CLUSTER ANALYSIS* OF SVIB PROFILE PATTERNS
OF ADULT WOMEN AND COLLEGE MEN
PHYLLIS YUNG-EOU
Lru
University of Californi::-, Los Angeles1
In 1943, Strong argued for separate blanks for men and women in that too many women enter an occupation as a stop-gap until marriage
, would not be there
if they had a choice, are rather a heterogeneous group and, thus, would not be an adequate criterion group used to develop differential occupational scales. (Strong, 1943) A quarter of a century has passed but similar position towards using separate blanks for men and women remains largely unchanged, though with some modification. (Strong, 1955; Laima, 63; Campbell, 1968).However, in view of the increasingly merging and overlapping activities of men and women, in education, in vocation and in all walks of life, the continuing separat assessment of men's and women's success and satisfaction in a shared world is questionable. Furtuermore, it has already been found that the interests of men and women are quite similar (correlating about .69); that authors, artists, and music teachers among men score the same on the MF scale as physicians and dentists among women; and that some men have the capacity to score almost as the most feminine of women. (Strong, 1955) Unfortunately, the SVIB Male form is widely used for women as well as for men. Counselors have also often used the SVIB Male form for women, because it contains more keys. The purpose of this study is to demonstrate that the application of SVIB-M to adult women does yield mean-ingful and comparable data, in the similarities and differences of the occupational scale scores, in the profile patterns, and in the occupational groupings obtained by cluster analysis.
METHûD
Subje刮目 SVIB-M data on adult women were drawn from the women enrolled in a UCLA Daytime Extension course designed primarily to help adu1t women find a meaningful new direction after marriage and children. Their typical expressed concerns are: (1) To get some kind of gratifying, meaningful, and financially rewarding job; (2) To launch an intersting personal career which wi11 allow for future growth and development; (3) To know what to do with increasing free time; (4) To be ful 日 led in making a contribution to society, other than in the role of wife and mother; and (5) To grow in knowledge and ability, thereby enriching self and family.
* Computing ass;stanc~ was obtained from th~ Health Scienc空 Computing Facility, UCLA, sponsored by NIH Grant, FR-3.
2 心理與教育
The SVIB-M data of college men were drawn from random samples supplied by counselors in the Student Counseling Center, UCLA. About
1/3 were graduate
students and 2/3 undergraduates. Most of them were unhappy or unsuccessful with their chosen majors and were not sure what they could do with their majors and would like to transfer to a more promising field. By and large, the college men in this study constituted a rather selective group of college students who were aware of and wi11ing to use the counseling service in their search for a career commitment. Treatment of the data1. Compare the means and standard deviations for all 54 occupational scales between two groups.
2. Compare the interest profile patterns of two groups.
3. Compare their occupational groupings obtained by the c1uster analysis帶 method ，
and together, with the SVIB's given c1 assi自 cation.
The c1uster analysis method used here is a weighted variable group method (Sokol & Sneath, 1963) using Spearman's sum of variable method for recomputing the correlation of coefficient. In c1ustering, the first step is to find the mutually highest correlation as the central point of the c1uster. Highest correlation means a correlation between any two scales which is higher than the correlation of these scales with any other scales.
After the first c1uster is formed, one can proceed to determine whether an additional new scale could join this c1uster by producing an average correlation between the newcomer and the estab!ished c1uster by meeting a certain criterion not lower than the previous level of junction. If three members have formed a c1uster, one wi11 have to calculate the average correlation of the three c1uster members with a fourth possible member in order to decide whether the cluster should cease or whether the fourth member should be admitted to the c1uster. If the fourth member is not admitted, a new member with the highest correlation with the fourth member wil1 be formed to establish a new c1uster, and so on, until a certain number of clusterings are finally estab!ished to inc1ude all scales under consideration.
