Transactional Analysis Journal

http://tax.sagepub.com/

Change in Two Settings

David Steere, Grayson Tucker and Ann Worth
Transactional Analysis Journal 1981 11: 222
DOI: 10.1177/036215378101100306

The online version of this article can be found at:

http://tax.sagepub.com/content/11/3/222

Published by:

http://www.sagepublications.com

On behalf of:

International Transactional Analysis Association

Additional services and information for Transactional Analysis Journal can be found at:

Email Alerts: http://tax.sagepub.com/cgi/alerts

Subscriptions: http://tax.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://tax.sagepub.com/content/11/3/222.refs.html

>> Version of Record - Jul 1, 1981

What is This?

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

 Change in Two Settings
David Steere
Grayson Tucker
Ann Worth

Abstract participants in a weekend marathon

Two groups of psycbotberapy patients
conducted by Bob and Mary Goulding. No
were compared as to tbe relative effective-
comparison was made to similar
ness of a 12-session on-going weekly therapy procedures employed in a regularly-spaced
time format.
group and an 1'-bour maratbon tberapy
group. The CatteD Sixteen Personality Purpose
Factor Questionnaire was administered pre-
and post-treatment to evaluate effects on
This study examined the impact of parti-
underlying personality structure, and a
cipation in time-extended (marathon)
groups and regularly-spaced (weekly)
Goal Attainment SCale utiUzed to report
bebavioral cbanges. Tbe bypotbesis tbat
groups using transactional analysis (TA)
the maratbon group would sbow greater
led by the same psychotherapist, David
Steere. Bach (1966) defined a marathon as
positive cbange initially was supported, as
meeting 12-24hours with or without pause.
was the bypotbesis tbat tbe weekly group
would evidence greater positive cbange In
Traditional weekly groups meet for one to
two hours once or twice during each week.
tbe latter two-thirds of tbe treatment
Steere's marathons meet for 18 hours
period. A bypotbesis tbat maratbon
distributed over a weekend from Friday
participants would sbow a moderate loss
evening to Sunday afternoon with normal
of an initially strong cbange toward
rest and sleep. His weekly groups meet for
bealtb was not confirmed.
one and one half hours each week. A 12
week period was selected for comparison
Debate continues over the relative with the marathon to strike the rough
effectiveness of a marathon or a weekly equivalent of 18 hours treatment in each
group to produce personal change. Some setting.
data suggest the marathon group may There were three hypotheses: (a) the
result in short-term gains but less durable marathon group would measure greater
change in the long run (Gibb, 1971). Others positive change initially (one week after the
claim the intensive character of such an experience) than the weekly group (after
experience can produce more personal four weeks of treatment); (b) the weekly
change in a weekend than months, even group would measure greater positive
years, of space diluted meetings (Lieber- change between Weeks 4 and 12; and (c)
man, Yalom, & Miles, 1973). King, Payne, during the period between the first posttest
& McIntire (1973) showed greater change and the final one (at the end of 12 weeks)
taking place among participants in positive change would increase substan-
marathon groups than in prolonged tially for the weekly group and decline for
groups, but refused to conclude that results the marathon participants. The last hypo-
measured immediately following such an thesis which predicted moderate loss of an
experience persisted over a long period of initially strong change toward health by
time. McNeel (1975) has documented marathon participants was the one
changes holding after six months among 17 hypothesis not confirmed by the data.

