The Application of theSentence Completion Method in Junior HighSchools in Taiwan 121
THE APPLICATION OF THE SENTENCE COMPLETIOM
METHOD IN JUNIOR HIGH
SCHOOlβINTAIWAN
CHIEN-HOU HWANG
INTRODUCTION
Since the extention of the public education system in the Republic of China
,
guidance program in junior high sen∞，Is has become an important part of the sch∞1 activities and has made some achievement in last three years. Psychological tests which were infrequ-ently used in high senωIs have also become routine instruments. In most of the scn∞Is， general ability tests are usually administered in the first year (or the seventh grade) while aptitude and personality tests are given in the second and third years (or the eighth and the nineth grade). It is an encouraging sign as far as the application of psychololgical tests is concerned; but it has alω ∞nstituteda problem. In Taiwan today,the supply of tests is still far from being sufficient. There are several intelligence tests and two to three 恤tteriesof aptitude t臼.t. In the case of p訂閱nality test,
however,
very little is available for adolescents of junior high sch∞I level. The demand of more suitable tests in that category is,
therefore,
very great.Psychologists in this country have realized this demand
,
they are working hard in preparing more tests to answer the need of the community. At the same time,
theyare aware of the fact that in m品.t scb∞Is， there is only one guidance worker who has some experiences in using tests while the enrollment of students is usually in thousands. In theωcasion of test administratio
n,
assistance of all other regular teachers will be needed.They will be the examiners though they have had no training in using tests. The present author often feel that for sch∞I guidance program today
,
a test should fulfill the following criteria:a. Easy to administer
,
preferrably group test.b. Can be easily scored and analyzed.
c. The result can be understαxl and useful to regular teachers.
Having reviewed a large number of tests, the present author has found that the Sentence
Completion method is onE of the techniques that are suitable for the use injunior high sch∞Is in Taiwan. He has made a trial; the result of which is summerized in this report.
THE SE~TENCE COMPLETION TECHNIQUE
A. Previous Studi個 of theSeDteuce Completion Me曲“
The Sentence completion Technique is a rather simple method. The subject is given a set of sentence fragments or stems and he is asked to complete them into full sentences. This method is frequently used in language t曲ching programs. Psychologists first used this method in measuring inteU但tual ca開cities (Ebbinghaus
,
1897; Kelly,
1917). In recent years,
however,
it has mainly us吋 for personality assessment since the pioneer attempt of Payne (1928) and Tendler (1930).Goldberg has made a very thorough review ofSentenceCompletion method in per甜n
ality ass個sment. He pointed out that this method had been used in a great variety of researches: in studying attitudes toward some specific groups of people or certain objects; in predicting achievement for specialized groups; in studying differences between a wide variety of contrasted groups; in evaluating adjustment of various clinical groups(Goldberg
,
1966)
Because of the great flexibility of the Sentence Completion meth吋， each research often developed his own form of SCT to meet his specific clinical or research purpose. However, there have been attempts to develop standard SCI' forms among which Sacks' Sacks Sentence Completion Test (SSCT) and Rotter's Incomplete Sentence Blank (ISB) are examples of two different approaches.
