Personality and Individual Differences 97 (2016) 146–150

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

Item selection and validation of a brief, 20-item version of the

Differentiation of Self Inventory—Revised
David Sloan a,⁎, Dirk van Dierendonck b
a
Whitworth University, USA
b
Erasmus University, Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: The 46-item Differentiation of Self Inventory—Revised (DSI-R) (Skowron & Schmitt, 2003) is perhaps the most
Received 30 December 2015 widely used and psychometrically validated instrument measuring the family systems theory (Bowen, 1978)
Received in revised form 7 March 2016 concept of differentiation of self. Given the substantial length of the DSI-R, Drake et al. (2015) offered the
Accepted 16 March 2016
Differentiation of Self Inventory—Short Form (DSI-SF). However, based on limitations of the DSI-SF, we
Available online 28 March 2016
empirically validated a brief version of the 4-factor DSI-R using data from two adult samples (Sample 1: n =
Keywords:
541; Sample 2: n = 203). First, we conducted an EFA on the full 46-item DSI-R scale on Sample 1 to help us
Differentiation of self identify items most strongly representing each factor. Then we conducted CFAs comparing the model fit between
Item selection our brief DSI, the DSI-SF, and the full DSI-R on both samples. Results indicated that the brief DSI had the strongest
Short model fit in both samples. Finally, a comparison of associations among the brief DSI, the DSI-SF, and full DSI-R and
Brief various relevant variables showed similar correlations. The brief DSI provides counselors and researchers in the
Abridged areas of psychology, organizational behavior, and business a useful measure of differentiation of self when
Validation circumstances prohibit the use of the full DSI-R.
Family systems theory
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction emotional reactivity, and become emotionally overinvolved with or

Because of an increasing interest in understanding how relational
systems affect psychological and behavioral functioning, scholars have
looked to Bowen's (1978) Family Systems Theory (FST). Articulated as
one of the most developed theories on human functioning (Skowron
& Friedlander, 1998), FST claims that differentiation of self (DOS) is
of key importance (Bowen, 1978). The concept has already been
widely examined in the area of counseling psychology; more recently,
researchers have begun to explore the application of DOS in the areas
of business and organizational behavior (Beebe & Frisch, 2009;
Cochran, 2011).
DOS refers to the degree to which an individual can balance emo-
tional with intellectual functioning and closeness with independence
in relationships with others (Bowen, 1978). Individuals high in DOS
are able to distinguish their thoughts from their feelings and are capable
of balancing degrees of autonomy and intimacy (Bowen, 1978; Skowron
& Friedlander, 1998). More specifically, individuals with high DOS have
the ability to maintain the I-positon, in which they keep a well-defined
sense of self and possess the capacity to follow convictions despite
experiencing pressure to do otherwise (Skowron & Friedlander, 1998).
Individuals low in DOS tend to be more emotionally cut off, engage in
⁎ Corresponding author at: Whitworth University, 300 W. Hawthorne Road, Spokane,
WA, USA.
E-mail address: DSloan@Whitworth.edu (D. Sloan).
“fused” with others in relationships (p. 239). The most widely used
and psychometrically validated instrument measuring DOS is the
46-item Differentiation of Self Inventory—Revised (DSI-R) (Skowron &
Schmitt, 2003).
Practical reasons warrant the use of a brief version of the DSI-R. The
length of the full 46-item scale takes time that could be used for other
purposes in a clinical setting, or that could be used for measuring
additional variables in a research setting (Roets & Van Hiel, 2011). The
large number of items may also introduce respondent fatigue or bore-
dom, which has potential to diminish the quality of the response set
(Credé, Harms, Niehorster, & Gaye-Valentine, 2012). The purpose of
this paper is to put forth a brief version of the four-factor DSI-R while
maintaining the validity of the full instrument.
1.1. Background to DSI-R
Researchers have found that DOS influences the psychological
health and functioning of emotional systems across various contexts
(e.g., Gushue & Constantine, 2003; Peleg, 2005; Tuason & Friedlander,
2000). Empirically, differentiation has been found to lower stress and
anxiety (Murdock & Gore, 2004), increase psychological well-being
(Skowron, Holmes, & Sabatelli, 2003), and improve relational function-
ing (Chung & Gale, 2009). Evidence also supports differentiation as
negatively related to attachment avoidance and anxiety while being
positively related to effortful control (Skowron & Dendy, 2004).