RESULTS
1. Comparison of the occup 在tional scales is presented in Table 1. These groups of adu1t women and college men are significantly different at the .001 level on 24 occupational scales. As a group, the adult women scored higher on Psychiatrist,
Psychologist, Personnel Director, Rehabilitation Counselor, Social Worker, Social Science Teacher, School Superintendent, Minister, Librarian, Music Teacher, Life Insurance Salesman, Advertising Man, Lawyer, Author-Journalist, Chamber of Com-merce and Business Education Teacher scales. Unlike the college men. the women scored significantly lower on Chemist, Army 0伍 cer， Air Force Officer, Math-Science
風尚 少
Cluster Analysis of sVm Profile Patterns 3
Tab!e 1. Comparison of Means and S. D丸。 f SVIB OccupationaI Scales between Adult W omen and College Men
~1A1也
VARIABLE ADUL T WOMEN (N-230)
DENTIST OSTEOPAT VETF..RINA PHYSIC工A PSYCH工AT PSYCHOIρ BIOIβ}IS ARCHITEC MA虫也\IAT PHYSIC1S CIlEM1ST ENG1N回H PRODUCTI AJlI.!Y OFF A1R FORC CARPENTE FOREST S FA且也H MA啞í SCI PRINTER POLICEr.lA PERSONNE PUBLIC A REHAB1LI YMCA SEC SOC工ALW SOC工AL S SCIIOOL S HIHISTER L1凶位R工A ARTIST MUS1CIAN Þ!USIC TE CPA Oln也 SXNIOR C ACCO叮叮A OFF工 CEW PURClIAS工 BANKER PlIARl吐ACI MORT1C工A SALESMA REAL EST L1FE 1NS ADVERT1S lA\lYFR J\UTHOR J PRESIDEN CREDIT t4 ClW4旭R rnYSICA.L COMPUT且R BUS1NESS CO~'飢UN工T OPl'OME1'R NO. -2 弓 h56789 叩斗 2 日的巧，的叮泊叩即位且月的的封閉門必叩初，立 nnω 莉， HJM 刃嗨， nMMMMUHHM 叮叮岫川岫均 m 戶戶戶自 UF 仿 阻AN (1) 26.35 2斗 .89 21，斗5 29.78 36.13 36.70 30.24 29~97 26.62 23.20 20.69 21,31 22.90 12.37 20.07 21.04 12.66 28.02 19.但 27.69 15.57 28.06 30.19 35.52 29.58 39.33 36.69 30.25 32.22 ι2.84 31.92 40.88 37.61 25.63 17.92 21.53 24.72 25.吋 24.02 25.18 29.26 28.04 35.95 35.38 40.30 38.07 40.61 25.99 28.27 38.21 25.27 25.64 32.72 25.31 37.19
-h
叮叮叩咒胡扯朽印戶戶版臼始呦，師仙訂仿仙插曲曲咒均插曲的自叫即相勾卯叮叮旬部口，如叫起仿扎叮∞，泊位松的 rhmv 袖且的印叫阿沛 3 •••••••••••• um 叩斗扭扭叩月叫“斗凶 2212 月臼斗月 98 斗扭扭斗 pmu 叩且巧斗凶口，斗叫起巧 MU 斗叫“ 92 斗斗斗口， 9MU22 月，的口坦政 MFAN (2) COLLEGE MEN (N-249) U 札凶用戶口師訂 m戶相叮喃仰自印呵，叮叮叮扭曲扭扭八仙呦，初∞叮叮 mWM 吋咒叩∞﹒帆的旬口出口∞且札泊位叮缸，叮封閉 U 均凶 Fm 巾，∞句話仿的。...
