本研究的主旨在於瞭解受性虐待與未受性虐待兒童，在Sidun & Cha
se（1985）所發展的畫人測驗檢核指標上之差異情形，並進一步探尋 具有區
辨力之圖畫指標。本研究採對照團體之實驗設計，控制樣本特性 與施測情
境，每位受試者皆接受畫人測驗之實施。受性虐待組（實驗組） 之樣本來自
北、中、南、東部之社會服務機構共35名女學童。未受性虐 待組之樣本係
根據實驗組樣本之性別、年齡、父母社經地位水準等樣本 特性選取來自北
縣兩所國小低、中、高年級學童共70名女學童，以兒童 情緒困擾行為評定
量表與教師提名方式，分為情緒困擾與適應良好之兩 組對照組。
研究者收集圖畫之後，再
由三位評分者根據Sidun-Chase畫人測驗 檢核工具，進行人物畫之檢
核。檢核後所得資料，分別以卡方考驗、t 考驗、區別分析等統計方法進
行分析。研究結果發現如下：
一、 性虐待兒童與未受性虐待兒童
，在Sidun-Chase畫人測驗部份檢核 指標上有顯著差異。

（一） 性虐待兒童與未受性虐待但有情緒困
擾之兒童，在Sidun-Chasec 畫人測驗部份檢核指標上有顯著差異存在。性虐
待兒童出現較多 比例的指標，共計有8項指標。其中差異達P < .01
顯著水準的 指標僅有「手指/缺漏」一項；而差異達至P <. 05顯
著水準的指 標有「線條的力道感/異常粗重」、「多個人物像」、「
耳朵/大或特 殊的」、「鼻子/強調」、「嘴巴/缺漏」、「手部/缺漏」
、「腰部區域/ 出現肚臍」共7項。
（二） 性虐待兒童與未受性虐待且適應良好之兒童，在Sidun-Ch
asec 畫人測驗部份檢核指標上有顯著差異存在。性虐待兒童出現較多
比例的指標，共計有11項指標。差異達P < .01顯著水準的指標
有「人物像的位置/左上」、「多個人物像」、「頸部/缺漏」、「手部/
缺漏」、「手指/缺漏」、「腰部區域/軀幹底部未封閉」共6項；而
差異達至P <. 05顯著水準的指標有「幾何圖狀的人物像」、「耳
朵/大或特殊的」、「嘴巴/缺漏」、「腰部區域/出現肚臍」、「衣服/
缺漏」共5項。 （
冉） 然而，研究發現兩組未受性虐組出現較高比例之特徵指標有「衣 服/強
調領帶」、「衣服/強調或精細畫出腰帶」、「線條的質地/堅定 的」、「
線條的質地/素描式的」、「不完整的身體結構」、「塗黑/足 部」、「塗
黑/軀幹」、「鼻子/缺漏」、「人物像的大小」、「人物像的 位置/中央」、
「以圖紙為基底的人物像」、「擦拭的數目」、「圓形 物數目」、「楔形
物」、「腿部/雙腿緊貼著」，合計15項，此結果 則留待後續進一步探
討。
二、 以Sidun-Chase畫人測
驗檢核指標區辨受性虐與未受性虐兩組，正 確區分率為96.19%。再以逐步
刪減法篩選出其中較具有區辨力之10 個指標為「人物像的大小」、「
手指/缺漏」、「圓形物數目」、「人物像 的性別特徵」、「腰部區域/出現肚
臍」、「鼻子/強調」、「擦拭的數目」、 「只有頭部」、「嘴巴/缺漏」、「擦
拭/骨盤區域」，其正確區辨率為 79.05%，其中又以「手指/缺漏」、「鼻
子/強調」、「腰部區域/出現肚 臍」三項指標最具區辨力。

最後，根據研究結果作進一步的探討，並提出建議
，以供從事有關 心理諮商領域實務工作者及未來研究之參考。

The purpose of the study was to investigate the difference between
sexually abused and non-sexually abused children on Human-Figure-
Drawings graphic indicators which was developed by Dr. Sidun and Dr. Chas
e in 1985. Furthermore, the researcher would like to find out the most po
werful predictors in case screening between sexually abused and non- sexuall
y abused group through the process of list-wise deletion of discrimin
ant analysis.
The coding results of Human-Figure-Drawings (HFDs) of 35 sexual
ly abused female children were compared with those of 35 matched
non-sexually abused emotionally disturbed children and 35 matched no kno
wn adjustment difficulties. The subjects from elementary schools in the
two comparison groups had no known sexually abused history and were ma
tched for sex, age, and parents' socioeconomic status. A questionnaire an
d teacher recommendations were used as the tools to select th
ese two comparison groups. The subjects of experimental group (sexua
lly abused children) were from social welfare agencies. The coding result
s were rated by three trained coder. T-test, chi-square test, and discrimin
ant analysis were used to analyze the data.
The results are as followings:
1. There are significantly different, b
etween sexually abused group and non- sexually abused but emotionally distu
rbed group, on several graphic indicators of the HFDs. The sexuall
y abused children tend to draw more of following : figures omitted, line
pressure/unusually heavy, multiple figures, large or unusual ears, nose emp
hasized, mouth omitted, hands omitted, and presence of navel. Non- sexua
lly abused but emotionally disturbed children tend to draw more of follow
ing : tie emphasized, belts elaborated or emphasized, line quality/
sketchy, poor body integration, feet shaded, nose omitted, and wedges.

2. There are signi
ficantly different, between sexually abused group and non- sexually abused a
nd well-adjusted group, on several graphic indicators of the HFDs. The s
exually abused children tend to draw more of following : placement of
figure/upper left quadrant, multiple figures, neck omitted, hands omitt
ed, figures omitted, bottom of torso not closed, geometric figure, large or
unusual ears, mouth omitted, presence of navel, and clothing omitted
. Non-sexually abused and well-adjusted children tend to draw more of fo
llowing : erasures, circles, line quality/sketchy, paper based figure, tor
so shaded, feet shaded, and leg pressed closely together.

3. Sidun-Chase scoring categories were able to correctly differentiate the
sexually abused children from non-sexually abused children. The percent of
correct classification is 96.19%. After the process of list-wise deletion
of discriminant analysis, there are left ten most powerful predictors, inc
luding figure size, figure omitted, circles, sexuality of figure, presence o
f navel, nose emphasized, erasures, head only, mouth omitted, erasure/pelvic
region. The percent of correct classification of the ten indicators is 79
.05% . Among the ten predictors, there are three indicators with the most po
werful prediction.

Based upon the above results, relevant assessment and
counseling implications were discussed and suggestions for future
research and practice wese presented.