ch01_straus_4370.

qxd 2/11/05 9:05 AM Page 507

Violence and Victims, Volume 19, Number 5, October 2004

A Short Form of the Revised Conflict

Tactics Scales, and Typologies for
Severity and Mutuality
Murray A. Straus
Emily M. Douglas
University of New Hampshire
Durham, NH

The revised Conflict Tactics Scales (CTS2) is the most widely used instrument for mea-
suring intimate partner violence. This article presents a short form to enable the CTS2 to
be used when testing time is very limited. It also presents procedures that can be used with
either the full test or the short form to classify individuals on the basis of severity of behav-
ior toward a partner or by a partner, and to classify couples on the basis of mutuality or
symmetry in the behaviors measured by the CTS2. The results indicate that the short form
is comparable in validity to the full CTS2. Although the short form does not identify as
many cases of partner violence as the full scale, it does identify a large number of cases
and if there is insufficient time for the full scale, can be a useful screening instrument.

T
he Conflict Tactics Scales (Straus, Hamby, Boney-McCoy, & Sugarman, 1996) is
the most widely used instrument in research on family violence. The instrument
includes scales to measure three tactics used when there is conflict in the rela-
tionships of dating, cohabiting, or marital couples: Negotiation, Physical Assault, and
Psychological Aggression. In addition, there are two supplemental scales: Injury from
Assault and Sexual Coercion. More than 200 papers and many books reporting results
based on administration of the CTS have been published (for bibliographies see Straus,
2004; Yodanis, Hill, & Straus, 1997), and between 5 and 10 papers are currently being
published each month. The purpose of this article is to further extend the utility of the
CTS by making available a short form of the Revised Conflict Tactics Scales (the CTS2)
and by providing procedures for use with both the short form and the full CTS2 to clas-
sify individuals on the basis of severity of behavior toward a partner or by a partner, and
to classify couples on the basis of mutuality or symmetry in the behaviors measured by
the CTS2.

Need for a Short Form and Typologies

Short Forms. The CTS2 short form (hereafter, the CTS2S) was created in response to
many requests since the CTS2 was introduced because the CTS2 is much longer than the
original CTS. The full CTS2 consists of 39 items, each of which is first asked for the
behavior of the respondent and then repeated for the behavior of the respondent’s partner,
making a total of 78 questions. The test administration time of 10 to 15 minutes is not a

© 2004 Springer Publishing Company 507

ch01_straus_4370.qxd 2/11/05 9:05 AM Page 508

508 Straus and Douglas

problem in many clinical and research applications, but there are also many applications
where this amount of time is not available. For example, most survey research is now con-
ducted by telephone. The recommended maximum length for a phone interview is 30 min-
utes. If, as is usually the case, many other aspects of family relationships must be measured,
it is not possible to devote a third to half of the interview time to the variables measured
by the CTS2. The research reported in this article was undertaken to create and evaluate a
20-question short form which would take approximately 3 minutes to administer.
Severity Level. The severity of aggression against a partner is a critical issue that needs
to be addressed in research on family violence. The CTS2 has separate subscales for minor
and severe levels of the Physical Assault, Injury, Psychological Aggression, and Sexual
Coercion scales. However, the minor subscale scores are confounded with the severe sub-
scale scores because almost everyone who engages in the more severe behavior also engages
in the less severe behavior. Partners who kick or punch are also likely to slap and shove.
The Severity Levels measure described in this article avoids this problem by classifying
respondents into three mutually exclusive types: none, minor only, and severe. These cat-
egories can be used as the dependent variable in multinomial logistic regression or dis-
criminate analysis. The severity level variable can also be used as a three-category ordinal
measure of each of the behaviors measured by the CTS2.
Mutuality Types. Another critical issue in measuring partner violence is the mutuality
of the abusive behavior. The CTS2 obtains data on the use of physical and psychological
aggression by both parties in a relationship. However, until now there has not been a stan-
dardized procedure to take into account the mutuality of violence. This article describes a
method of doing this by creating mutuality types. The need for such a typology is based
on the assumption that most research and clinical work on partner violence will benefit
from taking into account the behavior of both partners in a relationship. This applies even
when it might seem that only information on the behavior of one of the partners is needed,
such as when the CTS2 is used to measure progress in a treatment program for male bat-
terers. Research has shown that the cessation of violence by one partner is highly depen-
dent on whether the other partner also stops hitting (Feld & Straus, 1989; Gelles & Straus,
1988). Thus, when monitoring a treatment program, it is crucial to know the extent to
which the partner has also ceased acts of psychological and physical aggression. The
mutuality types provide one way to investigate this issue. They obtain the scores for each
partner to classify couples into three categories: male partner only, female partner only,
and both aggressive.