222 Transactional AnalysisJournal

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

 CHANGE IN TWO SETTINGS

Method A Goal Attainment Scale (GAS) with

SUBJECTS
instructions and practice pages was
designed by the researchers patterned after
Participants were drawn from a pool of a Goal Attainment Follow-up Guide
clients in psychotherapy with David Steere developed at the Hennepin County
in Louisville, Kentucky. A total of 22 Community Mental Health Center in
persons were included in Group 1 Minneapolis, Minnesota. This instrument
(Marathon Group). Twelve participants was used as a self-evaluation measurement
attended a marathon in December, 1974 by subjects to report behavioral changes.
and ten attended a marathon in February, These two instruments were selected in
1975. Group 2 (Ongoing Treatment Group accord with Steere's dual emphasis on
meeting weekly) included a total of 22 contracting for concrete changes in the
persons in their first two months of "here and now" coupled with direct
therapy: nine in Group A, eight in Group methods of exploring the way particular
B, and five in Group C. ego states got formed to facilitate self-
Group 1 included nine males and thirteen understanding and redecision. For example,
females ranging in age from 20 to 50. A, whose contract is to stop withdrawing
Group 2 included nine males and thirteen from conflicts, is encouraged to attend to
females ranging in age from 21 to 46. what happened in specific transactions with
Eleven subjects did not complete the the group when others began to argue and
experiment: six in Group 1 and five in raise their voices (transactional analysis).
Group 2. Some marathon participants lived The internal dialogue within A is explored
out of state and did not respond to the (structural analysis) and optional responses
posttest mailing. Some participants in encouraged. Attention is given to repetitive
Group 2 complained of a lack of time to patterns of withdrawal from friends and
complete the posttest. Treatment of the intimates (game analysis) and immediate
subjects was in accordance with the ethical changes toward self-assertion are sought
standards of the American Psychological and practiced. At the same time, formative
Association, 1973. scenes with parents and siblings where A
EXPERIMENTAL DESIGN (T} X T2' T} Y T2)
learned to behave this way are explored.
Redecision work, as outlined by Goulding
A quasi-experimental design was elected (1972), is undertaken around early scenes
as recommended by Campbell and Stanley which proved determinitive in the life course
(1963) when more vigorous methods such (script analysis).
as randomly assigning subjects to different This dual emphasis devotes equal
treatments are not feasible. There was no attention to immediate changes in specific
control for history or other change- behavior and to reorganizing the
producing events. The groups were, underlying structure of personality. Because
however, found to be roughly equivalent. TA has been criticized by some as
A difference between them on the initial "superficial" or a "head trip" that doesn't
distance from health was observable but tap fundamental emotional trends in the
not statistically significant (See Figure 1). person's life, the research design addressed
Both were similar in male-female ratio and both concerns. The GAS provides a fair
age range. measure at the level of specific behavioral
APPARATUS change and the 16PF provides a view of
underlying personality trends. Steere's use
Form A, 1967, of the Cattell Sixteen of TAwas a constant in both settings,
Personality Factor Questionnaire (16PF) providing comparative data on marathons
was administered to each participant as a and weeklygroups.
pre and posttest measurement of such
personality factors as apprehensiveness, PROCEDURE
tenseness, trustfulness, stability, anxiety, After research participation agreements
assertiveness, etc. and release forms were signed, the 16PF
Vol. 11, No.3, July 1981 223

