*3.1. Preliminary Analyses and Descriptive Statistics—PPAT Drawings* Table 2 about here.



Table 3 about here.

**Table 3.** Means, Standard Deviations of the SC-PPAT/c: Symbolic Contents in children's PPAT [51] (*N* = 156).



#### **Table 3.** *Cont.*

Based on Tables 2 and 3, the average drawing in this sample (see for example Figure 1) demonstrates the use of realistic color, with some of the drawn forms colored in. Children invested moderate energy in their drawings and used approximately 50% of the space. Most of the drawing reflect a development stage approaching the latency period; the items in the drawings were recognizable but simply drawn. The drawings contained only one unrecognizable item. The drawn person is holding an apple, but it is not clear how he/she obtained it. In addition to the person, tree, and apple, a horizon line/ground is perceptible, line quality is good (a controlled line), and there is integration of two to three elements. The drawn person is identifiable. *Children* **2022**, *9*, x FOR PEER REVIEW 9 of 25

**Figure 1.** An average PPAT drawing, in regard to form and content layers. **Figure 1.** An average PPAT drawing, in regard to form and content layers.

Observation of the content layer of the PPAT drawings reveals a fairly weak tree with 3 or 4 apples on it. The drawn person seems passive or is avoiding the picking process, and though the apple and person are close, there is no contact between them. The person is shorter than the tree by a ratio of 1:3. The tree truck is slightly inclined away from the person. Tree's branches or treetop are neutral in relation to the person; however, they are high and inaccessible to the picker. The apples are placed on the side farther away from the person or distributed equally. Confirmatory Factor Analysis (CFA) was conducted using AMOS software version Observation of the content layer of the PPAT drawings reveals a fairly weak tree with 3 or 4 apples on it. The drawn person seems passive or is avoiding the picking process, and though the apple and person are close, there is no contact between them. The person is shorter than the tree by a ratio of 1:3. The tree truck is slightly inclined away from the person. Tree's branches or treetop are neutral in relation to the person; however, they are high and inaccessible to the picker. The apples are placed on the side farther away from the person or distributed equally.

23 for the FEATS and for the SC-PPAT scoring systems that compares the theoretical

**Measure Factor Estimat**

Integration <--- As-Real-R 0.699 \*\*\* Realism <--- As-Real-R 0.714 \*\*\* Problem Solving <--- As-Real-R 0.306 \*\*\* Developmental Level <--- As-Real-R 0.878 \*\*\* Color Fit <--- As-Real-R 0.553 \*\*\* Person <--- As-Real-R 0.452 \*\*\* Line Quality <--- As-Real-R 0.369 \*\*\*

Prominence of Color <--- Vividness 0.419 \*\*\* Implied Energy <--- Vividness 0.719 \*\*\* Space <--- Vividness 0.666 \*\*\*

Details of Objects & Environment <--- Vividness 0.906 \*\*\*

(36) = 84.10, *p* < 0.0001, CFI = 0.936, RMSEA = 0.093, SRMR = 0.067. Correlations between the two

**e** 

**Table 4.** Confirmatory Factor Analysis of FEATS scales.

Table 4 about here.

latent variables r = 0.734. \*\*\* *p* < 0.001.

Table 5 about here.

χ 2

model and the empirical model (leaning on previous data, [39]).

Confirmatory Factor Analysis (CFA) was conducted using AMOS software version 23 for the FEATS and for the SC-PPAT scoring systems that compares the theoretical model and the empirical model (leaning on previous data, [39]). Table 4 about here.

**Table 4.** Confirmatory Factor Analysis of FEATS scales.


χ 2 (36) = 84.10, *p* < 0.0001, CFI = 0.936, RMSEA = 0.093, SRMR = 0.067. Correlations between the two latent variables r = 0.734. \*\*\* *p* < 0.001.

Table 5 about here. **Table 5.** Confirmatory Factor Analysis of SC-PPAT/c2 scales.

**Table 5.** Confirmatory Factor Analysis of SC-PPAT/c2 scales. **Measure Factor Estimate** 


χ 2 (13) = 22.85, *p* = 0.043, CFI = 0.98, RMSEA = 0.070, SRMR = 0.046. Correlations between the two latent variables r = 0.746. \*\*\* *p* < 0.001. χ2 (13) = 22.85, *p* = 0.043, CFI = 0.98, RMSEA = 0.070, SRMR = 0.046. Correlations between the two latent variables r = 0.746. \*\*\* *p* < 0.001.

Regarding the FEATS scores, the CFA yielded acceptable goodness of fit indices. The first factor, Vividness, was identical in scale composition to this factor in previous sample [39], incorporating the following scales: Prominence of Color; Implied Energy; Space; and Details of Objects and Environment (Cronbach's Alpha was 0.80). Vividness accounted for 40% of the common variance. Figures 2 and 3 illustrate high and low levels of this factor respectively. Regarding the FEATS scores, the CFA yielded acceptable goodness of fit indices. The first factor, Vividness, was identical in scale composition to this factor in previous sample [39], incorporating the following scales: Prominence of Color; Implied Energy; Space; and Details of Objects and Environment (Cronbach's Alpha was 0.80). Vividness accounted for 40% of the common variance. Figures 2 and 3 illustrate high and low levels of this factor respectively.