METHOD

Sample
The sample consists of students enrolled in introductory sociology and psychology courses
at a New England university in 1998, 1999, and 2000. The procedures to ensure informed
consent, privacy, and safety were reviewed and approved by the Institutional Review Board
of the University of New Hampshire. The questionnaires were distributed in class to all
students present. They were told that the questionnaire asked about beliefs and experiences
they may have had in a dating relationship. They were informed that the questionnaire
included questions on sensitive issues such as sex. The written and oral instructions
emphasized that answering the questionnaire was entirely voluntary, and that they could
also skip any question that they did not wish to answer. The students were instructed to put
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 509

Short Form of CTS2 509

their completed questionnaires in a slot in a box near the exit door and leave. Those who
did not wish to participate put a blank questionnaire in the box and were indistinguishable
from students who completed the questionnaire. Only 1 or 2 students out of each class of
50 to 80 students did not complete the questionnaire.
The questionnaire included the full CTS2 scale (not a short form) and a number of other
scales in addition to the CTS2. Most students completed the questionnaire in 40 to 45 min-
utes. Questionnaires for students who had not been in a dating relationship of at least
1-month duration in the past 12 months were excluded from the analyses reported in this
article. Married students were not included because there were not enough in these groups
to analyze separately. The resulting sample consisted of 1,157 cases (347 males and 810
females).

Procedure to Create the CTS2S

The following steps were followed to create the short form of the CTS2 (see the
Appendices for sample items).
1. For each of the five scales, 2 items were selected from the full scale. One of the 2 items was
chosen from the severe behavior subscale, and the other from the less severe behavior sub-
scale. (For the Negotiation scale, 1 item was chosen from the Cognitive subscale and the
other from the Emotional subscale.) When these 10 items are repeated for behavior of the
respondent and the partner, it results in an instrument with 20 questions, compared to 78
questions in the CTS2.
2. The specific minor and severe items were selected by computing the correlation of the items
in each scale with the total scale score. The item with the highest correlation with the total
scale score was chosen. When there were items with approximately equal correlations, the
item with the higher prevalence rate was chosen.
3. The CTS2 items ask about only one specific act, such as punching, which is an item in the
Severe Assault scale. This is an appropriate strategy for an instrument that uses multiple
items. However, when there is only one item to measure severe assaults, it will underesti-
mate the prevalence of this behavior. Consequently, the items selected by steps 1 and 2 were
augmented by adding behaviors measured by the unselected items in the scale. For example,
the Severe Assault scale item “Beat up my partner” was changed to make it “Punched or
kicked or beat up my partner.” It is usually desirable to avoid double-barreled questions like
this because it is impossible to know which of the behaviors in the question occurred.
However, in this context, the need to tap as many aspects of the domain as possible took
precedence.

There are several ways to score the CTS2, including annual prevalence rate, “ever
prevalence,” and annual chronicity for those who experienced partner violence (Straus,
2000a; Straus et al., 1996). The annual prevalence method of scoring was used for four
of the five scales, and the yearly frequency method was used for the Negotiation scale
because these are the recommended and most widely used methods of scoring these CTS
scales (see section on “Scoring” in Appendix B).

Procedure to Create Severity Level Variable

Scores on the Minor Assault scale of the CTS2 overlap with Severe Assaults because peo-
ple who severely attack a partner almost always also engage in less severe attacks. The sever-
ity level variable deals with this problem by classifying respondents into three mutually
exclusive categories: 0 = no violence, 1 = minor only, and 2 = severe. The respondents in
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 510

510 Straus and Douglas

the minor only category are those who reported one or more acts of minor violence and no
instances of severe violence. The severity level variable can be used as a three-category
typology or as a 3-level ordinal scale.