站叮叮立且又甜甜的且前句話缸品品巧旬叮 mN' 的泣如 mJ 旬泣旬叩均 MPUV 岫孔，巧叭的臼叮叮白白訕叮叮叮品站站站叩叮訕訕虫叮叮叮 O叫叫申
JTwdmM37
…
FHm5
仰
8
叫叫叫叫叫叫叫
MC
閃閃閃
mml
凹的叭叭
9
的叫♂引向
AFhvnwAnMA
斗 FOnqnMJrJLAFhvphv 汀 ρ品汀
EVFhHnUA
晶晶
A晶晶晶晶的心而神晶晶
MRUndJRUJJJ
」 ρdQdnM 扎 KVA m--2226hll2161387121 日 JRJlh5760OEJRJ5.lll22266h671367ιl 、自曲、 EA ηr 』 入叫 Nm 刀品 m戶晶戶口 HJM山的也叮竹日川叮峙的泊切句口叮叮飢旬仙們列問叫旬叮峙甜的心♂
am
叫喊川晶晶晶晶叫甘心品品
AA
耐心
mAAAAm
心 3 •• 1 ••••••••• 叫叩 98 斗扎斗 puunuu29 斗 9229ω99 斗叩 ununemμU 油油 9 叫“叫起 9 月 UU 叩 988ω899888 斗斗斗立江口Teacher, Senior CP A, Physical Therapist and Computer Programmer sca!es. At the .01 and .05 level of significance, although the women scored higher than the coIlege men on Mortician and Real Estate Man scales, they scored lower on Printer,
Sales Manager, Osteopath, Mathematician, Forest Service Man, Production Manager sca!es. Both adult women and college seem to reject the interests of Forest Service Man (scored 12.66 and 15.39) and PoIiceman (scored 40.88 and 40.43).
4 心理與教育
2. Comparison of the interest profile patterns of adult women and college men are quite simllar-correlating .8046, signi且cant at the .001 level. Collectively, the means of the scales for both groups seem to show Group VI, Creative .aesthetic, B+,
B, and B 一， to be their primary interest; Group V, Social Service; Group 1, Psychia-trist and Psychologist and Group X, Linguistic, B and B 一， to be their secondary interests. Both adult women and college men scored low in Group IV, Technical-trade and Group VIII, Business.Detail, indicating the areas of their rejection.
Chart 1. Comparison of SVIB Profile Patterns of Adult W omen and College Men
N AME V ARIABLE NO. D,EN'l'1ST 1 OSTEOPA.T 2 VET.血 WA 3 PHYS1CIA h PSYCHIA'l' 5 6 PSYCHOlρ B10lρG1S 7 ARCH1TEC 8 MA白lEI-lAT 9 PHYS1C1S 10 CHEHIST 11 ENG1NEE:R 12 PRODUCT M l1h 3 ARMY OFF1 AIR }'ORC 1L5 6 CARPENTE FOREST S 11 FARHER 18 MATH SC1 19 PR1NI'ER 20 POLICE M 21 PERsom個 22 PUBL正C A 23 REHAB1LI 24 YMCA SEC 226 5 SOCIAL W sα;IAL S 27 SCH∞ L S 28 MINISTER 29 LIBRAR工A 30 ART1ST 31 MUS1CIAN 333h 2 3 MUS1C TE CPA OWNE S間工OR C 335 6 ACCOUNTA OFFICE W 31 PURCHAS1 38 IWI1也R 39 PHARMAC1 40 MORTICIA 41 6ALES I>\A 42 REAL Es'r 43 LIFE 1NS 44 ADVERT1S hh6 5 IAWYER AlJl'llOR J 41 PRESIDEN 岫 CRF.:D1'l' M 49 Ci!A.J~血R 50 PHYS1CAL 51 C臼位1π目1 52 BUS ED TE 53 cαt.!.\UNIT 54 OPr<l1ETR c 1 •
.
iÞJ.
已先JLLi
' ,. .