The second factor was found to incorporate more scales than the previous model; specifically, it included: Color Fit, Integration, Realism, Problem Solving, Developmental Level, Line Quality, and Person, with a good reliability (Cronbach's Alpha was 0.73). This factor was named As-Real-R, because most of its scales serve the purpose of drawing recognizable images through the use of realistic colors, the relationships between the drawn objects, and a controlled line quality. R stands for Revised and different this factor from the original 'As-Real' factor in the study of Roth et al. [39]. Figures 4 and 5 demonstrate

**Figure 2.** PPAT drawing with a high Vividness factor. **Figure 2.** PPAT drawing with a high Vividness factor.

**Figure 3.** PPAT drawing with a low Vividness factor.

high and low levels of this factor respectively.

latent variables r = 0.746. \*\*\* *p* < 0.001.

factor respectively.

**Figure 3.** PPAT drawing with a low Vividness factor. **Figure 3.** PPAT drawing with a low Vividness factor.

**Figure 2.** PPAT drawing with a high Vividness factor.

*Children* **2022**, *9*, x FOR PEER REVIEW 10 of 25

**Measure Factor Estimate** 

picking <--- Person's Agency 0.985 \*\*\*

Degree of success in picking the apple <--- Person's Agency 0.862 \*\*\* Quantity of apples on the tree <--- Tree's Generosity 0.543 \*\*\* Strength vs. weakness of tree <--- Tree's Generosity 0.541 \*\*\*

or far from the person <--- Tree's Generosity 0.836 \*\*\*

(13) = 22.85, *p* = 0.043, CFI = 0.98, RMSEA = 0.070, SRMR = 0.046. Correlations between the two

Regarding the FEATS scores, the CFA yielded acceptable goodness of fit indices. The first factor, Vividness, was identical in scale composition to this factor in previous sample [39], incorporating the following scales: Prominence of Color; Implied Energy; Space; and Details of Objects and Environment (Cronbach's Alpha was 0.80). Vividness accounted for 40% of the common variance. Figures 2 and 3 illustrate high and low levels of this

Position of the tree trunk in relation to the person <--- Tree's Generosity 0.893 \*\*\* Branch/s placement in relation to the person <--- Tree's Generosity 0.890 \*\*\*

**Table 5.** Confirmatory Factor Analysis of SC-PPAT/c2 scales.

The degree to which the person is active/passive in apple-

The extent to which apples are scattered on the tree either close

χ2

The second factor was found to incorporate more scales than the previous model; specifically, it included: Color Fit, Integration, Realism, Problem Solving, Developmental Level, Line Quality, and Person, with a good reliability (Cronbach's Alpha was 0.73). This factor was named As-Real-R, because most of its scales serve the purpose of drawing recognizable images through the use of realistic colors, the relationships between the drawn objects, and a controlled line quality. R stands for Revised and different this factor from the original 'As-Real' factor in the study of Roth et al. [39]. Figures 4 and 5 demonstrate high and low levels of this factor respectively. The second factor was found to incorporate more scales than the previous model; specifically, it included: Color Fit, Integration, Realism, Problem Solving, Developmental Level, Line Quality, and Person, with a good reliability (Cronbach's Alpha was 0.73). This factor was named As-Real-R, because most of its scales serve the purpose of drawing recognizable images through the use of realistic colors, the relationships between the drawn objects, and a controlled line quality. R stands for Revised and different this factor from the original 'As-Real' factor in the study of Roth et al. [39]. Figures 4 and 5 demonstrate high and low levels of this factor respectively. *Children* **2022**, *9*, x FOR PEER REVIEW 11 of 25 *Children* **2022**, *9*, x FOR PEER REVIEW 11 of 25

**Figure 4.** PPAT drawing with a high As-Real-R factor. **Figure 4.** PPAT drawing with a high As-Real-R factor. **Figure 4.** PPAT drawing with a high As-Real-R factor.

**Figure 5.** PPAT drawing with a low As-Real-R factor. **Figure 5.** PPAT drawing with a low As-Real-R factor. **Figure 5.** PPAT drawing with a low As-Real-R factor.

Regarding the CFA of the SC-PPAT scores, we found no difference between the two models. As shown in Table 5, two main factors were obtained: 'Person's Agency' describes the degree a person is active/passive in the apple picking process, and the degree of applepicking success. Figures 6 and 7 demonstrate high and low levels of this factor, respectively. Regarding the CFA of the SC-PPAT scores, we found no difference between the two models. As shown in Table 5, two main factors were obtained: 'Person's Agency' describes the degree a person is active/passive in the apple picking process, and the degree of applepicking success. Figures 6 and 7 demonstrate high and low levels of this factor, respectively. Regarding the CFA of the SC-PPAT scores, we found no difference between the two models. As shown in Table 5, two main factors were obtained: 'Person's Agency' describes the degree a person is active/passive in the apple picking process, and the degree of applepicking success. Figures 6 and 7 demonstrate high and low levels of this factor, respectively.

'Tree Generosity' pertains to the tree's strength, its number of apples, and the tree's orientation is relation to the drawn person, including the trunk's inclination, and

'Tree Generosity' pertains to the tree's strength, its number of apples, and the tree's orientation is relation to the drawn person, including the trunk's inclination, and

**Figure 6.** PPAT drawing with a high Person Agency factor.

**Figure 6.** PPAT drawing with a high Person Agency factor.

**Figure 7.** PPAT drawing with a low Person Agency factor.

**Figure 7.** PPAT drawing with a low Person Agency factor.

tively.

tively.