Procedure to Create Mutuality Types

The mutuality types can be created because the CTS2 repeats each item for the behavior
of the respondent toward the partner and for the behavior of the partner toward the respon-
dent. Because the CTS2 measures behavior, not attitudes or motives, the respondent has
full knowledge of the partner’s behavior. This permits classifying couples into three mutu-
ality types even though only one partner is the respondent. For example, the three mutual-
ity types using the Physical Assault scale are 1 = male partner only, 2 = female partner
only, 3 = both. Because this typology is intended to aid research on the dynamics of mal-
treatment, it is not scored if there was no maltreatment of the type measured by a CTS2
scale. Therefore, among a clinical sample such as men in a batterer treatment program or
women in a shelter, all the subjects are classified according to the mutuality of assault, but
for a nonclinical sample the typology for physical assault is scored only for the relatively
small part of the sample in which a physical assault occurred.

Other Measures
Risk Factors For Partner Violence. The questionnaire included scales from the
Personal And Relationships Profile (Straus, Hamby, Boney-McCoy, & Sugarman, 1999;
Straus & Mouradian, 1999) to measure risk factors for partner violence. The five scales
used for this study, with their alpha coefficients of reliability, and an example of an item
from each scale are listed below. Additional psychometric data for these scales can be
found in Straus and Mouradian.
Anger Management (.66): “When I feel myself getting angry at my partner, I try to tell myself to
calm down”
Couple Conflict (.82): “My partner and I disagree about my friends and family”
Criminal History (.83): “Since age 15, I stole or tried to steal something worth more than $50.00”
Negative Attribution (.74): “When my partner is nice to me I wonder what my partner wants”
Violence Approval (.72): “A man should not walk away from a physical fight with another man”
Socioeconomic Status (SES). An SES scale was computed by summing items that
measure the education of the student’s father and mother (score range 1–7 for both vari-
ables) and the parents’ combined income (score range 1–9). The resulting scale has an
alpha coefficient of .68.
Social Desirability Scale. A modification of the Reynolds (1982) 13-item social desir-
ability scale was included in the questionnaire. This scale measures the degree to which
respondents tend to avoid disclosing socially undesirable behavior. The need to control for
social desirability is indicated by the following correlations with the above risk factor vari-
ables: Anger Management r = .49, Couple Conflict r = –.47, Criminal History r = 35,
Negative Attribution r = –.41, and Violence Approval r = –.39.

Data Analysis
Use of An Approximation Version of the CTS2S. A subset of the items from the full
CTS2 was used to approximate the short form given. Therefore, the results presented in
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 511

Short Form of CTS2 511

this article refer to what will be called an “approximation” of the CTS2S, not to the
CTS2S. The reason it is an approximation, as explained previously, is because some of
the items in the short form have been augmented by adding behaviors from full CTS2
items that are not in the short form. Because the CTS2S includes more of the behaviors
in the full CTS2, the construct validity coefficients based on the approximation of the
CTS2S can be thought of as lower-bound estimates. If that is correct, use of the short
form would show stronger correlations between the short form and the full CTS2 than are
reported in this article.
Reliability and Validity Estimates. Internal consistency reliability is usually one of the
first ways of evaluating an instrument. However, this could not be done for the CTS2S
because there is no total score. The instrument consists of five separate scales that are not
intended to be summed to obtain a total score. It would be desirable to compute reliability
coefficients for each of the five scales, but this is also not appropriate because each scale
consists of only two items.
Concurrent Validity. The correlation of the approximation version of the CTS2S (see
above) with the full CTS2 was used to measure concurrent validity. However, as suggested
previously, because some items were augmented by adding behaviors (see step 3 above),
the results of using the approximation of the CTS2S can be considered as lower bound esti-
mates of what might be obtained using the actual CTS2S.
Construct Validity. Preliminary data on the construct validity of the CTS2S was
obtained by correlating four risk factor scales listed in the Other Measures section with
the short form scales. These four risk factors were also correlated with the full CTS2. To the
extent that the short form is conceptually equivalent to the full CTS2, it should have
the same pattern of correlation with these variables as the full scale. Because the data on
risk factors for partner violence were obtained only for the respondent, the construct valid-
ity correlations were performed only for perpetration of the behaviors measured by the
CTS2. Partial correlation, controlling for scores of socioeconomic status and social desir-
ability scales and for gender, was used because of possible confounding of these variables
with the risk factor and the CTS2 variables.