http://dx.doi.org/10.1016/j.paid.2016.03.037
0191-8869/© 2016 Elsevier Ltd. All rights reserved.
 D. Sloan, D. van Dierendonck / Personality and Individual Differences 97 (2016) 146–150
Although multiple instruments measuring DOS exist (i.e., Bartle &
Sabatelli, 1995; Haber, 1993; Jankowski & Hooper, 2012; Licht &
Chabot, 2006; Skowron & Friedlander, 1998; Skowron & Schmitt,
2003), the 46-item DSI-R (Skowron & Schmitt, 2003) is one of the
most used and psychometrically validated instruments. It is the de
facto instrument for the majority of researchers and practitioners
interested in DOS.1
Skowron and Friedlander (1998) developed the Differentiation of
Self Inventory as a multidimensional construct measuring differentia-
tion with four subcategories: emotional cutoff (EC), emotional reactivity
(ER), I-position (IP), and fusion with others (FO). Although three of the
four subscales were psychometrically sound, the FO subscale lacked
reliability and validity. Skowron and Schmitt (2003) recognized this
weakness and revised the FO subscale to strengthen reliability and va-
lidity. Their efforts resulted in the construction of the Differentiation
of Self Inventory—Revised (DSI-R) with acceptable internal consisten-
cies measured using Cronbach's alpha ranging from .81 to .89 for sub-
scales and to .92 for the full scale. Skowron and Dendy (2004)
confirmed the four-factor structure and showed similar reliabilities.
However, Chung and Gale (2009) reported weaker internal consisten-
cies among college students in a Korean sample (ranging between .63
and .69 for the subscales; full scale = .85) and a European American
sample (ranging between .65 and .84; full scale = .87).
1.2. DSI-SF
The need for a shorter scale was recognized by Drake, Murdock,
Marszalek, and Barber (2015), who performed further analysis to
construct a short-form version of the DSI-R. They introduced the
Differentiation of Self Inventory—Short Form (DSI-SF) as a 20-item
instrument constructed using item response theory. However, whereas
the original DSI-R focused on adults (Skowron & Schmitt, 2003), the
DSI-SF was developed exclusively with a college-aged sample.
Certain considerations are important when relying solely on college-
age participants. First, several studies have reported differences
between a college-age sample and the society at large regarding
elements crucial for DOS. For example, college students in the
United States differed from their adult, nonstudent counterparts in
susceptibility to attitude change (Krosnick & Alwin, 1989), degree
of self-monitoring (Reifman, Klein, & Murphy, 1989), and social
influence (Pasupathi, 1999). In addition, Lamont's (2000) findings
suggest that adults without college educations differ from college
students in terms of the makeup of their social networks. These
findings bring into question the generalizability of the outcomes of
a validity study based exclusively on a college-age sample to the
general population (Henrich, Heine, & Norenzayan, 2010).
Second, the selection method for the DSI-SF items was item response
theory (IRT), and no factor analysis was performed. IRT is an excellent
method for testing for equivalence of item parameters within one factor
(Meade & Lautenschlager, 2004). However, what is missing is taking
into account in the overlap in item content between constructs. For
testing multidimensional models such as the four-factor construct of
differentiation (Skowron & Schmitt, 2003), factor analysis provides
that information (Wang & Russell, 2005). Regretfully, Drake et al.
(2015) conducted only a principal component analysis (PCA) to confirm
the unidimensionality of the separate subscales. They also did not report
any confirmatory factor analysis (CFA) results.
Third, the number of items kept in the DSI-SF subscales was not
equivalent. With three items for EC, five in FO, and six in ER and IP,
the overall scores for ER and IP are better represented than EC
and slightly better represented than FO. With this study, we aim to
introduce a shorter scale that takes into account all three criticisms.
1
Skowron, Friedlander, and Mallinckrodt's (2009) 43-item instrument is the original
DSI developed in 1998. The 2009 article simply reports an error correction found in
Skowron and Friedlander's (1998) study.