**Figure 6.** PPAT drawing with a high Person Agency factor. **Figure 6.** PPAT drawing with a high Person Agency factor. **Figure 6.** PPAT drawing with a high Person Agency factor.

*Children* **2022**, *9*, x FOR PEER REVIEW 11 of 25

*Children* **2022**, *9*, x FOR PEER REVIEW 11 of 25

**Figure 4.** PPAT drawing with a high As-Real-R factor.

**Figure 4.** PPAT drawing with a high As-Real-R factor.

**Figure 5.** PPAT drawing with a low As-Real-R factor.

**Figure 5.** PPAT drawing with a low As-Real-R factor.

'Tree Generosity' pertains to the tree's strength, its number of apples, and the tree's orientation is relation to the drawn person, including the trunk's inclination, and 'Tree Generosity' pertains to the tree's strength, its number of apples, and the tree's orientation is relation to the drawn person, including the trunk's inclination, and 'Tree Generosity' pertains to the tree's strength, its number of apples, and the tree's orientation is relation to the drawn person, including the trunk's inclination, and placement of branches. Figures 8 and 9 demonstrate high and low levels of this factor, respectively. placement of branches. Figures 8 and 9 demonstrate high and low levels of this factor, respectively. These factors yield a total of 84% of the explained variance. placement of branches. Figures 8 and 9 demonstrate high and low levels of this factor, respectively. These factors yield a total of 84% of the explained variance.

Regarding the CFA of the SC-PPAT scores, we found no difference between the two models. As shown in Table 5, two main factors were obtained: 'Person's Agency' describes the degree a person is active/passive in the apple picking process, and the degree of applepicking success. Figures 6 and 7 demonstrate high and low levels of this factor, respec-

Regarding the CFA of the SC-PPAT scores, we found no difference between the two models. As shown in Table 5, two main factors were obtained: 'Person's Agency' describes the degree a person is active/passive in the apple picking process, and the degree of applepicking success. Figures 6 and 7 demonstrate high and low levels of this factor, respec-

**Figure 8.** PPAT drawing with a high Tree Generosity factor. **Figure 8.** PPAT drawing with a high Tree Generosity factor. **Figure 8.** PPAT drawing with a high Tree Generosity factor.

**Figure 9.** PPAT drawing with a low Tree Generosity factor. **Figure 9.** PPAT drawing with a low Tree Generosity factor. **Figure 9.** PPAT drawing with a low Tree Generosity factor.

Table 6 shows frequency score distribution of the PPAT drawing's narrative focus. Table 6 shows frequency score distribution of the PPAT drawing's narrative focus. These factors yield a total of 84% of the explained variance.

As can be seen the PPAT narrative was observed to be the main narrative in only 43.6% of the drawings. In about 28% of the PPAT drawings, the PPAT script was absent due to lack of at least one of the objects (a tree, a person, and/or an apple) (see for example Figure 10); in about 33% of the drawings, the children depicted narratives that rival the picking script and that became the central narrative (see for example Figure 11). The Mean score As can be seen the PPAT narrative was observed to be the main narrative in only 43.6% of the drawings. In about 28% of the PPAT drawings, the PPAT script was absent due to lack of at least one of the objects (a tree, a person, and/or an apple) (see for example Figure 10); in about 33% of the drawings, the children depicted narratives that rival the picking script and that became the central narrative (see for example Figure 11). The Mean score Table 6 shows frequency score distribution of the PPAT drawing's narrative focus. As can be seen the PPAT narrative was observed to be the main narrative in only 43.6% of the drawings. In about 28% of the PPAT drawings, the PPAT script was absent due to lack of at least one of the objects (a tree, a person, and/or an apple) (see for example Figure 10); in

in this scale was 2.48 (SD = 2.03), i.e., on average the drawing showed a PPAT script that is limited to the picking script, with some degree of emphasis on one element of the script.

in this scale was 2.48 (SD = 2.03), i.e., on average the drawing showed a PPAT script that is limited to the picking script, with some degree of emphasis on one element of the script.

percentages from a sample of the same age from a non-war zone [39] *N* = 125 in parenthesis. Adap-

percentages from a sample of the same age from a non-war zone [39] *N* = 125 in parenthesis. Adap-

<sup>37</sup>23.7

<sup>37</sup>23.7

<sup>30</sup>19.2

<sup>30</sup>19.2

<sup>20</sup>12.8

<sup>20</sup>12.8

(12) 23.7

(12) 23.7

(20) 42.9

(20) 42.9

(28) 55.8

(28) 55.8

(4) 61.5

(4) 61.5

**Score Category Description Frequency Valid Percent Cumulative Percent**

**Score Category Description Frequency Valid Percent Cumulative Percent**

process sun, plants, flowers, etc.) 9 5.8

process sun, plants, flowers, etc.) 9 5.8

ted with permission from [39] 2020 Elsevier Ltd.

ted with permission from [39] 2020 Elsevier Ltd.

Partial or complete absence of picking aspects.

Partial or complete absence of picking aspects.