,γL
' -嗯. ‘ aduJ.t wα耳e且 4 開個DENTIST 工 到IYS工C 工A 4 OSTEDPAT 2 BIO r.ρGIS 7 PHYSICIS 10 CHEMIST 11 MATHE}.<AT 9 ENGmEER 12 ARCHITEC 8
Chart 2. Tree Ðiagram of Cluster Based on Weighted Averages Algorithm一230 Adu1t W omen
39 哼，， "58 55 55 61 7~/ 仙 5~ ，崎 52 日
,
MORl'IC且 41 47 37 42 65 82 55 9EV 制糊糊 65 88-/
句句 39
43 497年 /574/7580658O
LI阻四S 山 :lp87 位，， "65
3240/ 心柵欄的**
93 /"
/ ; 49 59 59 76 83/ 78 90/ '72 79 月 70 ** REAL EST 4328/ 如/名4 33 5~ /糊糊 糊仙** 8h ** / ---.戶， τ/ / 2月， '3VA于 "'5年句8 8己 /66 71 76 70 糊糊臥血S MA 42 3斗 /80 50 68/ "** ** ** ** 98 **怖 H / / / / / / 26//2月， '30/ 恥 89/ 』757626590 峭的 MFG PRES 崎 81 71 77/ 糊糊糊糊** ** ** ** 97/
/
/一- -.一 -r 41" '30 /57 7~" 72 71 92 77 86 ** 81 89 AUTHOR J 47 37/37,/87 ** 93 67 65 50 83 82 62 ** / / h5/1255/859196H9hHH98HmvEmzsh538/ 弘糊糊 56 69 57 86 ** 58 **訕 / / 668L/7h70786』 818662575359LAMYZRh689969688929lH998』 7679H----."
- "7 - ";"'"一-，-.- --一-:;一---23" "92 月 91 ** 73 糊 糊 糊 ** ** ** PSYCHOL 6 17/55/ 告。 59 90 67 62 峙，勻6 66 66 86/
/
/
,
ARl'IST 31 ** ** ** ** 79** ** ** ** ** ** 'l!-* PSYCH工AT 5 岫~"93 75 96 72 59 **/ '88 56 59 83 84
CARPENTE 16 FAlWER 18 VETERINA 3 FOREST 工7 PRIN'l'、ER 20 PHARMACI 40 POLICENA 21 PROD 1吐A.N 13 ACCOUNTA 36 SENR CPA 35 MA-SC TE 19 COMPUTER 52 AIR FORC 15 ARMY OFF 14 ---一..--τ戶 --τ----r---.戶---.--.- l炒到/叮/斗γ7:} ，刃 70 **兒竹 76 戶 / /
'
吭，山 /5q/ 肉 '54 76 糊 糊 67 79 64 56 / 47， γh 寸見 '55 58 99 89 77 94 82 73 70 / \ 89, '80,'42 58 84 79 82 63 73 64 61 72 / 77, '80 97 80 85 81 83 糊 糊 糊 糊糊 / 65 70 80 98 72 92 ** 92 抖的** 48 ----戶-.--_...耳---3了 "53 66, '93 71 70 61 52 70 ** / 62 79 39 47/ '81 81 74 65 47 86 86 53 75 81 2吃"** 68 97 94 73 87 65 78 74 60 ** ---呵~--** 52 37 43 39 81 55 55 63 55 糊 糊 59 58 70 6年， 99 輔 相 91 ** **柵欄---
/ 28 5:).' ' 47 "** ** ** ** **糊糊糊 糊---,,-
/
27 ，峙//糊糊糊糊糊 糊糊糊糊 H / / '68/** 85 糊** ** 柵 欄 99 糊 糊 糊 MINISTER 29 MUSIC TE 33 LIBRAR工A 30 MUSIC PE 32 SCH SUPE 28 SOC工ALw
26 REHABILI 24 訊JB AD~位 23 PERSONNE 22 SS TEACH 27"
/ 41 39 59 50 47 76/47 33 41 67 78 65 一一一一-一-一一/
27 31/ 叫 60 48/ /77 54 68 76 92 55 ** ./ / 27/35 35 峙，的“ 65 78 96 73 糊糊./
,
山 50 39，/峭的** ** 峭的**. ** ** 一一一一』一-"，- -7一一一--一一一一一/一一 山山版峙'"~9771 6452 56 7~/捕時-
一 一-
一 ，戶一 / 2}/41 47/'5ν75 81 42 53 仙/輔料恃/ / / 82 59 76 51/ ** ** ** ** / / / 30 331 4~戶 75 - -./一/ 23 乃5 名9 85 52 78 61" /** ** ** **卅 / / 62/ 加 70 51 訂戶，，***.*
** ** ** ** / 4 6 9 6 3 7 8 W H H H H H柵 欄
-
-，--一一一- -;;一 CREDIT 1>1 49 18" 峙 29 時/-lHE' 柵欄峙** ""*柵欄,