RESULTS

Concurrent Validity
Partial correlation of the short form scales with the corresponding full CTS2 scales, con-
trolling for scores on the SES and social desirability scales and for gender of the respon-
dent, resulted in the following correlations between the short form and the full scale.
Negotiation. r = .89 for Negotiation by the respondent, and r = .88 for Negotiation by the partner
Physical Assault. r = .72 for Assaults by the respondent on the partner, and r = .69 for Assaults
by the partner on the respondent
Injury. r = .94 for Injuries to the respondent, and r = .94 for Injuries to the partner as reported by
the respondent
Sexual Coercion. r = .65 for Sexual Coercion by the respondent, and r = .67 for Sexual Coercion
by the partner
Psychological Aggression. r = .77 for Psychological Aggression by the respondent, and r = .69
for Psychological Aggression by the partner.
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 512

512 Straus and Douglas

These coefficients indicate that measuring the five constructs by the CTS2S results
in scores that are highly correlated with the measures of these constructs using the full
CTS2.

Construct Validity
Partial correlations of five risk factors for partner violence with the short form CTS2 scales
and the full CTS2 scales are given in Table 1. These correlations controlled for scores on
the socioeconomic status and social desirability scales. The issue investigated by the cor-
relations in Table 1 is not the strength of the correlation between the risk factors and the
CTS2 variables but rather whether the results from using the short form parallel the results
when the full scale is used. This was evaluated by computing a test of the significance of
the differences between the short and long form for each of the pairs of correlations in
Table 1. Of the 25 pairs of correlations in Table 1, only one revealed a statistically signif-
icant difference in the results from using the short form and full scale versions of the
CTS2. That was for the correlation between negative attributions about the partner and
psychological aggression against the partner. The short form resulted in a significantly
lower correlation (r = .04) than the full scale (r = .19) at the .001 level. These results indi-
cate that, with one exception, the short form scales produce the same results as the full scale.

Prevalence Rates
One way of evaluating the validity of scales intended to measure illegal or reprehensible
behavior is the degree to which the instrument is successful in obtaining disclosure of
these behaviors. The rates for the short-form and the full CTS2 for the four reprehensible
behavior scales are given in Table 2. The full CTS2 results in prevalence rates that are from
20% greater to about double the rate obtained using this version of the short form. These

TABLE 1. Partial Correlation of CTS2 Short Form and Full CTS2 Scales With Five
Risk Factor Variables, Controlling for Social Desirability Response Set (N = 1,160)
Anger Couple Criminal Negative Violent
CTS Scale Management Conflict History Attributions Approval
Assault
Short -.18** .16** .05 .13** .11**
Full -.21** .17** .07* .14** .12**
Injury
Short -.08* .13** .10** .15** .12**
Full -.09* .15** .13** .19** .14**
Psych. Aggression
Short -.16** .12** -.02 .04 .00
Full -.21** .16** .05* .19** .07*
Sexual Coercion
Short -.09** .08* .13** .13** .17**
Full -.11** .14** .14** .15** .22**
Negotiation
Short -.04 .04 -.02 -.02 -.13**
Full -.02 .01 -.03 -.05 -.12**
*p ≤ .05, **p ≤ .01, one-tailed partial correlation coefficients.
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 513

Short Form of CTS2 513

TABLE 2. Prevalence Rates of Short-Form and Full CTS (N = 1,160)

Prevalence
Percentage Point Full Exceeds
CTS Scale Short Full Difference Short By
Assault
By partner 14.7 28.1 13.4 91.2
By respondent 16.7 29.6 12.9 77.2
Injury
By partner 6.6 8.0 1.4 21.2
By respondent 7.1 8.5 1.4 19.7
Psychological Aggression
By partner 55.7 76.7 21.0 37.7
By respondent 57.5 79.1 21.6 37.6
Sexual Coercion
By partner 13.0 25.6 12.6 97.0
By respondent 12.1 20.1 8.0 66.1

results are consistent with the principle that, for behavioral measures, the more different
forms of the behavior for which there are items in the scale, the higher the prevalence rate
(Straus, 1990).

Severity Levels
The comparison of the short and long form results for the severity level variables in Table 3
can be considered an extension of the comparison of prevalence rates for the short and full
scales in Table 2. The results extend those results by showing that the lower prevalence
rates from using the short form also apply to both the minor only and severe violence cat-
egories. All 12 of the comparisons in the column headed minor only show lower rates for
the short form, although often not much lower. The same applies to the comparisons in the
column for severe behavior.