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

DAVID STEERE, ORAYSON TUCKER AND ANN WORTH

was administered to all participants with the reliability of this process indicated that
the proper instructions. The GAS an estimate of .57 is appropriate (Sherman,
instructions and worksheet were given to 1974, p, 8).
each person for review prior to a scheduled Following the marathon, Group 1
goal-setting session. At this session, the participants were given their GAS, 16PF
researchers assisted participants in setting booklet, and answer sheet. They were
behavioral objectives in two or three areas instructed to complete the self-evaluation
surrounding their change contract. with someone who "knew them well," and
Behavioral patterns were established on a to retake the 16PF one week later, mailing
5-point scale (from - 2 to + 2) for each the material back. Group 2 took the
problem area. The present behavior pattern 16PF and repeated the same procedure
became the - 1 point on the scale. By way with the GAS after the fourth week. At the
of illustration, for A whose contract was to 12 week follow-up both groups took the
stop withdrawing from conflict, - 1 16PF and made their third evaluation with
equalled clamming up and/or leaving the the GAS.
scene when someone began to argue or
"HEALTH" ON THE 16PF
criticize. For the zero point on the scale,
participants were instructed to assume they No norms for "healthy" or "self
found some good ways to work on this actualizing" persons existed for the 16PF.
problem and to estimate what change could A health criterion was developed for each
be reasonably expected after three months of its scales by asking 24 "experts" to
(for A, to continue to talk with respond to the instrument in terms of their
argumentative persons for five minutes). conception of how a "healthy" or "well
The + 1 level on the scale reflected a integrated" person would respond.
pattern somewhat more optimistic in Fourteen of the respondents were clergy
outcome (A would argue back and assert with special training in pastoral counseling,
his own point of view for five minutes). two were social workers, five
The + 2 level was defined by describing psychologists, and three professors of
behavioral change with the most favorable theology. They were asked to omit Scale B
outcome likely (A would assert his own items which provide a rough IQ measure
point of view without having to leave the and to mark all items judged irrelevant to
scene of an argument emotionally or describing the healthy individual by "in
physically). Finally, the - 2 level described between" or "uncertain" responses which
a behavior pattern projected if things got are always "b" in the booklet.
worse (A would begin avoiding friends and The means of the 24 scores were then
intimates anticipating conflict). used to define a health norm for males and
With the GAS it was possible to females on each of the 15 remaining scales
designate one scale to be weighed of the 16PF. The reliability of the health
differently than the others, indicating criterion was assessed by the STET pattern
behavioral change of greater importance. similarity coefficient (Cattell, Ever, &
Subjects could also mark a point midway Tatsuoka, 1970, p. 137, 307) for the 276
between the five behavioral descriptions, pairwise combinations of the 24
permitting response on a 9-point scale. The respondents, taking the median of these co-
formula used for assigning numerical efficients and determining its probability of
values to a person's profile was developed occurence. An estimate of this, employing
by Kiresuk and Sherman (1968, p, 449): well over half of the combinations, was .23
with .26 required for significance at the .10
lO:EwjXi level. Although not used in this study, a
50+ more accurate health criterion was estab-
V .7 :Ew? + .3 (:Ewj}2 lished by eliminating six respondents who
where Xi is the level indicated on a were extreme on five or more scales. This
particular scale and Wi is the weight provided a median coefficient of
assigned that scale. Preliminary studies of approximately .33 which is significant at

224 Transactional Analysis Journal

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

 CHANGE IN TWO SETTINGS

Male Criteria Female Criteria

Standard Sten Standard Sten
Cattell Scale Mean Error Range Mean Error Range
A(Reserved vs. Easygoing) 6.50 1.61 4-10 6.12 1.65 4·10
C(Easily Upset vs. Stable) 6.66 1.85 3-10 7.04 1.96 3·10
E(Submissive vs. Dominant) 7.66 1.52 5-10 7.87 1.36 6-10
F(Serious vs. Happy-go-lucky) 6.75 1.39 3-9 7.29 1.54 4-10
G(Expedient vs. Conscientious) 3.41 1.44 1-7 3.. 50 1.56 1-7
H(Timid vs. Venturesome) 7.45 1.84 3-10 7.87 1.62 4-10
I (Tough Minded vs. Tender Minded) 7.45 1.38 5-10 4.91 1.44 3-8
L (Trusting vs. Suspicious) 5.04 2.11 2-9 5.70 1.98 3-10
M(Practical vs. Imaginative) 7.58 1.21 6-10 7.37 1.43 5-10
N(Clumsy vs. Polished) 4.62 1.55 1-8 3.66 1.46 1-7
o(Self-assured vs. Apprehensive)
°° 1 (Conservative vs. Experimenting)
2 (Other-directed vs. Self-sufficient)
4.04
5.87
5.83
1.98
1.89
1.63
1-7
2-9
4-10
3.54
6.95
5.83
1.88
1.96
t.63
1-7
3-10
4-10
03(Undisciplined vs. Disciplined) 3.83 1.83 1-8 4.33 2.05 1-8
04(Relaxed vs. Tense) 4.66 2.07 1-10 3.66 2.18 1-9
ANXIETY 4.23 1.99 1.0-9.1 3.50 1.95 1.0-8.6
NEUROTIC 4.39 1.80 1.9-7.4 2.99 1.78 1.0-6.8