The drawing focus solely on the PPAT story, with equal investment in all parts of the drawing

The drawing focus solely on the PPAT story, with equal investment in all parts of the drawing

The drawer emphasizes certain details about the person/tree/ picking/apples/ accessories/ environment (in this example the trunk).

The drawer emphasizes certain details about the person/tree/ picking/apples/ accessories/ environment (in this example the trunk).

The environment supports the picking

The environment supports the picking

<sup>0</sup>Absence of PPAT narrative message.

<sup>0</sup>Absence of PPAT narrative message.

<sup>2</sup>Emphasized

<sup>2</sup>Emphasized

Schematic and balanced PPAT narrative message.

Schematic and balanced PPAT narrative message.

narrative message.

narrative message.

Narrative message supported by the environment.

Narrative message supported by the environment.

1

1

3

3

narrative message.

about 33% of the drawings, the children depicted narratives that rival the picking script and that became the central narrative (see for example Figure 11). The Mean score in this scale was 2.48 (SD = 2.03), i.e., on average the drawing showed a PPAT script that is limited to the picking script, with some degree of emphasis on one element of the script.

**Table 6.** Frequencies of Degree of Narrative Focus [52] *N* = 156 in comparison to the prevalence in percentages from a sample of the same age from a non-war zone [39] *N* = 125 in parenthesis. Adapted with permission from [39] 2020 Elsevier Ltd.

(21) 100.0


approaching the picker).

attention of the apple picker (e.g., a snake approaching the picker).

**Figure 10.** PPAT drawing with a missing apple-picking narrative. **Figure 10.** PPAT drawing with a missing apple-picking narrative. **Figure 10.** PPAT drawing with a missing apple-picking narrative.

*3.2. Preliminary Analyses and Descriptive Statistics—Independent Variables*  **Figure 11.** PPAT with a rival narrative. **Figure 11.** PPAT with a rival narrative.

partly confirmed the first hypothesis.

partly confirmed the first hypothesis.

*3.2. Preliminary Analyses and Descriptive Statistics—Independent Variables* 

Pearson correlations were calculated between the child's variables with the goal of

symptoms, and hyperactivity), the BREIF total score summing the five scales, and the CAMP score of child's self-potency. The results in Table 7 show that a significant association was found between the emotional problems of the child reported by the mother and EF problems reported by the child's teacher (r = 0.438, *p* < 0.0001). Additionally, a medium significant association was found between the degree of maternal stress and the child's emotional symptoms reported by the mother (r = 0.273, *p* < 0.004). Non-significant associations were found between the child's self-potency, maternal stress, and child's emotional symptoms. We also calculated these correlations in each gender group and found two additional associations only among girls: a positive association between maternal stress degree and girls' EF problems (r = 0.351, *p* < 0.011), n = 52; and a negative correlation between girls' EF problems and self-potency (r = −0.27, *p* < 0.036), n = 60. These associations

nal stress, the SDQ sum of two scales that were found with high reliability (emotional symptoms, and hyperactivity), the BREIF total score summing the five scales, and the CAMP score of child's self-potency. The results in Table 7 show that a significant association was found between the emotional problems of the child reported by the mother and EF problems reported by the child's teacher (r = 0.438, *p* < 0.0001). Additionally, a medium significant association was found between the degree of maternal stress and the child's emotional symptoms reported by the mother (r = 0.273, *p* < 0.004). Non-significant associations were found between the child's self-potency, maternal stress, and child's emotional symptoms. We also calculated these correlations in each gender group and found two additional associations only among girls: a positive association between maternal stress degree and girls' EF problems (r = 0.351, *p* < 0.011), n = 52; and a negative correlation between girls' EF problems and self-potency (r = −0.27, *p* < 0.036), n = 60. These associations

#### *3.2. Preliminary Analyses and Descriptive Statistics—Independent Variables*

Pearson correlations were calculated between the child's variables with the goal of analyzing theoretical validity (Table 7), in specific, between the BSI global score of maternal stress, the SDQ sum of two scales that were found with high reliability (emotional symptoms, and hyperactivity), the BREIF total score summing the five scales, and the CAMP score of child's self-potency. The results in Table 7 show that a significant association was found between the emotional problems of the child reported by the mother and EF problems reported by the child's teacher (r = 0.438, *p* < 0.0001). Additionally, a medium significant association was found between the degree of maternal stress and the child's emotional symptoms reported by the mother (r = 0.273, *p* < 0.004). Non-significant associations were found between the child's self-potency, maternal stress, and child's emotional symptoms. We also calculated these correlations in each gender group and found two additional associations only among girls: a positive association between maternal stress degree and girls' EF problems (r = 0.351, *p* < 0.011), n = 52; and a negative correlation between girls' EF problems and self-potency (r = −0.27, *p* < 0.036), n = 60. These associations partly confirmed the first hypothesis.

**Table 7.** Associations between global scores of child's problems in EFs, emotional symptoms, selfpotency, and maternal stress, for the whole sample and for gender groups.


\*\* Correlation is significant at the 0.01 level (2-tailed).

Spearman correlations were calculated between maternal assessments of the amount of political violence events that children witnessed and maternal stress and the child's emotional and cognitive problems. As Table 8 shows, the number of events of political violence to which the child was exposed according to the mother's assessment was positively correlated with the emotional problems she reported about her child (r = 0.362, *p* < 0.0001) and with her level of maternal stress (r = 0.309, *p* < 0.002). The last result confirms the second hypothesis, and indicates that the more the mother assessed that her child was exposed to events of political violence, the more she experienced maternal stress.