Sacks, in compiling his test forms, grouped his sentence stems into four clinically significant categories: Family
,
Sex,
Interpersonal Relationships and Self Concept. These four areas are further divided into 15 sub-categories with four sentence stems in cach of them (Sacks and Levy,
1950). This seems to be a very sound appr個ch in which the responses of a subj配t can be readily classified. The information secured in this manner will fit the scheme that a researcher plans to collect. Meanwhile,with structured,content determining stems,
it will be difficult for a subj配t to make evasive responses.Rotter and his associates have chosen a different approach by using unstructured stems. They feel that there is no experimental evidence that structured stems would elicit more signi訂閱nt responses. On the contrary
,
well-structured stems often constitute threat to subjects. Their objective was to develop a group screening instrument that would provide a summary score of general psychological adjustment. With simple and general stems, Rotter's ISB has gained wide acceptance and has the most objective scoring system (Rotter and Rafferty,
1951).The Application of theSentence Completion Method in Junior HighSchools in Taiwan 123
The Chinese Sentence Completion Test (CSCT) used in the pr閏ent study is a revision of Rotter's ISB. As the original sterns are simple and short such as
“
I like....",“
Most girls... 人 etc. ， there is little cultural factor involved in them; so the translation of the ISB can be used in Taiwan without any unhealthy effect. However,
there are still two modifications: A. Two items (“The kind of person that I would like to be...." and“
The test... .") have been added and so there are 42 items altogether; B. The order of some of the items has been chang吋.There is no time limit for CSCT; for junior high sch∞I students in Taiwan
,
the Test can be completed in2ι30 minut個. Most of the young people work on it with great 臼se， and usually have no difficulty to understand the instructions. During the administration of C住了， two points are always emphasized: a. one should feel free to put down his true feelings,
and b. there is no right or wrong answer to any of the stems. Subjects are also told that they may use phonetic symbols when they find difficulty to write a certain Chinese character. In actual administration of 臼CT in the present study,
the Test is given with a short form of Semantic Differential Test which takes abωt 15 minutes.su凹ECTS OF THE PRESENT STUDY
Subjects included in the present study are 7th,8th and 9th graders drawn from three junior high schools in Taipei 訂閱. One of the three sch∞Is has male students exclusively the second is a girls sch∞I while the third is one of co-education. Over 5ωstudents were t臼ted. After the elimination of the incompleted ones
,
the records used for analysis numbered 449 the distribution of those subjects is given in Table 1.Table 1. Distribution of Grades and Sex of the Subjects
Sch∞l A B B C
Sex Girls Girls Boys Boys Total
1st Year 33 40 27 37 137
2nd Year 40 39 42 35 156
3rd Year 38 38 42 38 156
Total 111 117 111 110 449
ANALYSIS OF THE RESULT
A. The Objective S伽ing
One way of analysing the sentence completion records is the quantitative scoring meth01，有 score each respen田 according to a specific manual. There are several such
scoring systems,and the one developed by Rotter and Rafferty (1950) is often considered the most rigorous and objective one. According to their meth叫 each response is scored on a 7-point 民ale from 0 to 6 for its degree of ∞nflict. The principles of categorization
are as follows:
P3-0 point: Responses which indi個te a generalized positive f凹ling toward ，但ople， glαxl social adjustment
,
healthy family life,
optimism and humor.PZ-l point: Responses with clear-<ut glα地 natured humor, real optimism and warm ac-ceptance.
PI 2 仰ints: Responses which deal with positive attitude toward sch∞I hobbies
,
sports,
expressions of warm feelings toward some individual,
expressed interests in阱罰。pIe and so on.
N 3 個ints: Neutral responses not falling clearly into either positive or conflict categories, generally on a simple descriptive level.
CI 4 points: Responses in which 回me kind of problems
,
difficulties,
or complaints are expressed.CZ 5 points: Responses with expressions of inferiority
,
psychosomatic complaints,
concern over possible failure or vocational choice,difficulty in interpersonal relation-ships,
etc..C3 6 points: Expressions of severe conflict or indications of maladjustments such as suic一
idal ideas,severe family problems, and strong negative attitudestoward 戶均pIe， and so on.
In addition
,
Rotter and Rafferty have provided a manual in which examples of responses in each categories are given.While Rotter's principles of categorization are well formalized, his manual of scoring which is based on the respon記s made by American subjects is obviously not suitable for the use in Taiwan. The present author has to prepare a manual for scoring SCT respon-ses from Chinese subjects. This is a very time-<onsuming task and the scoring manual for each sex has to be prepared separately. In doing 凹" 50 records were selected by random. The r臼ponsesof those subjects to each item were carefullyexamin吋 and clas-sified into the above mention吋 categories (P3-C3). The judgment was made on the meaning of a response as well as its frequency among the 50 records. Three judges worked on this and fully agreement of three was r問uired to classify a response into a certain 臼tegory. It t∞k a long time to classify the respon時s to all 42 items. Some
{~xamples are given below:
To the stem
“
PEOPLE" (for female subjects)The Application of theSentence Completion Method in Junior High Schools in Taiwan 125
m
“...are kind",“...are hard working", "...are happy and active". PI “...are rational animals",‘'...help each other"; “...have ideals",N
“
...should c∞perate with one another" ，九....are born equal弋“
...are members of the community".CI “...are r間listie",“...are busy in their own business", “...are not always friendly".