147
1.3. The need for a brief version of the DSI-R
Clinical psychologists and researchers in the disciplines of
psychology, organizational behavior, and business would benefit
from having an efficient method to evaluate DOS that is generaliz-
able to populations outside college students. In this study we present
an instrument for this purpose: a brief version of the previously val-
idated DSI-R, which is psychometrically sound and captures each of
the four dimensions assessed in the full 46-item scale. In construct-
ing our brief version, we followed the recommendations by Smith,
McCarthy, and Anderson (2000), offering a clear intended use of
the short form and refraining from assuming that the reliability
and validity of the original full-length scale automatically applied
to shortened versions.
To develop the brief DSI, we combined insights gained from reliabil-
ity analysis in terms of internal consistency of the subscales with an
exploratory factor analysis (EFA) on the full 46-item DSI-R scale using
a single sample of adults. To reduce the scale, we kept those items
with the highest item-total analysis and removed those items with the
lowest loadings on their own dimension as well as loading on multiple
factors. In this way, we developed a brief version of the scale that
included the essential items from each subscale while removing those
items that were not clearly representative of one subdimension only.
To confirm our final solution, we conducted a CFA comparing the
model fit between our brief DSI, the DSI-SF, and the full DSI-R. We
repeated the same CFA on a second independent sample. Finally,
to test the construct validity of our short version, we compared the
associations among the brief DSI, the DSI-SF, and the full DSI-R with
other relevant variables.
Variables chosen for comparison were based on theoretical
and empirical associations. These variables included attachment
anxiety and avoidance (Bowlby, 1982) and effortful control
(Rothbart, Ahadi, & Evans, 2000). According to attachment theory
(Bowlby, 1982), anxiously attached individuals are hyperactive in
response to stressful systems. Individuals with avoidant characteristics
tend to deny the importance of intimacy and attachment in relation-
ships. The capacity for effortful control depends on a person having
been securely attached with caregivers. This capacity predicts social
functioning and psychological adjustment in adults (Skowron &
Dendy, 2004). According to Bowen (1978), DOS either constrains or en-
ables that individual's capacity for self-regulation, or effortful control, as
well as his or her tendency to exhibit emotionally reactive (anxious) or
emotionally cutoff (avoidant) behaviors.
Skowron and Dendy (2004) empirically tested the theoretical
associations among self-differentiation and attachment anxiety, attach-
ment avoidance, and effortful control. Their results confirmed that DOS
was significantly related to effortful control and to both attachment
dimensions. These variables were chosen to compare the magnitudes
of the brief DSI, the DSI-SF, and the full DSI-R scales.
2. Method
2.1. Participants
2.1.1. Sample 1
DSI-R data were collected from 614 fully employed adults living
in the United States through recruitment services provided by
SurveyMonkey Audience. 541 (85%) participants provided usable
data. Population demographics were as follows: 52.1% were male,
89.9% white/Caucasian, 2.9% Asian, 2.9% black/African American,
2.0% Latin/Hispanic, and 1.6% multiracial. Nearly twenty-three
percent (22.7%) of participants were between 25 and 39 years of
age, 23.4% were between 40 and 49, 35.5% were between 50 and 59,
and 18.5% were 60 or older. The average job tenure was 12.00 years
(SD = 9.94), ranging from less than 1 year to 60 years.
148 D. Sloan, D. van Dierendonck / Personality and Individual Differences 97 (2016) 146–150
2.1.2. Sample 2
DSI-R data were collected from 252 fully employed adults living in
the United States through recruitment services provided by Amazon's
Mechanical Turk. Slightly more than 200(203, or 81%) participants
provided usable data. Population demographics were as follows: 39.1%
male, 82.8% white/Caucasian, 2.0% Asian, 9.4% black/African American,
4.4% Latin/Hispanic, and 1.5% from other races. The average age of
participants was 37.47 years (SD = 12.52), ranging from 19 to
77 years, and the average job tenure was 6.22 years (SD = 5.85),
ranging from less than 1 year to 35 years.
2.2. Measures
2.2.1. Differentiation of Self Inventory—Revised
Differentiation was measured with the full 46-item Differentiation
of Self Inventory—Revised (DSI-R) (Skowron & Schmitt, 2003).