**Table 8.** Associations between the mother's assessment of political violence events, with global scores of maternal stress and child's variables.

\*\* Correlation is significant at the 0.01 level (2-tailed).

#### *3.3. Hypotheses Analyses Regarding the PPAT Drawings and Independent Variables*

To examine hypotheses no. 3–5, we computed Pearson correlations between the child's variables and the main factors of the drawings. Table 9 presents these associations. Table 9 about here.

**Table 9.** Associations between the PPAT drawings' main factors, and global scores of maternal stress and child's variables, for the whole sample and for gender groups.



**Table 9.** *Cont.*

\*\* Correlation is significant at the 0.01 level (2-tailed). \* Correlation is significant at the 0.05 level (2-tailed).

Hypothesis no. 3: As can be seen from Table 9, positive associations were found between the child's self-report of self-potency, and three of the main aspects of PPAT drawings. In specific, the more the child saw herself/himself as self-potent, the more the PPAT drawing depicted a more generous tree (r = 0.38, *p* < 0.0001), a more active drawn person (r = 0.30, *p* < 0.0001), and the drawing was drawn in a more realistic manner (r = 0.36, *p* < 0.0001). These results confirmed the third hypothesis. Two of the associations were significantly stronger among girls than boys, according to Fisher Z of the difference between the correlation coefficients: specifically, the more a girl reported higher self-potency, the more her drawing was higher in As-Real-R factor (r = 0.56, *p* < 0.0001) z = 2.67 (*p* < 0.007), and higher in Vividness (r = 0.24, *p* < 0.033) z = 1.94 (*p* < 0.05).

Hypothesis no. 4: Regarding children's EF problems, negative associations were found between the child's EF difficulties and the four main aspects of the PPAT drawing for the whole sample (associations range between −0.19 to −0.45). However, after a division according to gender, one significantly stronger association was found among boys (n = 59). Specifically, the more the teacher reported that a boy had EF difficulties, the more the boy's PPAT drawing was lower in As-Real-R factor (r = −0.56, *p* < 0.0001) z = 2.44 (*p* < 0.01).

In relation to the children's emotional problems, small negative associations were found between the total score of emotional problems and two PPAT aspects (tree generosity and the drawing's As-Real-R factor) for the whole sample. The correlations in each gender group were not significantly different according to Fisher Z. These associations (between EF and emotional problems and the PPAT drawings' main factors) partly confirmed the fourth hypothesis.

Hypothesis no. 5: As Table 9 shows, we found non-significant associations between maternal stress and the child's PPAT drawings, for the whole sample and for gender groups. For further examination, we also computed Pearson correlations for each sub-scale of the maternal stress measure, and we found a single positive association between the degree of maternal depression and the generosity of the boy's tree (r = 0.28 \*, *p* < 0.035). This means that the more the mother was depressed, the more her son drew a generous tree. This result showed that the fifth hypothesis was not confirmed.

Hypothesis no. 6: To examine the possibility of an association between the assessments of mothers regarding the degree of exposure of their children to events of political violence and PPAT drawings, we calculated Spearman correlations, and found that only among girls (n = 50), the higher the number of events of political violence that the daughter was exposed to according to her mother's assessment, the more the daughter tended to draw an active drawn person in her PPAT drawing (r = 0.49 \*\*, *p* < 0.0001), a generous tree (r = 0.28 \*, *p* < 0.048), and a realistic drawing (r = 0.33 \*, *p* < 0.016). The sixth hypothesis was disapproved.

Hierarchical linear regressions were carried out to examine the impact of children's gender, self-potency, and EF difficulties on the main factors of the children's PPAT drawings. Gender was entered in the first block; in the second, self-potency and EF difficulties were entered using the stepwise method. Two statistically significant models were found as explaining the drawing's As-Real-R factor, and the Tree Generosity (see Tables 10 and 11). The results of the hierarchical linear regression analysis revealed that in regard to the Tree Generosity factor, gender was not a statistically significant predictor (*p* > 0.05); however, in regard to the As-Real-R factor, gender was found to be a statistically significant predictor.



Gender 1 = Female. <sup>c</sup> CAMP Questionnaire; <sup>b</sup> BRIEF Questionnaire. \*\* *p* < 0.01; \*\*\* *p* < 0.001.

**Table 11.** Multiple Linear Regression Stepwise (Dependent Variable: Tree Generosity in PPAT).


Gender 1 = Female. <sup>d</sup> CAMP Questionnaire; <sup>c</sup> BRIEF Questionnaire. \* *p* < 0.05; \*\*\* *p* < 0.001.

Tables 10 and 11 show the influence of gender, self-potency and EF difficulties on the degree of the drawing's As-Real-R factor (F(3,115) = 18.55, *p* < 0.001) and the influence of self-potency and EF difficulties on the degree of Tree Generosity (F(3,115) = 19.27, *p* < 0.001).

Specifically, the R2 indicated that 33% of the variance in the drawing's As-Real-R degrees can be explained by variances in the three predictor variables: gender, self-potency and EF problems. For Tree Generosity, the R2 indicated that 34% of the variance in the Tree Generosity degree is explained by the child's self-potency and EF difficulties.