Mutuality Types
The mutuality types provide a way for researchers and clinicians to identify “gender sym-
metry” or asymmetry in abusive relationship behavior. In general, Table 4 shows that the
short form produced distributions of mutuality types that were similar to the full scale. For
example, for all four of the CTS2 scales in Table 4, when there was maltreatment of a part-
ner, the most prevalent pattern was for both partners to engage in the behavior. However,
for physical assault, the short form produced higher percentages in the “male only” and
“female only” categories.

DISCUSSION

This article makes available a 20-question short form of the CTS2. The article also
provides procedures for using either the short form or the full CTS2 to classify respon-
dents into (a) mutually exclusive categories according to the severity of maltreatment of
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 514

514 Straus and Douglas

TABLE 3. Severity Level of Perpetration By Gender (N = 1,069)

Percent in Each Severity Level
Scale None Minor Only Severe
Assault
All respondents
Short 81.9 13.0 5.1
Full 70.8 19.6 9.6
Males
Short 85.0 10.6 4.4
Full 73.2 18.9 7.9
Females
Short 82.1 13.3 5.5
Full 69.7 19.9 10.4
Injury
All respondents
Short 92.4 5.9 1.8
Full 91.5 6.1 2.4
Males
Short 92.8 5.6 1.6
Full 92.0 5.7 2.2
Females
Short 92.6 5.7 1.7
Full 91.7 6.0 2.3
Psychological aggression
All respondents
Short 37.7 55.5 6.8
Full 39.7 62.7 16.5
Males
Short 39.7 52.0 8.2
Full 23.4 60.6 15.9
Females
Short 38.0 55.1 6.9
Full 20.8 61.7 17.5
Sexual coercion
All respondents
Short 86.9 10.8 2.3
Full 79.9 15.5 4.6
Males
Short 86.6 11.4 2.0
Full 74.6 20.0 5.4
Females
Short 86.6 11.4 2.0
Full 80.1 15.5 4.4

a partner or by a partner, and (b) categories to identify whether there was asymmetry or
mutuality in violence. Administration time for the short form is about 3 minutes, which
makes this version of the CTS2 suitable for situations where the time available for clin-
ical screening or research interview is not sufficient for the full 78-item CTS2, which
takes 10 to 15 minutes.
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 515

Short Form of CTS2 515

TABLE 4. Mutuality Types, by Severity and Gender

Percent in Each Type
n a Male Only Female Only Both
Assault
Total Short 223 9.4 27.8 62.8
Full 361 8.9 19.7 71.5
Severe Short 71 15.5 33.8 50.7
Full 134 14.9 36.6 45.8
Injury
Total Short 95 15.8 16.8 67.4
Full 107 15.9 18.7 65.4
Severe Short 20 5.0 5.0 90.0
Full 28 10.7 10.7 78.6
Psychological aggression
Total Short 680 2.1 4.9 93.1
Full 869 1.7 4.9 93.3
Severe Short 110 32.7 19.1 48.2
Full 231 19.0 26.0 55.5
Sexual coercion
Total Short 177 17.5 18.1 64.4
Full 305 29.2 10.5 60.3
Severe Short 29 27.6 27.6 44.8
Full 68 27.9 13.2 58.8
aThe ns are lower than for the other analyses, because mutuality types are computed only
for cases where there is at least one violent incident.

Comparison of Short and Full Scale

Concurrent Validity. Concurrent validity, as measured by the correlation between the
short form and full scales, ranged from .77 to .89 for perpetration of the behavior measured
by each scale, and from .65 to .94 for being victimized by a partner who engaged in these
behaviors. These are inflated concurrent validity coefficients because the items for the
short form were selected by taking the items that had the highest correlation with the total
scale and because they are part-whole correlations (Cohen, Cohen, West, & Aiken, 2003).
Nevertheless, they indicate that the part does reflect the whole, even though the precise
degree cannot be estimated using the current data.
Construct Validity. Given the high correlation between the short and long form version
of each CTS2 scale, it is not surprising that the construct validity analyses found that the
correlations of five risk factors for partner violence with the maltreatment of a partner mea-
sured by the CTS2S scales were generally parallel to the correlation of these risk factors
with maltreatment measured by the full CTS2.
Sensitivity. While there is strong evidence indicating the concurrent and construct
validity of the CTS2S, the short form has a much lower sensitivity, as evidenced by lower
estimates of the prevalence of each of the five behaviors measured by the CTS2. This
occurs because the CTS2 is a behavioral measure, as compared, for example, to an attitude
measure. For this type of measure, each additional behavior in the scale results in an increase
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 516