Respondents rated items on a 6-point Likert-type scale on which 1
represents not at all true of me and 6 represents very true of me. The
DSI-R is a self-report measure focusing on adults and their significant
relationships that measures four factors: emotional cutoff (e.g., “I
have difficulty expressing my feelings to people I care for”), emotional
reactivity (e.g., “People have remarked that I'm overly emotional”),
fusion with others (e.g., “I usually need a lot of encouragement from
others when starting a big job or task”), and the ability to speak from
the I-position (e.g., “I tend to remain pretty calm even under stress”).
The DSI-R was included in both samples. Internal consistencies as
reported by Skowron and Schmitt (2003) reflected the full scale and
its subscales (Cronbach's α: DSI-R full scale = .92, ER = .89, IP = .81,
and FO = .85).
2.2.2. Experiences in close relationships scale
Attachment anxiety and attachment avoidance were measured
in Sample 2 using the 36-item self-report experiences in close relation-
ships scale (ECR) (Brennan, Clark, & Shaver, 1998). Participants rated
items on a 7-point Likert-type scale on which 1 represents disagree
strongly and 7 represents agree strongly. An example of an item measur-
ing attachment anxiety is “I worry a lot about my relationships.” An item
example measuring attachment avoidance is “I am nervous when
partners get too close to me.” The internal consistency of the experi-
ences in close relationships scale in the current sample was good
(Cronbach's α: ECR Full Scale = .96, Anxiety = .95, Avoidance = .93).
2.2.3. Effortful control scale
Effortful control was measured in Sample 2 using the 19-item
ATO-S-EC scale from the Adult Temperament Questionnaire—Short
Form (ATO-S) (Rothbart et al., 2000). The ATO-S-EC is a self-report
measure where participants rate items on a 7-point Likert-type
scale on which 1 represents extremely untrue of you and 7 represents
extremely true of you. Effortful control measures three factors, including
activation control (e.g., “I can keep performing a task even when I would
rather not do it”), attention control (e.g., “When interrupted or
distracted, I usually can easily shift my attention back to whatever I
was doing before”), and inhibitory control (e.g., “It is easy for me to
inhibit fun behavior that would be inappropriate”). The internal
consistency of the effortful control scale in the current sample was
good (Cronbach's α = .86).
3. Results
In the selection of items from the DSI-R for the brief DSI, (a) we
focused on strong internal consistency by checking for high item-total
correlations and a strong factorial validity based on the results of an
EFA; (b) we checked the content validity of individual items, such that
items chosen represented the breadth and scope of differentiation
constructs; and (c) we tried to maintain the reliabilities of the 20-item
brief DSI and its subscales. We confirmed the proposed brief DSI with
CFA results comparing the brief DSI, the DSI-SR, and the full DSI-R in
two samples.
In step one, using Sample 1, we calculated the item/total correlation
with internal consistency. We also conducted an EFA on the full 46-item
scale, eliminating items that loaded relevantly low (less than .3) on their
own factors and relatively high (greater than .6) on one or more of the
other factors. This analysis allowed us to reduce the 46-item instrument
to a 20-item brief DSI scale. The extraction method was Principle
Component Analysis and the rotation method was Varimax with Kaiser
Normalization. The items selected for the 20-item brief DSI scale includ-
ed emotional cutoff (items 3, 20, 24, 28, & 39); emotional reactivity
(items 14, 18, 21, 30, & 34); fusion with others (items 13, 17, 25, 33, &
44); and I-positon (items 7, 15, 19, 31, & 43). The FO subscale is
the only subscale with items loading on more than one factor. We
expected these results, given that in the literature the FO subscale has
consistently shown weaker validity and less reliability than the other
scales (e.g., Drake et al., 2015; Jankowski & Hooper, 2012; Johnson,
Thorngren, & Smith, 2001; Skowron, 2004).
We subsequently calculated the internal consistency of the full
scale and the four subscales. Our brief DSI had acceptable internal
consistencies measured using Cronbach's alpha: brief DSI (Full
Scale = .90; EC = .76; ER = .85; FO = .76; IP = .70).
Next, on the dataset from Sample 1, we conducted a CFA using MPlus
on the 20-item brief DSI scale (MLM X2[164] = 413.364; SRMR = .046,
RMSEA = .053, CFI = .918). We then compared the results with a CFA
on the DSI-SF (Drake et al., 2015) (MLM X2[164] = 520.717; SRMR =
.061, RMSEA = .064, CFI = .888) and on the full 46-item DSI-R
(Skowron & Schmitt, 2003) (MLM X2[983] = 3479.846; SRMR = .084,
RMSEA = .069, CFI = .701). Of the three versions, the factor structure
of our brief DSI clearly showed the strongest fit with the data.