C2 “...are stupid"，“...are 甜lfish"，“...donot res開ct me". C3
“
are∞ld-bloodanimals".To the item "AT HOME" (f叮 male subjects)
P3 ‘“.……..…....1 am v呃er叮y happy"弋，‘“‘九.……..……..……..…..i址t i泌s V，惆'er叮y wa缸rm"
n
‘“‘.………...I p抖la叮yw叭it由h brothers and s馳i泌st切er吋ss，"，，f.“‘.……..……..…….川…...I' m a gαM滔:i b切oy"九.
PI ...1 help the house work", ...I can have a sound sleep" ，‘ ·...1 個n watch TV comfortably".
N “...1 am the oldest child",“...1 watch TV every night",“...1have a pet". Cl ‘·...it is not very interesting",“...1 cannot relax myself".
C2 ‘“‘仁..…...…..I often get i削11"九，“九..…….叮…....…..my p開aren此ts are sometimes angry a前t me"刊， C
臼3 ‘“‘..………....1.. am always the sc臼a脾go咽at"九， “叭九...…….川……..…..1f間叫1 司v珊F甩er叮y 10∞n】ely"
B. Data of 曲e Present Study
1. Reliabillity
While all 449 CSCT records are scored by the author himself
,
98 of them are also scored by a second scorer to determine the reliability of the test. The coefficient is found to be 0.82 which can be ωnsidered satisfactory as it falls within the range of the reliability coefficients (0.80-0.94) achieved by judges who used a scoring manual to assign responses to different categories (Rotter and Refferty,
1950).2. Validity
The scores of CSCT have been validated against two other measures: (的 Eysenck's
Maudzley Per切nality Inventory, neurotic scale; and (b) Teachers' ratings on general adjustment of students. The CSCT score has a correlation of 0.28 with the former. AI-though it is not very high
,
it still indicates a positive relationship. Sincecscr
aims to measure general adjustment instead of neurotism alone, low correlationis probably what one should expect.The CSCT score cOrfl~lates even lower with Teachers' Ratings and the coefficient is 0.23. The Ratings made by the teachers are actually the sum of the ratings on ten su
b-民ales of which the important ones are: Academic Achievement
,
Relationship with Peers,
Relationships with t聞chers， Participation in Group activities,
Diciplinery problems,
and General Health ωnditions.When examined closely, academic achievement seems to be the key scale which has dominated the whole picture. Teachers tend to all other 民al臼 high when a subject is high in achievement sca峙， and when a subject has low achievement, all other scales will be rated low. The halo effect is very obvious. This generalization is not completely groundless
,
however. In high sch∞Is today,
academic achievement d個s have effect on other aspects of on的 life. Koh,
in a study of under-achievers,
finds that 晶晶e students have 阱。r adjustment in sch∞1 (Koh,
1972). But there is still insufficient evidence to say that academic achievement means everything. When one realiz田 that teachers' ratings have been immensely influenc吋 by students' grades,
he would not be surprised to see the low correlation the CSCT has with them. Perhaps the best thing we 臼n 胡y now is that CSCT is measuring something different from what teachers are rating.3. The Distribution of the 臼CTSωr臼:
For the 449 students tested in the present study
,
the distribution of their 臼CT scores is given in Table 2.Table 2. Distribution of CSCT Scores.
Score Female Male
150-154 3 145-149 3 2 1壘。-144 4 2 135-139 6 6 130-134 20 7 125-129 21 21 120-124 29 35 115-119 29 49 110-114 48 29 105-109 25 36 100-104 24 23 95- 99 12 9 90- 94 1 1 85- 89 2 Mean 116.55 115.50 S.D. 12.42 9.18
The ApplicationQ£theSentenceCompletion Method in Junior HighSchools in Taiwan 127
It is interesting to find that the mean scores of both sexes are very close to one another and they can be regarded as the normative data for girls and 1:的ys in junior sch∞Is in Taiwan today. However
,
the author has no intention at this moment to suggest a cutting s∞re which can be used as a dividing line between adjustment and maladjustment. Fur-ther study is n自d吋 for the determination of such a cutting s∞reoC. The Impre鷗ionistie Approach
In contrast to the obj也tive scoring
,
the handling of sentence completion material can also follow the impressionistic approach. This method is specially favored by clinical psychologists who feel that although a more 句bjective handling of the data sounds like a reasonable desire,"“. . .
