Meta-Analysis of the Factor Structures of the Beck Depression Inventory

Abstract

Two meta-analyses examined the factor structure of the Beck Depression Inventory–II (BDI-II). Study 1, which meta-analyzed 51 studies comprising 62 samples (N = 20,475) providing pattern matrices, determined that the two-factor solution comprising Cognitive and Somatic-Affective factors was supported for the full sample. The two-factor solution was also supported for subgroups of studies. As the factor structure varied somewhat between subgroups of studies, the strength of relationships between scale items and their underlying depressive symptoms varied. Hence, comparisons of mean BDI-II scores across subgroups can be misleading. Study 2 meta-analyzed 13 studies consisting of 16 samples (N = 5,128) providing covariance matrices among the 21 BDI-II items. The two-factor solution was again supported in Study 2. Nevertheless, the existence of a general depression factor was supported by the good fit of the one-factor model.

Keywords

Beck Depression Inventory–II meta-analysis factor structure pattern matrix covariance matrix factor analysis

Depression can seriously impair an individual’s functioning and is frequently associated with other mental disorders (Ohayon & Schatzberg, 2010) and compromised physical health (Shulz, Beach, Ives, Martire, Ariyo, & Kop, 2000). The Beck Depression Inventory–II (BDI-II) is a popular scale to assess the presence and severity of depression for adolescents and adults aged ≥13 in the past 2 weeks. Performing factor analysis to examine the factor structure of the BDI-II has both research and practical implications. From a practical perspective, treatment programs may not be equally effective for different depression subfactors. Hence, the identification of the factor structure of depression can improve comprehensive evaluations of intervention programs. From a theoretical standpoint, subgroup means are frequently examined and invariance of the factor structure is a prerequisite of group mean comparisons. Furthermore, the broad depression index lacks specificity. If subdimensions of depression are differentially associated with antecedents and consequences, then the use of a global depression dimension may account for the inconsistent findings in literature.

The Beck Depression Inventory–Second Edition

The BDI-II is a revised version of the BDI based on criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994). Four of 21 BDI items (Body Image Change, Work Difficulty, Weight Loss, and Somatic Preoccupation) were replaced with new items (Agitation, Worthlessness, Loss of Energy, and Concentration Difficulty) in the BDI-II. Nineteen of the 21 BDI-II items consist of four statements arranged on a continuum from less depressed to more depressed. Items of Changes in Sleeping Pattern and Changes in Appetite contain seven statements to differentiate an increase or decrease in sleep hours and appetite. That is, the seven options were 0, 1a, 1b, 2a, 2b, 3a, and 3b, with a higher rated option indicating an increase or decrease in either symptom. These modifications were mainly to address the DSM-IV criterion.

Although no items have been assigned formally to the BDI-II subscales, numerous studies were conducted to examine the dimensionality of the BDI-II. However, the results obtained by these studies were inconsistent. A meta-analysis that quantitatively synthesizes the factor structures of the BDI-II was therefore needed.

Previous Meta-Analyses

A meta-analysis by Shafer (2006) identified 33 studies reporting BDI pattern matrices. Studies from 1966 to 1999 yielded 13,643 participants. A three-factor solution seems best in terms of having an interpretable factor structure. The first factor, Negative Attitudes toward Self, was defined by 11 items. One of these 11 items, Body Image Change, was replaced in the BDI-II. The second factor, Performance Impairment, was defined by seven items. Two of these seven items, Work Difficulty and Somatic Preoccupation, were replaced in the BDI-II. The third factor, Somatic Concerns, was defined by three items. One (Weight Loss) of these three items was replaced in the BDI-II. The BDI-II is a refined version of the BDI, but with sufficient modifications that may potentially affect the factor structure. Whether the three-factor solution applies to the BDI-II is thus of interest. Furthermore, the factor structure of the BDI-II may vary as a function of various moderators. However, moderator analysis of the factor structure of depressive symptoms was not conducted by Shafer (2006).

Effect Sizes in Meta-Analysis of Exploratory Factor Analysis

Pattern matrix, correlation matrix, and covariance matrix are three effect sizes that can be used in a meta-analysis of factor analysis. The pattern matrix is an item-by-factor matrix that contains factor loadings indicating which items are involved in which factors and to what extent. Factor loadings are interpreted in a manner similar to how regression coefficients are interpreted in multiple regression analysis. Shafer (2005, 2006) used pattern matrices to construct co-occurrence matrices that were then subjected to exploratory factor analysis.

Covariance matrices are effect size statistics that can be applied in meta-analysis of factor analysis. Beretvas and Furlow (2006) developed a method for analyzing covariance matrices. After identifying studies that reported covariance matrices, raw covariance matrices were weighted and aggregated to construct a mean covariance matrix. The mean covariance matrix was then subjected to confirmatory factor analysis. As the covariance matrix among items is rarely reported in empirical studies, a meta-analysis of the covariance matrix is rare.