516 Straus and Douglas

in the percentage meeting the criterion of having experienced the behavior (Straus, 1990).
The lower sensitivity of the short form is a serious deficiency. Thus, users of the CTS2S
need to recognize that there may be significantly lower estimates of partner violence than
would be found with the full CTS2.
If there is insufficient time for the full CTS2 and estimating prevalence is crucial, a pos-
sible alternative is to use all the questions for the scale for which prevalence estimates are
most needed and omit the questions for the other scales. This, however, raises the question
of the extent to which asking only the questions for one scale, when they are not embed-
ded in random order with the questions for the other scales, is equivalent to the results from
the full CTS2.

Severity Levels and Mutuality Types

Severity Level. The severity level variable classifies cases as none, minor only, and
severe. These categories provide a way of avoiding the confounding of minor and severe
acts of maltreatment that occurs because some of the respondents who used or experienced
minor forms of maltreatment also experience or perpetrate more severe forms of that type
of maltreatment.
Mutuality Types. The CTS2 asks respondents about their own behavior and the behav-
ior of their partner. Usually, asking a respondent about the behavior of someone else is a
doubtful procedure unless the focus is on the respondent’s perception. However, in the case
of the CTS2, the questions ask about the behavior of the partner toward the respondent.
Thus, except for the injury scale, the respondent has full knowledge of each behavior in the
CTS2 by the partner. This makes it possible to classify the couples into mutuality types. The
mutuality types provide an elementary but important way of taking into account the fact that
behaviors measured by the CTS2 are embedded in a system of interaction. The mutuality
types classify couples into the following categories for each CTS2 scale: male partner only,
female partner only, both. The percentage of the sample classified into these three categories
by the short form and the full CTS2 is similar.
A problem with this method of obtaining mutuality types is that, unless the CTS2 is
administered to both partners, the validity of the classification depends on the accuracy of
one partner’s report. Moreover, the possible bias is confounded with the gender of the
respondent. Most studies have shown little difference in prevalence rates reported by males
and females (Archer, 2000). However, enough studies have shown a tendency for males to
underreport both perpetration and victimization to make it desirable to test both partners
or, if that is not possible, to exercise caution in conclusions based on the report of only one
partner. Although administering the CTS2 to both partners is desirable, it also requires a
procedure to use when the reports of the partners disagree. One approach is to use the
higher of the two reports.

Limitations
A major limitation is that in order to avoid having to obtain data for an entirely new sample,
the study analyzed a version of the short form that was computed from existing data. Con-
sequently, the results are only an approximation of what might be obtained using the short
form. We believe that the results based on this approximation of the short form are lower-
bound estimates. This is because the approximation to the short form lacked the additional
behaviors that are included in the augmented short form items. If the version of the CTS2S
using the items that have been augmented to include more of the behaviors that are in the
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 517

Short Form of CTS2 517

full scale is used, it may result in greater correspondence with the full CTS2 than was
found using the approximation items. However, even if that proves to be correct, the short
form is very likely to still have a lower sensitivity because the augmented items could not
include all the behaviors in the full CTS2.
Another limitation of the study is that the data refer entirely to the behavior toward a
partner by university students. This problem is mitigated somewhat by the fact that stud-
ies of students have generally found results that are parallel to the results of studies of gen-
eral population samples, with the important exception that, because of their youth, student
samples always have much higher scores on the Physical Assault scale.
A potential problem from using the approximation CTS2S, consisting of a subset of
items answered as part of the full scale, is the possibility that results from using the CTS2S,
which consists of only 20 items, will be different because respondents may react differently
when the items are not in the context of the larger set of items in the full CTS2.