To further verify our results, we repeated the CFA with the second
sample. Although the data from this sample indicated a lesser fitting
model compared with the first sample, our brief DSI scale, again, consis-
tently showed a better fit (MLM X2[164] = 325.014; SRMR = .068,
RMSEA = .070, CFI = .879) than the DSI-SF (MLM X2[164] = 374.562;
SRMR = .080, RMSEA = .080, CFI = .870) and the full DSI-R (MLM
X 2[983] = 3084.433; SRMR = .114, RMSEA = .103, CFI = .524).
Table 1 illustrates the comparative CFA results for the three scales.
Additional analysis showed that it would be possible to improve the
overall fit by removing FO items 21 and 33, which loaded on more than
one factor. The CFA results of the further reduced instrument showed a
better fit in both Sample one (MLM X2[129] = 271.776; SRMR = .039,
RMSEA = .045, CFI = .945) and Sample two (MLM X2[129] = 249.964;
SRMR = .066, RMSEA = .068, CFI = .898). However, based on our
rationale of wanting to keep the number of items representing
each construct consistent, and because of the results showing that the
20-item brief DSI consistently produced a stronger model fit than
the DSI-SF and the DSI-R, we prefer, for now, to keep the proposed
20-item brief DSI.
To test for construct validity, using Sample 2 we calculated correla-
tions between the brief DSI, the DSI-SF, and full DSI-R scales and
variables previously connected to DOS, such as attachment anxiety
Table 1
Goodness-of-Fit Indicators for the brief DSI, DSI-SF, and the full DSI-R on Sample 1 and
Sample 2.
Model χ2 df RMSEA CFI SRMR
Sample 1 (n = 541)
Brief DSI-R 413.36⁎⁎⁎ 164 .053 .918 .046
DSI-SF 753.42⁎⁎⁎ 165 .082 .816 .104
Full DSI-R 3479.85⁎⁎⁎ 983 .069 .701 .084
Sample 2 (n = 203)
Brief DSI-R 325.01⁎⁎⁎ 164 .070 .879 .068
DSI-SF 599.05⁎⁎⁎ 165 .114 .637 .118
Full DSI-R 3084.43⁎⁎⁎ 983 .103 .524 .114
⁎⁎⁎ p b .001.
 D. Sloan, D. van Dierendonck / Personality and Individual Differences 97 (2016) 146–150 149

Table 2
Correlations among and descriptive statistics for the brief DSI, DSI-SF, full DSI-R, attachment anxiety, attachment avoidance, and effortful control.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1. ER brief DSI –
2. ER DSI-SF .99⁎⁎ –
3. ER DSI-R .96⁎⁎ .97⁎⁎ –
4. EC brief DSI .57⁎⁎ .57⁎⁎ .58⁎⁎ –
5. EC DSI-SF .45⁎⁎ .44⁎⁎ .45⁎⁎ .78⁎⁎ –
6. EC DSI-R .51⁎⁎ .52⁎⁎ .53⁎⁎ .95⁎⁎ .86⁎⁎ –
7. FO brief DSI .79⁎⁎ .81⁎⁎ .80⁎⁎ .62⁎⁎ .43⁎⁎ .59⁎⁎ –
8. FO DSI-SF .80⁎⁎ .82⁎⁎ .80⁎⁎ .62⁎⁎ .44⁎⁎ .58⁎⁎ .97⁎⁎ –
9. FO DSI-R .73⁎⁎ .74⁎⁎ .73⁎⁎ .40⁎⁎ .30⁎⁎ .33⁎⁎ .77⁎⁎ .75⁎⁎ –
10. IP brief DSI .43⁎⁎ .41⁎⁎ .41⁎⁎ .25⁎⁎ .18⁎ .21⁎⁎ .41⁎⁎ .41⁎⁎ .40⁎⁎ –
11. IP DSI-SF .49⁎⁎ .46⁎⁎ .46⁎⁎ .24⁎⁎ .14⁎ .19⁎⁎ .44⁎⁎ .44⁎⁎ .40⁎⁎ .84⁎⁎ –
12. IP DSI-R .52⁎⁎ .49⁎⁎ .49⁎⁎ .30⁎⁎ .18⁎⁎ .25⁎⁎ .54⁎⁎ .54⁎⁎ .48⁎⁎ .92⁎⁎ .93⁎⁎ –
13. ECR anxiety −.54⁎⁎ −.56⁎⁎ −.58⁎⁎ −.66⁎⁎ −.48⁎⁎ −.65⁎⁎ −.64⁎⁎ −.63⁎⁎ −.44⁎⁎ −.27⁎⁎ −.28⁎⁎ −.37⁎⁎ –
14. ECR avoid −.37⁎⁎ −.38⁎⁎ −.38⁎⁎ −.70⁎⁎ −.66⁎⁎ −.75⁎⁎ −.41⁎⁎ −.41⁎⁎ −.20⁎⁎ −.18⁎⁎ −.19⁎⁎ −.24⁎⁎ .60⁎⁎ –
15. ATO-S
Effortful control .53⁎⁎ .51⁎⁎ .52⁎⁎ .46⁎⁎ .42⁎⁎ .43⁎⁎ .53⁎⁎ .53⁎⁎ .43⁎⁎ .49⁎⁎ .54⁎⁎ .57⁎⁎ −.50⁎⁎ −.35⁎⁎ –

n = 203.
⁎ p b .05.
⁎⁎ p b .01.