an effort in that direction at the pressent time would be prema-ture;" and they believe that“
more and better material can be acquired by a pr囚el 阻 of interpretation sentence by sentence until an acceptable glo加1 description is achieved" (Holsopple and Miale, 1954). Stein (1947), Levy (1950),and Fover (19印) are the out-standing scholars that favor the non-quantitative interpretation of sentence completion material even though there is Iittle empirical support for the impressionistic approach in叩otent analysis.
In the present study
,
the author examines this problem from a rather practical viewp-oint. I沁ubtlessly ，there are many advantages in objective scoring. However,we see clearly that even with a manual,scoring of sentence completion re∞rds still requires good trai-ning because the manual provides only a few examples
,
and in many occasions,
it is not 個sy to assign a response into a certain category. Besides,
it takes a glα地 deal of time to score a large number of records. For tne use of sentence completion method in sch∞Is in Taiwan t吋ay， these are the matters that should be considered. While each sch∞l often has thousands of students,there is one guidance worker who may have some training in testing. Even if he can handle the scoring all by himself,
other teachers would still have difficulty to understand what the score means. As Class Tutors are often invited 切 participate in many of the guidance activities,
it would be a glαxl idea 切 ask them to read the 臼CT records and to pick out the significant respon阻s. Being able to contribute more than just reading a score marked by some1x滔y else,
these Class Tutors may feel better and consequently may be∞me more interested in test programs.To test the feasibility of this idea
,
two Class Tutors in a girls sch∞1 have been asked to examine 50 CSCT records and to select from 聞ch record these responses that might indicate adjustive problems. They seem to have done the job very satisfactorily. Though these responses are not“
significant" from the view point of depth analysis,
they are indications of emotional concerns. They are very useful in counseling because those are the problems that a subject is ready to admit and willing to cope with. Below are several examples:A. Subject CF1321 Girl
,
9th Grade,
Total SCT score: 126IN SCHα)L I am not an outstanding student.
WHAT PAINS ME is that on one in my family knows my trouble. THIS TEST has made me aware of many problems.
Though her total score is not very high
,
the above mentioned responses suggest a state of general anxiety and inferiority feelings. Relations with family membersal田 deserve special attention. A teacher or counselor can easily see the n自d to discuss these with the subject.
B. Subject CF1332 Girl 9th Grade Total CSCT 興。re 132
THE BEST THING is to become taller and more beautiful
,
but not to gain any more weight.I FEEL that I am short and fat.
I HATE that I cannot be taller and prettier.
For this girl
,
her great concern is very obvious. Surely she needs help in regard to the matter of weight control and to the development of healthy attitude toward one's physique. Development of self acceptance seems to be the g明I of counseling.C. Subject CF1352 Girl (9th Grade) Total 臼CTscore 136
A MOTHER is dead.
AT HOME
,
it is dull and lonely. MY FATHER is d個d.IN SCHOOL
,
I act as if I were free from, worry.A teacher would natually want to know more about the family of this subject; the latter's need of affection and support is obvious.
D. Subject CF1128 Girl 9th Grade Total CSCT score 124
THE BEST food is what my mother has prepared for me. I SECRETLY put my pro臼entin mother's hand bag. AT BEDTIME I always bid ‘G∞n night' to my mother.
Positive feelings among these response indicate a healthy relationship with mother. This certainly tells more than a single score. But one may also be interested to know her attitude toward father.