CONCLUSIONS

Despite these limitations, the results are sufficiently promising to make it appropriate to
investigate the validity of the short form using nonstudent and clinical samples. Because
the short form has much lower sensitivity than the full CTS2, use of the short form results
in a much higher rate of false negatives. Nevertheless, because of the high sensitivity of
the full CTS2 (Straus, 1990, 1999), if the full CTS2 cannot be used, the short form is likely
to identify a large number of cases of partner violence.
For research focused on testing theories, such as testing the hypothesis that a certain
risk factor is associated with one or more of the aspects of partner maltreatment measured
by the CTS2, the concurrent and construct validity analyses in this article suggest that the
short form is likely to produce results that are sufficiently parallel to the results from
the full CTS2 to consider using the short form when using the full CTS2 is not possible.
Regardless of whether the full scale or the short form is used, the severity level types
are recommended because partners who engage in the less severe forms of maltreatment
may also engage in more severe forms. The severity level types avoid this confounding.
The mutuality types are recommended because at least half of the victims of maltreatment
are also perpetrators. The mutuality types enable identification of cases of mutual violence,
male-only violence, and female-only violence, and thus enable a first step in investigating
the dynamics of partner violence.

REFERENCES

Archer, J. (2000). Sex differences in aggression between heterosexual partners: A meta-analytic

review. Psychological Bulletin, 126(5), 651–680.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation
analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum.
Feld, S. L., & Straus, M. A. (1989). Escalation and desistance of wife assault in marriage.
Criminology, 27(1), 141–161.
Gelles, R. J., & Straus, M. A. (1988). Intimate violence. New York: Simon & Schuster.
Reynolds, W. M. (1982). Development of reliable and valid short forms of the Marlowe-Crowne
Social Desirability Scale. Journal of Clinical Psychology, 38(1), 119–125.
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 518

518 Straus and Douglas

Straus, M. A. (1990). The Conflict Tactics Scales and its critics: An evaluation and new data on valid-
ity and reliability. In M. A. Straus & R. J. Gelles (Eds.), Physical violence in American fami-
lies: Risk factors and adaptations to violence in 8,145 families (pp. 49–73). New Brunswick,
NJ: Transaction.
Straus, M. A. (1999). The controversy over domestic violence by women: A methodological, theo-
retical, and sociology of science analysis. In X. Arriaga & S. Oskamp (Eds.), Violence in inti-
mate relationships (pp. 17–44). Thousand Oaks, CA: Sage.
Straus, M. A. (2000a). Scoring the CTS2 and CTSPC. Durham, NH: University of New Hampshire,
Family Research Laboratory. Available at: http://pubpages.unh.edu/~mas2/
Straus, M. A. (2002). Revised Conflict Tactics Scales (CTS2) variable name system and SPSS syntax
for scoring. Durham, NH: Family Research Laboratory, University of New Hampshire.
Straus, M. A. (2004). Bibliography and tabular summary of publications on the revised Conflict
Tactics Scales (CTS2 and CTSPC). Durham, NH: Family Research Laboratory, University of
New Hampshire.
Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. B. (1996). The revised Conflict
Tactics Scales (CTS2): Development and preliminary psychometric data. Journal of Family
Issues, 17(3), 283–316.
Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. (1999). The personal and relation-
ships profile (PRP). Durham, NH: University of New Hampshire, Family Research Laboratory.
Available at: http://pubpages.unh.edu/~mas2/
Straus, M. A., & Mouradian, V. E. (1999, November 19). Preliminary psychometric data for the per-
sonal and relationships profile (PRP): A multi-scale tool for clinical screening and research on
partner violence. Paper presented at the meeting of the American Society of Criminology,
Toronto, Ontario, Canada.
Yodanis, C., Hill, K., & Straus, M. A. (1997). Tabular summary of methodological characteristics of
research using the Conflict Tactics Scales. Durham, NH: University of New Hampshire, Family
Research Laboratory. Available at http://pubpages.unh.edu/~mas2/

Offprints. Requests for offprints should be directed to Murray A. Straus, PhD, Family Research
Laboratory, University of New Hampshire, 126 Horton SSC, Durham, NH 03824. E-mail:
[email protected]

APPENDIX A: THE CTS2S SHORT FORM

Sample items from the CTS2S copyright © 2003, 2004 by Western Psychological Services.
Reprinted by permission of the publisher. No additional use without the prior written
authorization of the publisher. All rights reserved. For permission to use this instrument,
contact WPS, Attn: Rights & Permissions, 12031 Wilshire Boulevard, Los Angeles, CA
90025, USA.