"
( , / BUSED TE 53 49 35 56/ 恃糊糊 糊卅糊 糊糊糊 一;---~ / cm叫UNIT 戶 3ν32/'** ** ** ** ** *,,- ** ** ** ** / / CHAM COl~ 50 峙/JH H H H H 糊 糊 糊糊糊 糊,
,
RiYS THE 5 l h H H HHHH 糊糊糊糊 糊 YMCA SEC 25 糊糊糊糊糊 糊 糊輔糊 糊糊糊
CPA 叫NE 訕 78 70 日 77 94 85 79 83 **恥的 89
玖玖CH.ASI 38 24 46 47 80 71 53 52 ** ** ** ** **
B朋包R 39 32 32 62 日 51 65 ** 84 ** ** ** ** OFF WORK 37 36 79 65 58 77 峭的** ** ** ** 93
l
Chart 3. Tree Diagram of Cluster Based on Weighted Averages Algorithm 249 Male Col1ege Students, Ucla
D間TIST 30 3ιJ弱的 34 肉， 36 30 3哇，但扭扭
-
- - / 立γ3β30 28 3年/岫 76 9、/如 76 以此 / 31 44 28 仙過 /32 36 68" 74 46 58 92 92---r-
-,---I -,---I 8 ,'í6 ，-24，訕仙 62 ~'(8 5898
xx xx 恥 / 恤，告， isho664+ι6 84 扭扭 74 62 之/旬出川紅 80 扭扭扭扭 94 xx '/"
22. 岫訕/晶 36 60 96 xx AR臼UT:自 8 ART工ST ClID!IST H'I1~IC工ST 10 ENGINEER '. 12 t.fATHTh!AT 9 68 44 96 72 - - -.,. -J - - " BIOLOGIS 7 20 ，旬，丘。 28 62 的口 56 58 泣的口"
PHYSICIAι2ι ，仙 50 78 84 泣如 6o xx76xxrJXX , , OS'l'EOPA'r 2 58 80 xx 94 xx 78 xx xx xx xx "xx 82 凹，---J..- 一-一- - -/ PSYCHIAT 5 22 ,'48
42 66 印， 14 y.x 70 90,5cx xx xx LIR.'iAR工A / ih5880 哎， 46 82 60 60,
/;扭扭"
30 28 3紗，泣， /44 白血與"扭扭扭扭扭-....-,
29 峙，恥 '74 80 72 33 3~ '72 64 /"
32 50 76 扭扭/扭扭扭 98 扭扭扭扭 PSYCHOIρ6 xx xx M工NISTFÆ 扭扭 扭扭 扭扭 扭扭 扭扭 /扭扭口//扭
扭/
門 u no MUSIC 'I'E MUSIC工AN COMPUTER MATH AC 工 AlR FORC Al'..r-IY OFF PRODl氾TI PRES工D血f HORTIC工A REA.L ESI' SALES HA LIFE INS S聞工OR C ACCOUNTA OFFICE W PURCHAS工 IW丑CER PHYS工CAL AUTHOR / -....- - :-，-~-47 2年仙，'7于扭扭扭扭扭扭扭扭扭 SOCIAL W/
,
均鉤，但/立扭扭扭扭扭扭扭扭扭 REHI四日 ADVERTIS f / , IAWYER ι6 峙"扭扭扭扭扭扭 xx 扭扭扭扭 SCROOL S CPAαINE 3ι 扭扭扭扭扭扭扭 94 86 xx 扭扭 PUDLIC A 。目α.1ETR 55 扭扭扭扭扭 xx xx xx xx xx xx xx PERSON祖 - - 7-7 一一一，一一一/一一-一一-一一- --CARPEN咽屆時，它/ '42 熱切旬，76 58 80 82 52 96 CREDIT M / FARMER 13 3ν38 6已 "34 7、加 76 96 90 58 82 92 BUSINESS /"
/ PRI盯間的自~ /72 80" '96 70 扭扭 76 92 86 96 s∞工AL S / FOHEST S 17 峙 /46 80/ 拇指 74 臼 62 xx98
xx xx CHAMBER / / POLICEMA 21 50 52/' xx 74 58 50 46 xx 62 84 8298
CC棚切口T VETEHlNA 3 3~ /江偵 96 92 66 86 56 70 帥的 9扯到由自C , PITARMACI 40 口 80 82 88 58 66 36 62 56 74 72 58 52 峙 "32 52 ~，'扭扭江扭扭 58 扭扭 / l9§612-gqrD二 xx xx xx xx 62 78 位 96 巧品， '40 ，'94 xx 江 98 xx 插曲 711. 72 泣 的話，如 96 xx 86 xx 52 58 臼 61~ 8h 64 13 臼 70 86 臼 xx 48 38 50 的 50 88 xx 岫 58 岫 36 “//78 -56 -æ
44- -;丸，斗;口 xx 一-- -
-一-/ 41 26 28 30 / 90 6斗斗2 36 30. /92 98 90 82 ---,--/, / 43 14/16/90 70 51~ h6 40, /xx 訕 90 82 84 / / 42 2宅 /78 52 lj2 38ιo /xx 90 82 74“
48"