To avoid the limitations of previous meta-analyses, both pattern and covariance matrices were used as effect sizes. Exploratory factor analysis was applied to analyze mean standardized co-occurrence matrices derived from pattern matrices, while confirmatory factor analysis was applied to mean covariance matrices.

Moderators

Language Version

The meta-analysis by Kim, DeCoster, Huang, and Bryant (2013), which examined the effect of language on the factor structure of the Geriatric Depression Scale, analyzed 26 studies (N =14,669). Support for the effect of language was strong. A four-factor structure was supported for the full sample, while models with four to six factors were supported for the 10 languages. Penley (2001) compared the factor structures of the English and Spanish versions of the BDI-II using a sample of bilingual college students. The effect of language was small. As cross-cultural research (Al-Musawi, 2001; Chang, 2005) is especially interested in determining whether the factor structure obtained for Western cultures can be applied to non-Western cultures, one must explore the effect of language on the factor structure of the BDI-II.

Risk Level

The factor structure of the BDI-II must be invariant across groups such that the means of groups with different levels of depression risk can be compared. Beck, Steer, and Brown (1996) reported different factor structures for a student sample and a clinical sample. The Cognitive and Somatic-Affective factors were for the clinical participants, while the Cognitive-Affective and Somatic factors were for the student sample. The factor structure differed for normal and clinical samples for the Turkish version of the BDI-II in the study by Kapci, Uslu, Turkcapar, and Karoglan (2008), who sampled 203 normal adults and 176 depressive disorder outpatients. A Performance Difficulty-Somatic and Negative Attitude factors were found for the normal sample, while Somatic-Affective and Cognitive factors were for the outpatient sample. As primary research used the BDI-II to assess depression for general populations such as students (Al-Musawi, 2001; Chang, 2005; Dozois, Dobson, & Ahnberg, 1998), high-risk participants such as patients with chronic fatigue syndrome (Brown, Kaplan, & Jason, 2011), and clinical participants with a mental disorder/illness (Cole, Grossman, Prilliman, & Hunsaker, 2003), the factor structures of the BDI-II in these studies were inconsistent and warranted comparisons of the factor structures of the BDI-II across groups with differential risk factors for depression.

Age

Studies comparing the factor structure of the BDI-II across age groups were scarce, and the study by Segal, Coolidge, Cahill, and O’Riley (2008) was an exception. They compared the factor structure for 229 young adults (mean age = 19.60 years) recruited from psychology courses and 147 adults living in a community (mean age = 70.30 years). Although the one-factor solution was acceptable for both age groups, the salience of items varied. Nonetheless, the number of factors underlying the BDI-II varied in studies by Steer and colleagues (Steer, Ball, Ranieri, & Beck, 1999; Steer, Kumar, Ranieri, & Beck, 1998). More specifically, Steer et al. (1999) sampled 210 depressed adults (mean age = 41.29 years) and found that the two-factor solution was supported. On the other hand, Steer et al. (1998) sampled psychiatric patients (mean age = 15.17 years) and found that the three-factor solution was supported. Because empirical research has rarely examined the age effect on the factor structure of the BDI-II, this meta-analysis examined the possible effect of age.

Confirmatory Factor Analysis of the BDI-II

Several factor structure models, including the two-factor, three-factor, bifactor (specific and general depression factors), and first-order with second-order depression factors models, were tested in primary research. However, findings regarding the support of these models were inconsistent. For example, Titov et al. (2011) surveyed 172 depressed patients and found that the two-factor solution with the Cognitive and Somatic factors had a good fit to data. Corbière et al. (2011), who surveyed 206 patients with chronic pain, found that the three-factor solution with Cognitive, Affective, and Somatic factors had the best fit to data. Vanheule, Desmet, Groenvynck, Rosseel, and Fontaine (2008) tested the two models in Beck et al. (1996) as well as 10 alternative models using a clinical sample of 404 patients and a general sample of 695 adults. They found that the fit of the two-factor solution with Cognitive and Somatic-Affective factors in the clinical sample by Beck et al. (1996) and the three-factor solution with Cognitive, Affective, and Somatic factors by Buckley, Parker, and Heggie (2001) were acceptable.