Couple Conflicts
No matter how well a couple gets along, there are times when they disagree, get annoyed
with the other person, want different things from each other, or just have spats or fights
because they are in a bad mood, are tired, or for some other reason. Couples also have
many different ways of trying to settle their differences. This is a list of things that might
happen when you have differences. Please mark how many times you did each of these
things in the past year, and how many times your partner did them in the past year. If you
or your partner did not do one of these things in the past year, but it happened before that,
mark a “7” for that question. If it never happened, mark an “8.”
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 519

Short Form of CTS2 519

How often did this happen?

1 = Once in the past year
2 = Twice in the past year
3 = 3–5 times in the past year
4 = 6–10 times in the past year
5 = 11–20 times in the past year
6 = More than 20 times in the past year
7 = Not in the past year, but it did happen before
8 = This has never happened

1. I explained my side or suggested a compromise for a

disagreement with my partner 1 2 3 4 5 6 7 8
9. I pushed, shoved, or slapped my partner 1 2 3 4 5 6 7 8
10. My partner pushed, shoved, or slapped me 1 2 3 4 5 6 7 8
11. I punched or kicked or beat-up my partner 1 2 3 4 5 6 7 8
12. My partner punched or kicked or beat-me-up 1 2 3 4 5 6 7 8

APPENDIX B: SCORING AND TRANSFORMING

THE CTS2 SHORT FORM

Scoring Method
Scoring the short form is identical to scoring the full CTS2. There are many ways to score
the CTS2. These are described in Straus et al., 1996 and Straus, 2000b. The optimal
method depends on the characteristics of the sample and the purpose for which the scales
will be used. The most usual and recommended scoring method for the Physical Assault,
Injury, and Sexual Coercion scales is to create a dummy variable for “prevalence” by
assigning a score of 1 if one or more instances of the items were reported to have occurred
in the past year and 0 if no instances were reported. For the Negotiation scale, the recom-
mended scoring method is to sum the number of times each behavior was reported. To do
this, the answer categories must be converted from 0 to 7 to the midpoint of the range of
scores in each category (Straus et al., 1996).

Transforming the CTS2 or CTS2S Scores to Compute Severity Levels

The procedure to create severity types uses the prevalence scoring (0 = no instance of the
behavior measured by the scale; 1 = one or more instances). The procedure is illustrated by
the following syntax to compute severity types for assaultive behavior by male respondents:
IF (MaleMinor = 0 AND MaleSever = 0) MaleSeverityTyp = 0.
IF (MaleMinor = 1 AND MaleSever = 0) Male SeverityTyp = 1.
IF (MaleSever = 1) MaleSeverityType = 2.
VAR LABELS MaleSeverityType ‘SHORT FRM ASSAULT BY MALE-SEVERITY TYPE.’
VALUE LABELS MaleSeverityType 0 ‘NO VIOL’ 1 ‘MINOR ONLY’ 2 ‘SEVERE.’
Where:
MaleMinor = Minor assault by male partner
ch01_straus_4370.qxd 2/11/05 9:05 AM Page 520

520 Straus and Douglas

MaleSever = Severe assault by male partner

MaleSeverityLevel = Assault severity level by male partner

Syntax following the above pattern should be used to create Assault Severity Levels
for the female partner. Severity levels can also be computed for Injury, Psychological
Aggression, and Sexual Coercion by male and female partners—a total of eight Severity
Level variables.

Transforming the CTS2 or CTS2S Scores to Compute Mutuality Types

Below is the syntax that can be used to compute “Mutuality Types”, which allows
researchers to determine gender symmetry among violent relationships.
IF (SevereMale LE 0 AND SevereFemale LE 0) SevereCouple = SYMIS.
IF (SevereMale GE 1 AND SevereFemale GE 1) SevereCouple = 3.
IF (SevereMale GE 1 AND SevereFemale = 0) SevereCouple = 1.
IF (SevereMale = 0 AND SevereFemale GE 1) SevereCouple = 2.
VAR LABELS SevereCouple ‘ASSAULT SEVERE, MUTUALITY TYPES’.
VALUE LABELS SevereCouple 1 ‘Male Only’ 2 ‘Female Only’ 3 ‘Both’.