Hypothesis no. 7. As can be seen in Table 6, in drawings of children from the same age group from a non-war zone (n = 125) [39], the main narrative in the majority of the drawings (66%) was that of a person picking an apple from a tree (the percentage-sum of categories number 1 to 4); however, the current study revealed different results. Specifically, only 43.6% of the drawings depicted a central narrative of a person picking an apple. To calculate the difference of comparison of the proportions (of the two categories of missing picking narrative and rival narrative), a significant Chi-Squared of 14,084 was found *p* < 0.002.

The Focus of the Narrative scale was recoded into three main categories: (a) drawings lacking a picking scene (category 0; n = 35); (b) drawings depicting a picking scene as a the main narrative (categories 1–4; n = 68); and (c) drawings that depict a rival narrative as the central narrative (category no. 5; n = 51). We first conducted Spearman correlations between the three narrative categories and the child's variables, including maternal stress, and then conducted further One-Way ANOVA tests (see Table 12), which revealed significant differences between the three narrative focus groups in terms of their self-potency scores F (2,115) = 12.05, *p* < 0.0001, Partial Eta Squared was 0.163. (Scores of self-potency for each narrative groups: Group 1. Missing picking narrative-Mean: 13.67, SD = 3.5; Group 2. A focused picking narrative-Mean: 16.60, SD = 2.24; Group 3. A rival narrative-Mean: 16.41, SD = 2.74). A post hoc Bonferroni test showed significant differences between Group 1 and the other two groups, i.e., children that drew a PPAT drawing without a picking narrative had the lower self-potency scores in comparison to children who drew a picking scene as the main narrative or a picking scene with a rival narrative.

Table 12 about here.

**Table 12.** Univariant ANOVA for self-potency differences withing the three narrative categories.


<sup>a</sup> R Squared = 0.163 (Adjusted R Squared = 0.152).

#### **4. Discussion**

The present study sought to investigate the experiences of preschool children living in areas of prolonged political violence through examination of their PPAT drawings in relation to various child and maternal variables.

Regarding the children's mothers' perspective, significant associations were found between mothers' reports on the degree to which their children were exposed to events of political violence and the degree of maternal stress. According to caregiving theory, one of the parent's roles is to keep the child safe for the purpose of her/his survival [53], as well as to serve as a secure haven, helping the child in times of need to cope with stress and anxiety [54]. However, within the context of political violence, the parent may feel that his/her ability to fulfill the two roles is compromised, and thus experience heightened emotional distress (e.g., [55]). This may trigger a cycle in which the maternal caregiver's functions- for instance maternal responsiveness and sensitivity- are limited due to higher maternal stress, and this in turn may have a negative effect on the child's well-being (e.g., [10,55]). Accordingly, the current study found that the more mothers reported experiencing stress, the more they tended to report that their child suffers from emotional problems. Although parent's reports may be biased by their own distress and symptomatology [56], significant associations were found in the present study between the mother's and teacher's reports about children's difficulties. Thus, the current study confirms the presence of associations between maternal stress and the child's emotional problems. In regard to maternal stress and child's variables, we found that among girls only, the more the mother felt stress, the more EF problems her daughter has, and the more girls had EF problems, the lower their sense of self-potency. Bearing in mind the small sample size of the current study, particularly in the context of the effects of gender, further research is required in order to determine whether girls' executive functioning is vulnerable to the effects of maternal stress or distress.

Regarding children's self-potency, which was assessed from the child's perspective in this study, we found a surprising finding of no association between maternal stress and the child's self-potency. We would like to present three possible explanations for this finding. First, from a developmental perspective, the child's sense of self-agency is formed early in her/his development, and this may show continuity despite contextual liabilities over time (e.g., [57]). Second, the present study did not measure additional parental aspects that are known to have an impact on the child's variables in the context of conflict, for example parental support and monitoring [58,59]. In addition, the present study did not include paternal variables that have a crucial influence in the context of the triadic relationship, and in particularly during stressful events, when the child might calm down as a result of reaching out to the other parent, or even by watching her/his parents in a calm dialogue or discussion about the situation [60]. The third explanation leans on the ecological perspective which takes into account the potential of the community to mitigate the effects of armed conflict [61,62]. Specifically, in the present study, the children and their families live within mainly small collective communities that have already established social and community initiatives and services designed to help their members cope with the daily events of violence [63]. Thus, we can suppose that the community, which also includes members of the kindergarten staff, functions as a buffer and as a protective factor in regard to the children's sense of self-potency. This speculation is in accordance with one of the dimensions of the self-potency concept- which includes the individual's perception of the social envelope as a source of support in times of need [28].

#### *4.1. The Drawings as a Source of Insight*

The current study sought to examine associations between PPAT drawings of normative and functioning children who live in an area of persistent political violence and children's and maternal variables. The study's results illuminate the potential clinicalassessment value of careful observation of positive and negative aspects of children's PPAT drawings in relation to emotional strengths and cognitive difficulties. The drawings were analyzed and coded quantitatively into art formal elements, content aspects and categories of narrative focus.