Table 1
Health Criteria Established by 24 Respondents

Marathon Group Marathon Group Marathon Group

Scale Time 1-2 Time 2·3 Time 1-3
A - .21 - .12 .37 .32 .11 .18
C .41 .39 .99 .11 1.35 .54
E -.01 -.24 .27 .59 .26 .35
F - .16 - .09 .25 .32 .10 .22
G .50 .07 .31 -.35 .69 -.16
H .19 .28 .33 .14 .54 .41
I -.17 .07 - .44 .34 -.58 .41
L .03 .35 .68 .47 .73 .80
M .68 .11 o -.15 .54 .11
N - .29 .09 .17 .36 .05 .28
o 1.02 .15 .71 .26 1.30 .71
°1 .14
.38
- .12
.17
- .51
- .16
-.72
.19
- .45
.20
-.77
.38
°2°3 .55 -.17 -.35 -.06 .10 -.08
1.10 .46 -.06 .50 .86 1.09
°4
ANXIETY .83 .31 .56 .40 1.26 .83
NEUROTIC .31 .15 .21 .32 .47 .53
SUM A-0 4 4.30 1.20 2.64 2.34 6.22 4.25

Table 2
Adjusted Mean Changes Toward Health In Cattell Sten Scores

Vol. II, No.3, July 1981 225

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

DAVID STEERE, GRAYSON TUCKER AND ANN WORTH

the .05 level. (This more accurate criterion another, also have varying intercorrela-
is available upon request.) tions (Cattell et aI., 1970). Therefore,
The computer printout of the 16PF also analyses of variance were run to check for
provides scores on selected second-order significance in change toward health for
factors which are linear combinations of the 15 means as a group. Results of this
the basic scales. Two of these, measures of analysis are provided in Table 3.
anxiety and neuroticism, were used in the The pattern of change in Groups 1 and 2
experiment. The complete health criteria may be seen graphically in Figure 1
are given in Table 1. This enabled us to depicting the mean sum of distances from
compute a distance from health on 17 health on the 15 primary source traits of
scales and a sum of distances from health the 16PF. This presents a general picture of
on 15 basic scales for each subject at each the findings. An initial strong measure of
testing. These distances from health were change among marathon participants held
made positive numbers by taking the with slight increase over the three month
absolute value of each difference between a perod. A more steady pattern of change
subject's score and that of the criterion. toward health is observed among weekly
Movement toward or away from the health group members with the final posttest at 12
norm could then be determined for each weeks. The third hypothesis, which pre-
participant over the designated time dicted a moderate loss of initial gains in a
intervals. "post-marathon high" over the same
Results period, proved false.
Mean levels for Group 1 and Group 2
Analysis of variance and t-tests using an participants using the GAS are given in
adjusted change toward health in the Table 4. The results have a pattern similar
different time periods were applied to the to the Cattell data. Marathon participants
Cattell data. Mean changes toward health reported better achievement than group
on the different scales were adjusted to participants of individually set goals in the
compensate for the varying effect of period Time 1 to Time 2. Comparisons for
regression toward the mean due to the Time 2 to 3 and 1 to 3 were not significant.
correlaton of change scores with initial In both marathon and regular group
distance from health. The formula used for treatment, changes were different from
this adjustment from Blalock (1972, p. 482) zero at .005 or better for each of the three
was: time periods.
tl -- = Further analysis of the data revealed the
1\
y = _y - b (X - x) where following personal changes in each treat-
Y= adjusted mean change for a given scale ment modality and comparisons between
y = mean change from the same scale taken from the them that were significant at the .05 levelor
data better using the one-tailed t test:
b= mean slope for Group 1 and Group 2 of regression
I. Over the entire period, members of
of change score on initial distance from health
x = mean initial distance from health in Group 1 or both Group 1 and Group 2 exhibited
_ in Group 2 change toward health on 16PF scales C
x = mean initial distance from health for both Group (Easily Upset vs. Calm, Stable), L
1 and Group 2.
(Trusting vs. Suspicious), 0 (Self-Assured
The adjusted mean changes toward vs. Apprehensive), Q4 (Relaxed vs. Tense,
health among marathon participants Driven), Anxiety, and the sum of scales A
(Group I) and weekly group participants through Q4' Group 1 showed more change
(Group 2) are given in Table 2. One may on Scale G (Expedient vs. Conscientious)
expect one out of 20 t-tests to show signifi- and Scale C (Easily Upset vs. Calm, Stable)
cance at the .05 level simply by chance. than Group 2. On two scales, one for each
Further, the 15 scales (A through Q~ of group, changes were recorded that moved
primary source traits, while constructed to away from the health criterion: Scale I
show a degree of independence from one (Tough Minded vs. Tender Minded) for
226 Transactional Analysis Journal