/ 仙 xx 90 70 76 60/ xx 66 品 56 74 58 54 一 ---7 35l836MKE/扭扭扭扭扭扭臼 72 / / 36lf 心4 28/ 扭扭扭扭的鴨山訕 86 / 37 24 2今 96 xx96
96 84 70 30 34 58 10 / 38 1吃/扭扭扭扭阱 82 56 6斗 90 的路 39 xx 玖玖玖 96 訕訕訕 66 86 9696
只 36 66 92 74 扭曲 70 80 82 46 46 xx / / 26 10' 26,
34 岫/ 62 58 38 30 32 3已/也 xx " / / / 訕訕，'30 3♀/恬岫 36 28 30 30 /玖玖玖 / / / 283033/565hk236 峙峙/扭扭扭扭 / 23 比 /38 46 58 26 30 1咚扭扭扭扭扭 / 且 30 岫 52 包括峙，包扭扭扭扭扭 - - - r , -/ 峙泣 /36/22 30 34/ xx 扭扭扭扭扭扭 / ~..." -, / 53 26, /玖玖 3~ 扭扭扭扭扭扭扭扭 / 2736323Ol b 扭扭扭扭扭扭扭扭 h呵- -5016 泊/么扭扭扭扭扭扭扭扭扭
- - r 6/扭扭扭扭扭扭扭扭扭扭扭 / 54 25 扭扭扭扭扭扭扭扭扭扭扭扭Cluster Analysis of SVIB Protìle Patterns 5
Adu1t women and college men are not only much alike in the variability between different occupational groups; their pattern of fiuctuation between different scales within the same occupational groups is also similar. In Group 1, both scored high on Psychiatrist and Psychologist, and low on Veteranarian scales; in Group 11, high on Architect and low on Physicist and Engineer scales; in Group IV, higher on Printer and Farmer and lower on Forest Service Man; and Group VIII, higher on Mortician and lower on Accountant.
3. Cluster analysis of adult women's and college men's data are shown on Chart 2 and Chart 3. The average of all correlations between members of each and all of the occupational c1usters are listed in Table 2. A comparison of adu1t women's and college men's occupational c1usters with that of the SVIB's c1assification is presented in Table 3.
Explanation of the cIusters on Chart 2 and Chart 3
The c1uster consists of a distance matrix intersected by lines representing a tree. (Hartigan, 1967) A sample tree found in adult women is:
Author-Journalis七 (47) 37/τr3-7--J -', / / J Adver七 ising m叫1 (45 ) /38 ,/ / (46) /
r.a
vycr也---The number 37 means that the distance between Author-J ourna1ist and Adver-tising man and Lawyer is 37, corresponding to a correlation 1-.37 = .63. The number 38 means that the distance between Advertising man and Lawyer is 38, correspoding to a correlation 1-.38 = .62. A similar tree is found in college men's data:
h'1.wyer (47) (句) (46) 一一-7 - - - -" 24/-2拉/ ,/ ,/ / /28 / / / Au七hor-Joul"l1alis七 Ad-v'ertis 勾18m也1
.::_---Accordingly, the distance between Author-Journalist and Advertising man and Lawyer is .24, corresponding to a correlation of .76; and the distance between Advertising man and the Lawyer is .28, corresponding to a correlation of .72.