and avoidance (ECR) (Fraley, Waller, & Brennan, 2000) and effortful
control (ATO-S-EC) (Rothbart et al., 2000). As shown in Table 2, results
for all three versions were consistent with the findings reported by
Skowron and Dendy (2004).
Next, we compared how our brief DSI and the DSI-SF subscales
related to the full DSI-R. The correlations between the brief DSI and
the full DSI-R and between the DSI-SF and the full DSI-R were almost
identical under each subscale except emotional cutoff. On that subscale,
the correlation between the brief DSI subscale and the full instrument
was slightly stronger than that between the DSI-SF and the full
instrument (r = .95, r = .86, respectively).
4. Discussion
The purpose of this study was to provide clinical psychologists
and researchers in the areas of psychology, organizational behavior,
and business with an empirically validated brief version of the Differen-
tiation of Self Inventory—Revised (Skowron & Schmitt, 2003). A total of
20 items were selected based on their component loadings, including
five items from each DSI-R subscale. CFAs on two samples revealed
that our brief DSI showed a better model fit than the short form version
by Drake et al. (2015) and the full DSI-R scale while keeping the internal
consistency of the subscales intact. Finally, variables associated with the
DSI-R in the literature showed similar relationships with both the
full DSI-R and our brief DSI. The results of this study suggest that our
20-item brief DSI may be used as a valid alternative to the DOS
construct. We would like to emphasize that the brief DSI does not
attempt to replace the full 46-item scale under all circumstances. The
brief DSI is designed to measure DOS only when practical circumstances
prohibit clinical psychologists and researchers from using the full DSI-R.
The brief DSI version differed from the DSI-SF (Drake et al., 2015) in
8 out of 20 items. In our selection of items, we attempted to capture the
complexity of all four factors. For example, whereas the items measur-
ing emotional cutoff in the DSI-SF emphasized a sense of discomfort in
close, intense relationships, the brief DSI emotional cutoff subscale in-
cludes items related to concern about loss of personal independence,
confidence regarding the outcomes of talking about stressful issues,
and discomfort in overly intense relationships. In the fusion with others
subscale, although both versions captured the need for acceptance and
personal discomfort with criticism, only the brief DSI included items re-
garding a lack of confidence when making decisions alone. When mea-
suring the I-position, both instruments addressed individual sense of
self, personal stability, and concern for approval of others. Only the
brief DSI measured the ability to accept unchangeable circumstances
and the tendency to separate thoughts about a person from the issue.
Both versions used virtually identical items for the emotional reactivity
subscale. Based on these differences, the brief DSI is a good option for
researchers interested in a shortened version of the instrument that
also captures the complexity of each subscale.