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

 CHANGE IN TWO SETTINGS

Time 1-2 Time 2-3 Time 1-3

F-Ratio Sig. Level F-Ratio Sig. Level F-Ratio Sig. Level
Group Compared 2.86 N.S. 2.93 N.S. 6.93 .02
to no-change
Marathon Compared 6.06 .03 2.33 N.S. 7.33 .01
to no-change
Marathon Compared 2.15 N.S. .01 N.S. .23 N.S .

Table 3
Analyses of Variance for 15Adjusted Means

Group 1, and for Group 2, Scale Q I

(Conservative vs. Experimenting). Health Score Time 1 Time 2 Time 3
2. From Time 1 to Time 2 (one week
after the marathon for Group 1 and four 25
, ..
weeks later for Group 2), persons in Group 26 ..... »> .... 25.16
1 exhibited change toward health on Scales
27 , 26.15
C (Easily Upset vs. Calm, Stable), M /
(Practical vs, Imaginative), 0 (Self- 28 /
/

,,
/
Assured vs. Apprehensive), Q3
(Undisiciplined vs. Disciplined), Q4
29
~
,.
(Relaxed vs. Tense, Driven), Anxiety, and 30 29.98 29.28
the sum of Scales A through Q4' Those in 31
Group 2 showed significant change on
none of the scales for this period. 32
3. From Time 2 to Time 3, Group 1 33
participants showed change on Scales C
(Easily Upset vs. Calm, Stable) and L 34
(Trusting vs, Suspicious). Group 2 35 34.85
participants showed change on Scales C
Marathon
(Submissive vs, Dominant), L (Trusting vs.
Suspicious), and the sum of Scales A Weekly Group - - -
through Q4' Figure 1
Discussion Comparison of Marathon and Group
Mean Sums Distances from Health
Comparable personal change at both the
behavioral level and the level of underlying Scale A-Q4
personality structure does occur in TA
groups whether treatment hours are massed period in weekly group meetings. Neither
in one weekend as in the marathon format setting is inherently superior to the other.
or regularly spaced over a three month Our apparent success in establishing
workable health norms on the Cattell 16PF
needs further investigation and refinement.
Time 1 Time 2 Time 3 The fact that we could establish them
MARATHON 37.00 51.20 59.22 through experts positing answers for the
GROUP 37.00 46.31 56.70 healthy individual was surprising. The fact
that we could record statistically significant
Table 4 changes in the direction of these established
Goal Attainment Scale Means norms among our subjects in treatment was
significant.
Vol. 11, No.3, July 1981 227