The c1uster consists of a set of scales which may be reached by moving left on dashed lines from nodes (intersections of dashed lines). The c1uster here contains three variables (occupational scales) 47，峙， and 46, and it may be defined by the boundaries once the order of variables in specified.
Order of Variables 吋 tFhdnb A 佳 A 生 A 生 Other Boundary 49 46 47 Cluster 47, 45, 46 45, 46 46, 45, 47
6 心血與教育
Table 2. Comparison of SVIB Occupational Scale-Grouping
V 22. Personnel director 23. Public administrater 24. Rehabilation counselor I 25. YMCA secretary 26. Social worker 27. Social Sci. teacher 28. School superintedent 29. Minister Vl I 30. Librarian ! 31. Artist I 32. Music performer 33. Music teacher vll I 34. CP A owner 油( X I
XII
沮| lX 35. Senior CP A 36. Accountant 37. Officer worker 38. Purchasing agent 39. Banker 40. Pharmacist 41. Mortician 42. Sales manager43. Real estate salesman
44. Life ins. salesman 45. Advertising man 46. Lawyer 47. Author-journalist 48. Presid 巴nt MFG 49. Credit manager
50. Chamber o[ Com. exec.
51. PhysiC3.1 therapist
52. Computer programmer
53. Business Ed. Te.
54. Community Rec. admin.!
* ... ~same as
I
249
Clusters (Male College Students) T 'l V Psychiatrist 'l Psychologist Vl Minister T -G -Rλ -M
一羽
AU -M T-u………
T…
M-rMH. -1n-e]-m
仙一泣
-EU • rtAM --K 沿 AUA -dnp 口 -PPCP心……一卜……
T
…
mmwmIXI
CPA ownerT
x
lXYMCA sxretary YMCA secret 口 ry
T 'l Social Sci. Te.
Cluster Analysis of SVIB Profile Patterns 7
The construction algorithm is such that the first variable always appears first in the tree. The distance in the matrix should increase from left to right, and should be approximately equal in the parallelepipeds bounded by the dashed lines.
By visual inspection, the size of intra.c1uster correlation ranges from .45 to .88,
with most correlations scattering around .50. For adult women, the highest cor. relation is .88, in a single small c1uster formed by Psychiatrist, Psychologist and Minister. For college men, the highest correlation is found in the c1uster tentatively labeled Community Action, consisting of Business Education Teacher, Social Science Teacher, Chamber of Commerce, Community Recreation and YMCA secretary.
The c1uster groupings obtained by the c1uster analysis method seem to demon. strate that adult women and college men do have a common framework of reference in responding to di芷erential interests of men of various occupations. Artists' interests are seen to be similar to those of the Architect's, changed from Group V to Group 1. Neither adult women or col1ege men c1early di 芷erentiate between life science and physical science, though together the sciences might be subdivided into several small c1usters. The Computer programmer and Physical Therapist join in with SVIB Group III, probably having in common regimentation, precision and contro1. The Veteranarian's interests are not associated with those who are involved with life science, Group 1; nor with the Pharmacist, the Business.detail, Group VIII. Instead,
they are in c10ser association with the interests of Carpenter, Farmer, or Printer,
Group IV. SVIB's Group V remains almost the same, except that YMCA secretary was c1ustered, instead, with Credit Manager, Chamber of Commerce, Community Recreation Director-a group of community action men. Psychiatrist and Psycholo司
gist, unfortunately, are not c1ustered in life science. For women, they are definitely associated with the Ministers; and for meny somewhat c10ser to the interests of the Minister and Librarians. Both women and men do not c1ear1y di旺erentiate
between Business Contact and Business Detail as SVIB has suggested, though women perceived Senior CP A, Accountant, Policeman, and Production manager as having something in common, forming a separate c1uster. SVIB's Group X, Linguistic remains a cohesive c1uster, joined by CPA owner, however, accordi
DISCUSSION
The above findings suggest some possibiliities and also raise some questions regarding the application of SVIB.M to women.