Our brief DSI shows good construct validity by combining a strongly
improved factorial validity with good internal consistencies for sub-
scales and a content validity comparable to that of the full DSI-R. Most
notable is the improvement of the factorial validity compared with
both the full DSI-R scale and the DSI-SF. The brief DSI measures four
factors: emotional cutoff (e.g., “When things go wrong, talking about
them usually makes it worse”), emotional reactivity (e.g., “I'm overly
sensitive to criticism”), fusion with others (e.g., “I feel a need for
approval from virtually everyone in my life”), and the ability to
speak from the I-position (e.g., “When I am having an argument with
someone, I can separate my thoughts about the issue”).
4.1. Limitations
Despite the advantages of having a validated brief scale, we rec-
ognize limitations of this study. First, the FO subscale contained
items loading on more than one factor; the weakness and lack of
clarity of this subscale, as noted by other studies (Drake et al.,
2015; Jankowski & Hooper, 2012; Johnson, Thorngren, & Smith,
2001; Skowron, 2004) warrants further research strengthening the
FO subscale, perhaps through the addition of new items. Despite hav-
ing two items with double loadings, the final result remains the
strongest fitting model compared with the DSI-SF and full DSI-R.
A second limitation is that the ethnic/racial make-up of our partici-
pant pool was disproportionately Caucasian (89.9% in Sample one
and 82.8% in Sample two). However, we propose that our sample of
working-age adults, regardless of ethnic/racial makeup is a more
appropriate for developing a brief version of the DSI-R than is a stu-
dent sample. A third limitation is that the first sample consisted of
working adults. We attempt to address this limitation by including
unemployed and retired adults in Sample 2. Regardless, results of this
study should be generalized with caution to the greater population of
adults, as it is questionable whether those of differing ethnic/racial
backgrounds and those without full-time work would have pro-
duced results different from those of the participants in our study.
Nevertheless, the use of two adult samples is a strong point of our
study and gives us confidence that our results can be replicated.
Conducting research using the brief DSI with diverse people groups
is encouraged to expand the generalizability of the instrument
(Rosenthal & Rosnow, 1991).
150 D. Sloan, D. van Dierendonck / Personality and Individual Differences 97 (2016) 146–150
References
Bartle, S. E., & Sabatelli, R. M. (1995). The behavioral and emotional reactivity index:
Preliminary evidence for construct validity from three studies. Family Relations, 44,
267–277.
Beebe, R., & Frisch, N. (2009). Development of the differentiation of self and role invento-
ry for nurses (DSRI-RN): A tool to measure internal dimensions of workplace stress.
Nursing Outlook, 57(5), 240–245.
Bowen, M. (1978). Family therapy in clinical practice. New York, NY: J. Aronson.
Bowlby, J. (1982). Attachment and loss: Volume 1. Attachment. New York, NY: Basic Books.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult
attachment: An integrative overview. In J. A. Simpson, & W. S. Rholes (Eds.),
Attachment theory and close relationships (pp. 46–76). New York, NY: Guilford Press.
Chung, H., & Gale, J. (2009). Family functioning and self-differentiation: A cross-cultural
examination. Contemporary Family Therapy, 31(1), 19–33.
Cochran, N. B. (2011). Bowen family systems theory and its relationship to teachers: Does
differentiation of self predict teacher job satisfaction? (Doctoral disssertation)
Hattiesburg, MS: The University of Southern Mississippi (ProQuest Dissertations
and Theses).
Credé, Harms, Niehorster, & Gaye-Valentine (2012). An evaluation of the consequences of
using short measures of the big five personality traits. Journal of Personality and Social
Psychology, 102(4), 874–888.
Drake, J. R., Murdock, N. L., Marszalek, J. M., & Barber, C. E. (2015). Differentiation of self
inventory—Short form: Development and preliminary validation. Contemporary
Family Therapy, 1–12.
Fraley, R. C., Waller, N. G., & Brennan, K. A. (2000). An item response theory analysis
of self-report measures of adult attachment. Journal of Personality and Social
Psychology, 78(2), 350–365.
Gushue, G. V., & Constantine, M. G. (2003). Examining individualism, collectivism, and
self-differentiation in African American college women. Journal of Mental Health
Counseling, 25(1), 1–16.
Haber, J. (1993). A construct validity study of a differentiation of self scale. Scholarly
Inquiry for Nursing Practice, 7(3), 165–178.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?
Behavioral and Brain Sciences, 33(2–3), 61–135.