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014

DAVID STEERE, GRAYSON TUCKER AND ANN WORTH

The main purpose of this study was to David Steere, PhD, CTM, is currently
compare results in the two temporal Professor of Pastoral Care and Counseling
formats employed. Whereas a cursory at Louisville Presbyterian Theological
glance at Figure 1 could suggest that the Seminary.
marathon is superior to weekly group Professor Grayson Tucker, STD, is
treatment, the data do not establish this. Dean of Louisville Presbyterian Theologi-
The experiment did, however, add further cal Seminary and a Fellow with the
weight against the argument that marathon American Association of Pastoral
gains are short lived. Counselors.
Each treatment setting appears to have Ann Worth, MA, is a psychologist and a
its own unique advantages. In a marathon, PhD student in psychology at East Texas
successivehours create a "building effect" State University.
from one person's work to another's. The
disengagement and withdrawal from life REFERENCES
Bach, G.R. The marathon group: Intensive practice
has many of the focusing qualities of the of intimate interaction. Psychological Reports,
retreat practiced from time immemorial in 1966, 18, 995-1002.
religious communities, providing a pivotal Blalock, H.M. Social statistics. New York: McGraw
setting for redecision and change. Hill,1972.
Psychotherapists who restrict themselves to Campbell, D.T., & Stanley, J.D. Experimental and
traditional time formats ignore an quasi-experimental designsfor research. New York:
established and powerful medium for Rand McNally and company, 1963. (Reprinted
from Handbook of Research on Teaching. New
change. York: American Educational Research Associa-
On the other hand, those therapists who tion, Rand McNalleyand Company, 1963.)
work only in the marathon setting may Cattell, R.B., Eber, H.W., & Tatsuoka, M.M. Hand-
miss the long-range picture of treatment book for the sixteen PF. Champaign: Institute for
Personality and Ability Testing, 1970.
which invades regularly the daily routines
Gibb, J.R. The effects of human relations training.
of self-defeating behavior. The rhythm of Handbook ofPsychotherapy and Behavior Change.
following clients week to week keeps the New York: John Wiley, 1971,839-862.
therapist close to the locus of realistic Goulding, R., M.D. New directions in transactional
change. All of our participants in regular analysis: Creating an environment for redecision
groups were in treatment for more than the and change. In Sager, C.J., & H.S. Kaplan, (Eds.)
three month span of this study. The Progress in Group and Family Theapy. New York:
Brunner/Mazel, 1972, 105-134.
average "stay" in Steere's groups is 8.2 King, M.E., Payne, D.C., & Mclntire, W.G. The im-
months. We have every reason to believe pact of marathon and prolonged sensitivity training
that change measured over that period of on self-acceptance. Small Group Behavior, 1973,
time would exceed the changes registered 4(4),414-423.
over this shorter period of research. Kiresuk, J., & Sherman, R.E. Goal attainment scale:
A general method for evaluating comprehensive
In all likelihood, we may find some community mental health programs. Community
combinations of the two temporal formats Mental Health Journal, 1968,4,443-453.
optimal. Marathons provide excellent Lieberman, M.A., /Yalom, I.D., & Miles, M.B.
formative experiences during the early Encounter groups: First facts. New 'York: Basic
stages of group treatment. On several Books, Inc., 1973, 13.
occasions, we have noted rapid McNeel, J .R. Redecisions in psychotherapy: A study
improvement and successful termination of the effects of an intensive weekend group work-
shop. Unpublished Doctoral Dissertation. San
among long-term regular group members San Francisco, The California School of Profes-
upon engaging in a marathon experience sional Psychology, 1975.
after eight to ten months in a weekly setting. Sherman, R.E., Baxter, J.W., & Audette, D.M. An
Whatever other implications this study examination of the reliability of the Kiresuk-Sher-
man oral attainment score by means of components
may have, it points to liberal experimenta- of variance. Minneapolis: Program Evaluation Pro-
tion with different combinations of time ject. (501 Park Avenue South), Report 4, August,
formats among group therapists. 1974.

228 Transactional Analysis Journal

Downloaded from tax.sagepub.com by Ana Farcas on May 22, 2014