Although the form and content aspects were similar to those found in a normative sample of the same age group that lives in a non-war zone in Israel [39], when we examined the focus of the drawings' narrative, we found significant differences: the current sample had more drawings lacking a picking scene narrative, and more drawings with a rival narrative diverting the observer's attention away from the apple-picking scene. These differences may be explained by the unique context that may have interfered and reduced the children's focus on the task's requirements. This could happen for example when children felt preoccupied with other issues, and/ or were overwhelmed with feelings of unsafety, or were vigilant or fearful in the context of living in an area of political violence. Bearing in mind that the drawing task was conducted in the kindergarten setting with an unfamiliar adult (the researcher), it could have been experienced by the children as a stressful event (e.g., [64,65]). Although the same procedure was used in the previous study [38], it may be speculated that the children in the conflict zone suffer from the cumulative effects of stress and this affected their performance. Psychoanalyticallyspeaking, war interferes with the continuous flow of everyday life and impairs one's sense of personal safety; thus, it may impact attention and perception [66].

Sandler [67] has introduced the concept of 'background of safety,' as a feeling generated from our perceptions:

"the successful act of perception is an act of integration that is accompanied by a definite feeling of safety—a feeling so much a part of us that we take it for granted as a background to our everyday experience". (p. 353)

Gampel [66]), who researches the experiences of traumatized individuals, has identified the opposite experience, the 'background of the uncanny' which emerges from situations of living in contexts of social violence. Overwhelmed by horror and pain, our precepting and symbolization abilities are impaired. Neuroscience has shown that long-term and chronic stress are related to the heightened secretion of cortisol levels, which may negatively affect cognitive functions by changing processes of neurotransmission in the hippocampus and the prefrontal cortex [68]. We can thus speculate the abilities of 56% of the children who took part in the present study and who drew a missing picking scene and a rival narrative were compromised, and they could not respond accurately to the requirements of the task. Furthermore, it can be speculated that these children communicated authentic subjective experiences in their drawings, such as the experience that something is missing, and thus fragmented, as in the cases of the missing picking narrative; they also communicated the need to process other "burning" issues that they were coping with, such as by drawing PPAT drawings with a rival narrative.

An additional important finding in this study was that children who drew PPAT drawings with a missing picking narrative reported the lowest self-potency in this sample. This could indicate that these children, whose PPAT drawings revealed a deficient triad (absence of a person, or tree, or apple), implicitly communicated the basic experience of being unable to reach a goal, or to be part of an effective triadic alliance. These are the children that might be in a need of therapeutic intervention to strengthen their sense of self-potency.

The present study also showed evidence of associations between PPAT drawings and emotional and cognitive aspects; in some cases, gender was an intervening variable. The child's experience of self-potency was found to be positively associated with positive aspects in the PPAT drawings; specifically, the more the child perceived herself/himself as self-potent, the more the drawn tree was generous, the drawn person was active, and the drawing was realistic. In addition, the PPAT drawings tended to be more vivid among girls with stronger self-potency. The differences in the expressions of positive experiences of potency between the two genders may be related to meta-analysis reviews showing that girls express more positive emotions than boys (e.g., [69,70]). As mentioned above, in a previous study with the same age-group in Israel in a non war-zone [38]. we found associations between self-potency and positive content in the PPAT drawings only among girls.

Children's difficulties, as assessed by teachers and mothers, were inversely associated with the PPAT drawing's main factors; specifically, as much as the child coped with cognitive and emotional problems, the less generous the drawn tree was, and the less realistic the drawing. EF difficulties were also found to be associated with a less active drawn person, and a less vivid drawing. The present study used a regression model for calculating the explained variance of the PPAT drawings by cognitive and emotional child's variables. Findings show that the children's cognitive difficulties and their selfpotency experiences explain only about 30% of the variance of the As-Real-R and the Tree Generosity factors. This indicates that there are additional aspects—not measured in the current study—that may explain the variances of these drawing, for example, creativeness; motivation; cognitive abilities; social abilities; and environmental variables, such as art, education, and more.

In summary, these results may be described as a parallel-track visual communication pattern, in which children express their personal strengths through positive form and content aspects of the drawing, and express difficulties through negative form, content and narrative aspects of the drawing.

This pattern was found in many studies that showed that positive visual drawing elements were associated with child's strengths and abilities, and negative visual drawing element were associated with children's emotional difficulties (e.g., [71–73]). Evidence of this parallel-track visual communication was also found in studies of children's PPAT drawings, for example, poor line quality and less color prominence were associated with children with ADHD [74]; a less active drawn person was associated with EF difficulties, while an active drawn person was associated with higher self-potency in preschoolers [38] and with maternal hostility/aggression among school-age children [37].

The present study, however, also identified an inverse pattern, where negative child/maternal variables were associated with positive expressions in the PPAT drawings. These may be understood as hope expressions, correcting or compensating efforts, and/or wish fulfillments representations in the face of experiences of adversity and helplessness. In this pattern, we found gender to be an intervening variable. Specifically, the more the mother reported feeling depressed, the more her son tended to draw a more generous tree, and the higher the exposure of the daughter to events of political violence according to the

mother, the more the daughter drew a more active person, a more generous tree, and a more realistic drawing. Psychoanalytic theory perceives compensating expressions/behaviors for example, the phenomenon of an imaginary friend that functions as a buffer against loneliness—as non-adaptive; moreover, when the child relates to the imaginary friend in a concrete way, it also prevents him/her from establishing real relationships [75]. In the same vein, artistic products are interpreted as expressing contents of the unconsciousness as wishes and expressions of the forbidden (e.g., [76]). Furthermore, children's drawings may demonstrate children's defense mechanisms [77]. However, drawing could serve as a space in which the child could engage in intellectual play [78], and/or immerse himself or herself in the potential space and move along the bridge that connects reality and phantasy [79]. Accordingly, the child could feel he or she is a problem solver, a creator and one who controls the drawing medium [80]. Symbolic expressions of hope and overcoming may be understood as child's attempts to self-soothe, regulate emotions, have fun and make the shift from helplessness to agency. Thus, further research is recommended to better understand the nature of the positive images and visual elements in the PPAT drawing in the face of negative life contexts; specifically, do they serve as non-adaptive compensations defenses, or expressions of hope and positive thinking that may be part of adaptive efforts.