The Application of theSentenceC咽.pletion Metb吋 inJunior High schools in Taiwan 129
After 旭.ving spotted th個e signif胸前 r個po阻臼， the Tutors h叫 interviews with t胸e subjects and disc凶sed with the latter the matters related to th個e responses. The discussions were held in friendly atmosphere in which the students express吋 freely their f聞自 and anxieties. With the suppo此 of the Tutor
,
those subj配ts could release their個叫詢nal tension and gain more insight about their problems. With the exception of thr甜甜Ibj配ts who had to be referred to the Counselor
,
all others 記em，吋切 havebeen benefited by these discussions with the tutor. In a report,
the Tutors later pointed out 儡做個) In most of theCSCT
r郎。rds， one can getωme information which is not avail... a恥泌 other resources in school files,
(b) In general,
thestl.臨的 are willingωdisc胸 伽tr problems that have 加en revealed in 臼CT， and (c) the experienc晶 are rewarding to 00ththe Tutor and the students.Thepr船ent author has no intention to limit the u揖 of CSCT to impressionistic aPl• roach, all 加 wants 個 pointωtis that 油 the pr臼ent schc沁11 situation
,
this method 甜ems 給 hav~ practical value and can beused
along with the obj臨的ve scoring.缸珀GES啞ONS IN REGARD TO THE USE OF CSCT 閥割沮00凶
Tt蛤 PI'自ent study has shown thatSentence臼mpletionmethod can be used in Chine揖 junior high seh∞Is. This technique has many advantag臼us features among whicb the follow時 pointsare of importaI悶:
1. Most of the subj民ts are interested in taking the CSCT. They can see that it is not an ability 能st and COl閻明ent!y causes little “t閏tanxiety". Mol'ωvel'， the CSCT gives stl.則ents nearlyωmplete freedom in making their respon臨s， they can express t加ir feelings in their own words. It is no surpri記 that over 75% of the subjects incl叫剖 in this study give Positive respon隨 tothe item
‘
THIS TEST'.2. The CSCT can provide a g，αxl deal of information a切ut the students so that teachers
個n unders個nd them better. Although the pI"個ent form has not arranged the stems in pre-arranged a categories
,
it is still 開sy for a teacher to see the generaladjust-U鳴叫 of a st叫ent.
3. The CSCT is 組sy 個 administer and the time for testing is relatively short. It dl間S not ∞st much as the Test can be prin切d or mim個Igraphedon one single sheet. Te<立uu個lIy， it does not r·問uire much s抖出cialtraining to give 血et曲已 regularteachers can handle its administration quite 祖tisfactorily. In other words
,
the 臼CTis a very 前onomical t配hniquein regard to the labor and finance required for itsappli個tion in sch∞Is.they know very little about those instruments. This will not be so in the case of the C此T which I∞ks so plain and so simple,and does not have any element of threat. As a result, most teachers would like the CSCT better than other tests.
Based on the findings of the present study,the author likes to make the following
suggestions: ∞nceming the application of the CSCT in junior high scb∞Is:
1. Both Obj自tiveScoring and Impressionistic Appr個l ch can be adopted. As it has been mentioned before
,
objective scoring can provide an over-all index of general adjust-ment; but the scoring process does require some training. In she∞Is where a there are trained workers, the scoring can be done without difficulty. If, however, there is no trainedper田nnelin the scb∞lorheistl∞ busy to do all the scoring,the CSCT record can be treated by Impressionistic approach. By reading over a r自ord， teachers can often identify those signicant responses which indi個te worries, anxieties, or unhealthy attitudes of a subject. These responses can be u甜d as a starting point for discussion in individual counseling.2. In ordinary sch∞I situation, it is not necessary to treat the CSCT material with depth analysis. The expressed conflicts in a record often provide sufficient informa-tion that is u記ful to a guidance worker. Generally
,
a teacher should not try to push t∞ hard on a certain topic when the subject is not ready to discuss it. B曲idi缸， de醉h analysis is a task for clinical psychologists,
not for teacher counselors.3. Just like all psychological tests
,
the CSCT records should not be interpreted indepen-dently. On the ∞ntrary， the informrtion obtained from the CSCT should be put together with those drawn fr個1 other resources to get a total picture of the general adjustment of a student. In the same token,no single r臼ponse of the CSCT is to be examined separately. The general over-all atmosphere of a record is often more important. It is also advisable to evaluate a response against the normative da個 ofthe group to which a subject belongs.
