high stress coping resource scores. In the case of the CG, most of the pictures showed a lot of rain, or rain was drawn in various shapes such as dots, lines, circles, and drops. In almost all the paintings, the protagonist was shown to be in direct contact with the rain. Wind, puddles, lightning, and dark clouds were also depicted a lot, and the stress was high. In addition, there was no protection or it appeared inappropriately; the protagonist had also omitted body parts or only the back of the character was shown. The protagonist in the picture looked tired and stressed, and depicted exuded a negative vibe. The stress coping resources were low (see Table 9).

The comparison revealed that the score for “stress” significantly decreased in the EG, as shown in Figure 5. Specifically, the EG’s pre-test mean score of 14.67 decreased to 5.27 in the post-test; this difference is statistically significantly different (p < 0.01). Meanwhile, the CG’s pre-test mean score of 11.87 increased to 15.73 in the post-test; this difference is also statistically significantly different (p < 0.05) (see Figure 5).

The stress coping ability of the participants who participated in the group art therapy session was improved by comparing the pre-post of the PITR picture in the actual box group. It seems like the stress was reduced after assessing the items, which comprised a stress measure. Looking closely, the “amount of rain” significantly decreased for all 15 participants from 1 to 15, and the “contact of rain” in the pictures of participants 1, 3, 4, 6, 12, 13, and 14 showed that the “wind” item in the pictures of participants 12 and 14, “hole” item in the pictures of participants 2, 3, 4, 6, 7, 8, 9 and “lighting” item in the pictures of participants 1, 5, 7, 8, 9 disappeared. The “cloud” item in the pictures of participants 3, 4, 5, 6, 8, and 9 showed that the cloud disappeared or the dark cloud became a cloud. It seems that the ability to cope with stress increased throughout the EG through the items, which are measures of coping ability. Looking closely, the pictures of participants 1, 2, 3, 6, 7, 11, and 12 included “protective objects,” and the pictures of participants 3, 12, and 14 showed “protective objects” appropriately so that people in the drawing do not get rained on. The “character” in the pictures of participants 2, 3, 4, 5, 6, 11, and 14 were drawn from the back or side; however, the entire figure could be seen more completely as it was from the front side of human body. Moreover, the main character of the paintings of participants 1, 2, 3, 5, 6, 8, 10, 11, 12, and 15 had a smiling expression. As a result, the main character in the picture turned negative emotions into positive ones (see Figures 6, 7).

The “stress resources” of the CG did not change—the group’s pre-test and post-test mean scores for this domain were both 0.33 to a post-test mean score of −0.33; this result accordingly did not reveal a statistically significant difference (p > 0.05) (see Figure 8).

By comparing the pre-post of the PITR picture in the control group, it was found that the stress and coping ability scales did not change significantly. The stress scale items, namely the “amount of rain,” “contact of rain,” “wind,” “punch,” “lightning,” and “cloud” did not decrease or disappear. The coping ability scale showed that no protection appeared or it was used inappropriately. Furthermore, most of the protagonists in the picture were not complete or appeared in the back. Overall, this showed that the pre-post of the PITR picture in the CG did not change much. Finally, “coping ability” increased in the EG—the EG’s pre-test mean score for this domain was −12.53 and its post-test mean score was 1.67, as shown in Figure 7; this difference was statistically significant (p < 0.01). In the CG, the coping ability mean score increased from −11.53 in the pre-test to–16.07 in the post test (p > 0.05), which was not significant. Ultimately, these results show that the group art therapy program decreased stress and increased stress resources and coping abilities among Chinese students in Korea (see Figure 9).