The present study's findings—that of a parallel-track communication pattern reflecting child's difficulties and strengths through negative and positive aspects in their drawings versus compensating positive expressions despite experiences of deficit and stress—may demonstrate the complex nature of the individual's drawing as it relates to multiple and sometimes contradictory phenomena. These findings may also reinforce McGrath & Carroll's [81] tenet that projective tasks, including projective drawings, may serve as a portal for the broadband implicit representations of the individual, and not as valid personality tests. This implies that extra caution is needed when endeavoring to draw conclusions from children's drawings.

#### *4.2. Study's Limitations, Future Recommendations, and Clinical Implications*

The present study's sample size was 156 children, however teachers and mothers' compliance rates were low, so sample size was reduced when calculating some of the associations and regression tests. We thus recommend conducting this research with a larger sample size, and in particularly, to examine the gender differences that were found in the current study. Secondly, we did not measure paternal variables, and this may be a limitation due to the fact that both parents affect the way children function in many areas. Another limitation is that we did not measure child's behavior factors in times of stressful events, for example behavior of comfort-seeking was found as associated with child's better functioning in a war-zone [82]. This article, because of space limitations, does not include a description of the processes of drawings and the accompanying verbal stories of these children that were also documented.

Regarding clinical implications, the present study makes three main contributions. The first is the ability of PPAT drawing to communicate issues of self-potency, particular drawings that lack the picking narrative. The second is the potential to gain insight about children by observing the narrative of their PPAT drawings. In cases of a missing picking narrative or the presence of a rival narrative, the clinician should ask herself/himself: What is going on? Why is the drawing incomplete? What is the meaning of the absence of an element? What is the meaning of the additional predominant narrative? Thus, the narrative focus may serve as a source of questions and associations for the clinician, as part of her/his effort to better understand the child. The third contribution is addressed towards the psychotherapist who works with children's drawings. The findings which reveal the presence of a parallel-track communication pattern, along with compensating and/or hope expression pattern, demonstrate that each drawing may be the individual child's idiosyncratic expression. This means that in our endeavor to understand the messages the child communicates through his/her drawing, we need to carefully observe each nuance in the drawing, whether it is in the form or content layers, and determine if it illustrates a

positive aspect reflecting the child's strengths or a compensating or hope expression in the face of adversity. During the therapeutic process, after the child draws the PPAT drawing, the clinician may ask questions or even invite the child in a playful manner to draw, for example: "What happened next?"; "what happened before?". These questions/invitations may serve as therapeutic interventions in the case of a picking narrative that is lacking, or the depiction of a passive figure, and may thus enliven the narrative and encourage the child's expressions of self-potency.

#### **5. Conclusions**

The present study shows that the PPAT drawings of preschool children who live in area of persistent political violence convey the children's strengths and difficulties through layers of form, content, and narrative. The narrative focus's analysis of these drawings revealed that these children express much more either competing narratives and either the absence of a picking script, than children who live in a non-war zone. These specific narratives may be interpreted as a call for further attention to the child's self-potency and unique needs as they may arise in the context of the 'background of the uncanny' [66]. The associations between the PPAT drawings main factors and the child's variables that were found in the present study reveal a complex communication dynamic. Specifically there are parallel patterns pertaining to the child's strengths and difficulties as they were associated with positive and negative visual aspects in the drawings, respectively, as well as the presence of compensating expressions, reflected in the positive visual aspects created despite adverse context's variables. These expressions may be interpreted as resisting responses against helplessness, and/or restoring actions towards safety, at least in the child's creative space. Further research is required to illuminate on these two patterns; for example, what are the characteristics of children who lean more on parallel versus compensation patterns in their PPAT drawings? What is the role of gender in this respect? Is resiliency related to these forms of communication? Finally, if it is, can we use children's drawings to learn more about how they are coping in the face of stressful event? And how can we further foster their coping skills through art-based interventions?

**Author Contributions:** Conceptualization, M.B.O. and R.I.; Data curation, N.B. and O.S.; Formal analysis, N.B. and O.S.; Investigation, M.B.O. and R.I.; Methodology, M.B.O. and R.I.; Project administration, N.B. and O.S.; Software, M.B.O.; Visualization, M.B.O.; Writing—original draft, N.B.; Writing—review & editing, M.B.O. and R.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was approved by the the Israeli Ministry of Education (protocol code 10352 date of approval 14 August 2018), and by the Ethics Committee of the Faculty of Social Welfare and Health Sciences at the University of Haifa (protocol code 459/18 date of approval 12 December 2018).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