4. The CSCT can be used as a means to investigate the social attitudes of a group. By analysing the CSCT records of a group, one can get some valuable material in regard to the central tendencies of the responses of its members. Such normative data will be very important reference in evaluating individual records. Furthermore,differenc明 between groups can be determined. All th臼e will be useful 旭 sch∞I guidance
prog-rams.
風.lMMARY
The Application of theSentenceCompletion Method in Junior HighSchools in Taiwan 131
completion technique in junior high schools in Taiwan. A Chinese Sentence Completion Test (CSCT) has been developed. The Test was used in three sch∞Is 泊 Taipei area ∞vering449 students who are equally divided by sex. In adopting the ObjectiveScoring appr個晶， a storing manual based-on 出e material coIl配ted in this study has been pre
p-ared. The Manual was used as a guide for assigning quantitative scores to responses made by the students. The interscorer reliability is found to be satisfactory (0.82). For validity study
,
the CSCT scores have positive correlations with the Neutoric Scale of the MPI and with Teachers Ratings on general adjustment; but with only moderate coefficients. The mean 也or，個 for boys and girls are comparable (115.50 and 116.55). Further studyis ne剖ed to establish a cutting score for the discrimination of the well-adjust吋 andthe maladjusted.
The Impressionistic appr個ch of treating the CSCT material has also been tried in this study. Under the present circumstance
,
this nαHIu個ti個tive method seems to have practical value in sch∞1 guidance programs.REFERENCES
Forer, B.京之Sentence Completi∞. In A. C. Carr (E旬， The prediction of Over
tJ>e
haviorthrough the u揖 of projective technique. Springfield
,
Ill: CharI自 Thomas，凹的，pp.6-17.
Goldberg
,
P. A. Sentence Completion Meth吋s in Personality Assessment. J. Proj. Tech. 駒， 12-45.Holsopp峙. J. O. & Miale, F. R. Sentence Completion: a proj配tive meth吋 for the study of perω>Dlity. Springfield,Ill.: Charles Thomas. 1954.
Kelly
,
T. L. In dividual testing with completion test exercises. Teachers College Record,
1917
,
18,
371-382.K嗨， S. Y. A Psychological Studyof Under-achievers in Junior High Sch∞Is in Taiwan. Unpublished master thesis,National Taiwan Normal University,1972.
Payne,A. F. Sentence Completions. New York: N. Y. Guidance Clinic. 1928.
Rotter
,
J. B. & Rafferty,
J. E. Manual: the Rotter Incomplete Sentence Blank. New York: Psychol. Corp. 1950.Rotter
,
J. B. Word association and sentence completion methods. In H. H. Anderson &Gladys H. Anderson (Eds),An introduction to projective technique. New York: Printice-Hall. 1951.
Sacks,J. M. & Levy, S. TheSentence Completion T臼t. In I.E. Abt & L. Bellak (Eds), Projective Teachnique. New York: Knopf. 1950.
Stein
,
M. I. The use of a sentence completion test for the diagnosis of personality. J. Clin. Psychol., 1947，妞.， 46-56.Tendler,A. D. A preliminary report on a test for emotional insight. J. Appl. Psycho1.,
語甸完成測驗在我國國民中學的
可應用性
黃堅厚 摘要 本研究之目的在於探討語句完成測驗的技術是否可適應於我國國民中學的學生。為了進行此項 研究，乃先蝙製一種語句完成測驗(簡稱 CSCT) ，並初步試用於台北地區三所國民中學的 4494屆 學生(男女各半〉。扭攘此項研究所得資料，並騙好評分手珊，以作為評卦時的參考。評分係據用 客觀評分法，可對學生的每一反應加以量的評定。 本翻驗的信度根撮兩評分者評分結果之間的相關來決定，其信度係數連 .82 。在教度研究方面，則發現本測驗與 MPI 中的 ~eurotie ~ale ﹒以及與教師對-般適應的評定好數之間﹒有正相關 存在( .28和 .23 )。至於良好適應與不良適應的劉分點，則讀繼續進一步研寬，方能決定。
驗客觀音字分之外，本研究還輯用非暈性的印象法 (impr翩翩嚕He al歌唱eh) 哥拉解釋學生的