First, the di旺 erences of the scale scores between adult women and col1ege men remind the users of SVIB to explore more fully the variables that might infiuence the scores of the occupational scales. In his 1966 revision of SVIB, Campbell (1966) has noticed that his recent1y tested samples score higher on recent1y developed
8 心理與教育
scales, no matter what the occupation is. Wi11iam, Kirk, and Frank (1968), on the other hand, found discrepancies between scores obtained on the new form as com-pared with the old form and the tendency for the scores obtained on the new form to be lower than the old form, in magnitude and location. No adequate explanation has been given for those differences, although it appears that variables such as form (old or new), sex, or sampling population could all operate to produce significant differences between scale scores under different circumstances.
Second, the similarities and differences of the interest profiles between the two groups in this study pose an interesting question. In general, the adult women indicated that they are interested in helping people (Group V, Uplift) through verbal and persuasive communication (Group V, Linguistic), and in a professional (group 1,
Professional Scientists) and cultural setting (Group VI, Musician) to function more effectively in the world of personal service (Group IX, Business Contact) and inter-personal relations (Group XI). With a .80 correlation of coefficient, the interest profile of the college men looks rather similar to that of the adult women. Could we infer the career direction of those college men seeking career counseling to be similar to that of the adult women?
However, the women's profile is sharper. According to Strong, a sharper profile was interpreted as having c1earer direction. For instance, the engineer or other professional school students in his sample had sharper profiles than did the business majors, because the former supposedly had little misgivings about their choice of a field and less difficulty in finding themselves. A recent study by Herkenhogg, who compared older and younger women, also revealed that women thirty years or older have more c1early defined interests in SVIB. Should we then, imply that the adult women have stronger preferences regarding what they wi11 and will not enjoy doing? Or, though inexperienced in career and employment have the adult women developed more definite interest patterns? Or, as a group, the adu1t women are more homogeneous, than the younger females and college men?
Third, the meaning of the occupational groupings needs to be further examined. Thurstone, Guilford and others have shown that four or five factors (or groups) are sufficient to account mathematically for all or nearly the variation in interest
Cluster Analysis of SVIB Profile Patterns 9
in responding to differential interests of men in various occupations. Al1 in al1,
the comparison of interest scales, interest profiles, and interest clusters points to
the possibility and feasibility of integrating two separate SVIB blanks into one
,
inthe near future.
REFERENCES
( 1) CAMPBELL, D. P.: “The 1966 Rεvision of the Strong Vocation31 InteresL Blank." Personnel
& GuidancεJournal ， 1966, 4<1, 7在 4-747.
(2) CAMPBELL, D. P.: Manual for SVfB for Men & Women, 1968.
( 3) HERKENHOFF, L. A.:“A Comparison of Older & Youngεr Women Smdents at San Jose
City Collegεwith Implication for Curriculum & Stud 己nt Personnel Service."
( 4) HARTIGAN, J. A.:“Present且 tion of Simil 且rity Metrices by Tree." Journal of American
Statistical Association, 1967.
( 5) LAIMA, B.: ‘Women's Score on the M & F Forms of the SVIB". Vocational Guidance
12: 116-8, Winter 63-64.
( 6) SOKOL, R. R. & SNEATH, P. A.: Principles of Numerical Taxin()my. San Francisco: W. H.
Fresdman & Co. 1963.
(7) STRONG, E. K.: Vocalional fnleresls of Men & Women, Stanford, California: Stanford
University Press. 1943.
(8) STRONG, E. K.: Vocational fnle何 sls 18 Years Afler College. Minneopolis: University of
Minnes~ta Press, 1955.
( 9) WILLIAM, P. A., KlRK, B. A. FRANK, A. C.: “New Men's SVIB: A Comparison with the
10