Jankowski, P. J., & Hooper, L. M. (2012). Differentiation of self: A validation study of the
Bowen theory construct. Couple and Family Psychology: Research and Practice, 1,
226–243.
Johnson, P., Thorngren, W. C., & Smith, A. J. (2001). Parental divorce and family functioning:
Effects on differential levels of young adults. The Family Journal, 9, 265–272.
Krosnick, J. A., & Alwin, D. F. (1989). Aging and susceptibility to attitude change. Journal of
Personality and Social Psychology, 57, 416–425.
Lamont, M. (2000). The dignity of working men. New York, NY: Russell Sage Foundation.
Licht, C., & Chabot, D. (2006). The Chabot emotional differentiation scale: A theoretically
and psychometrically sound instrument for measuring Bowen's intrapsychic aspect
of differentiation. Journal of Marital and Family Therapy, 32(2), 167–180.
Meade, A. W., & Lautenschlager, G. J. (2004). A comparison of item response theory
and confirmatory factor analytic methodologies for establishing measurement
equivalence/invariance. Organizational Research Methods, 7(4), 361–388.
Murdock, N. L., & Gore, P. A., Jr. (2004). Stress, coping, and differentiation of self: A test of
Bowen theory. Contemporary Family Therapy: An International Journal, 26(3),
319–335.
Pasupathi, M. (1999). Age differences in response to conformity pressure for emotional
and nonemotional material. Psychology and Aging, 14, 170–174.
Peleg, O. (2005). The relation between differentiation and social anxiety: What can be
learned from students and their parents? American Journal of Family Therapy, 33(2),
167–183.
Reifman, A., Klein, J., & Murphy, S. (1989). Self-monitoring and age. Psychology and Aging,
4, 245–246.
Roets, A., & Van Hiel, A. (2011). Item selection and validation of a brief, 15-item version of
the need for closure scale. Personality and Individual Differences, 50(1), 90–94.
Rosenthal, R., & Rosnow, R. (1991). Essentials of behavioral research: Methods and data
analysis. New York, NY: McGraw-Hill.
Rothbart, M. K., Ahadi, S. A., & Evans, D. E. (2000). Temperament and personality: Origins
and outcomes. Journal of Personality and Social Psychology, 78, 122–135.
Skowron, E. A., & Dendy, A. K. (2004). Differentiation of self and attachment in adulthood:
Relational correlates of effortful control. Contemporary Family Therapy, 26(3),
337–357.
Skowron, E. A., & Friedlander, M. L. (1998). The differentiation of self inventory:
Development and initial validation. Journal of Counseling Psychology, 45, 235–246.
Skowron, E. A., & Schmitt, T. A. (2003). Assessing interpersonal fusion: Reliability and
validity of a new DSI fusion with others subscale. Journal of Marital and Family
Therapy, 29(2), 209–222.
Skowron, E. A., Holmes, S. E., & Sabatelli, R. M. (2003). Deconstructing differentiation: Self
regulation, interdependent relating, and well-being in adulthood. Contemporary
Family Therapy, 25(1), 111–129.
Skowron, E. R. (2004). Differentiation of self mediates college stress and adjustment.
Journal of Counseling and Development, 82(1), 69–78.
Skowron, E., Friedlander, M., & Mallinckrodt, B. (2009). Errata. Journal of Counseling
Psychology, 56(4), 597–598.
Smith, G. T., McCarthy, D. M., & Anderson, K. G. (2000). On the sins of short-form
development. Psychological Assessment, 12(1), 102–111.
Tuason, M. T., & Friedlander, M. L. (2000). Do parents' differentiation levels predict those
of their adult children? And other tests of Bowen theory in a Philippine sample.
Journal of Counseling Psychology, 47(1), 27–35.
Wang, M., & Russell, S. S. (2005). Measurement equivalence of the job descriptive index
across Chinese and American workers: Results from confirmatory factor analysis
and item response theory. Educational and Psychological Measurement, 65, 709–732.
Dr. Sloan is a visiting assistant professor of Management in the School of Business
at Whitworth University. His areas of emphasis are leadership, business psychology,
organizational behavior, and human relations.
Dr. van Dierendonck is professor of Human Resource Management at Rotterdam
School of Management, Erasmus University (RSM). His areas of expertise include human
resource management, leadership and leadership development, positive organizational
scholarship, and measurement development.