Ward (2006) reanalyzed data from five studies consisting of six samples and found that the three-factor solution with a general factor and two specific depression factors best represented data. The solution with a general factor with two specific factors was also supported by Quilty, Zhang, and Bagby (2010), who surveyed 425 adult patients with major depression and found that the general factor with Somatic- and Cognitive factors fit data well. The solution with a general factor and three specific factors, the Cognitive, Affective, and Somatic factors, was supported by Arnarson, Ólason, Smári, and Sigurđsson (2008), who tested the factor structure of the Icelandic BDI-II using 1,206 students and 248 patients. Cross-country support for the presence of general depression with specific depression factors was also demonstrated by Brouwer, Meijer, and Zevalkink (2013), who surveyed a large sample of 1,530 patients to test the Dutch BDI-II, and Al-Turkait and Ohaeri (2010), who used 624 college students to test the factor structure of the Arabic BDI-II. Along a similar line, Harris and D’Eon (2008), who analyzed 481 Canadian patients with chronic pain, found that the fit to data of the model with three first-order factors, Negative Attitude, Performance Difficulty, and Somatic Elements, and a second-order factor was good. Given findings range widely, a meta-analysis is needed because it can identify the overall factor structure of the BDI-II. This study combined covariance matrices obtained by previous studies to extend and clarify previous findings.

Study 1

Method

Literature Search

To locate potential studies, the ERIC, PsycINFO, PubMed, and ProQuest Dissertations and Theses databases were searched up to March 2013. The search terms were the Beck Depression Inventory Second Edition, Beck Depression Inventory–II, and BDI-II combined with factor analysis, factor analyses, principal component analysis, principal component analyses, factor structure, factorial, validity, validation, item analysis, dimension, dimensionality, or psychometric. Included studies must meet the following criteria. First, studies must report correlation matrices or pattern matrices from exploratory factor analysis of the BDI-II. For studies reporting the pattern matrix, they must provide at least the highest factor loadings for BDII-II items. Furthermore, studies must be published in English. The studies included are listed in the references prefixed with an asterisk.

Analysis

For the use of pattern matrix as an effect size, the sample-size-weighted mean co-occurrence matrix, which served as the correlation matrix, was subjected to exploratory factor analysis. To compute a sample-size-weighted mean co-occurrence matrix, a raw co-occurrence matrix for each pair of the 21 BDI-II items was computed. When a pair of BDI-II items had their highest loading on the same factor and their factor loading was ≥.4, they were given a value of 1; otherwise, a value of 0 was assigned.

Some researchers did not administer all 21 BDI-II items to study participants. Specifically, Byrne, Stewart, and Lee, (2004), Lindsay and Skene (2007), and Palmer and Binks (2008) did not administer the item of Loss of Interest in Sex. Hence, sample size for Loss of Interest in Sex was smaller than those for the remaining 20 items. To compare co-occurrence matrices across studies, standardized co-occurrence was acquired by calculating the proportion of each pair of BDI-II items with the highest loadings on the same factor. That is, the standardized co-occurrence matrix was derived by dividing the number of times a pair of BDI-II items had their highest loadings on the same factor by the total number of times that a pair of BDI-II items was measured. The mean standardized co-occurrence matrix was then computed by averaging the sample-size-weighted standardized co-occurrence matrices. Conway and Huffcutt (2003) suggested that principal axis was generally a good factor extraction model and the oblique rotation was a preferred choice. As Promax rotation was the most commonly used rotation method in the included studies, factor analysis was performed via principal axis analysis followed by Promax rotation in the present study.

Results

Included Studies

This meta-analysis included 51 studies consisting of 62 independent samples involving 20,475 participants. Notably, some studies had overlapping samples. About half of the data Bedi (2001) analyzed was also analyzed in the studies by Beck et al. (1996) and Steer et al. (1999). In the study by Bos et al. (2009), 331 pregnant women were surveyed in their last trimester of pregnancy. Of all 331 women, 130 women and another new 184 mothers were tested at 3 months postpartum. As splitting overlapping participants among studies was not possible, these studies were considered independent.

Of the 51 studies, 42 were journal articles, 6 were doctoral dissertations, 2 were master’s theses, and 1 was a test manual. Table 1 lists sample size, mean age, gender, population type, language version, reliability coefficient in terms of α, factor extraction method, rotation method, number of factors retained, and percentage of variance explained. Average sample size was 330.24 participants (range = 51-2,095), and mean age was 32.86 years (range = 15-76). Seven samples were male only, 12 were female only, and 43 were mixed. Seventeen samples comprised college students. Of the 62 samples, the BDI-II version used and sample numbers were as follows: English BDI-II, 42 samples; Chinese BDI-II, 4 samples; Turkish BDI-II, 4 samples; Portuguese BDI-II, 3 samples; Spanish BDI-II, 3 samples; Arabic BDI-II, 2 samples; Japanese BDI-II, 1 sample; Croatian BDI-II, 1 sample; Malay BDI-II, 1 sample; and Xhosa BDI-II, 1 sample. The 21-item BDI-II was administered to 59 samples.

Table 1.

Included Studies.

Study	N	Age	Gender	Population type	Language version	α	Extraction method	Rotation	No. of factors retained	% of variance explained
Al-Musawi (2001)	200	23	Mixed	General	Arabic	.84	PA	Promax	3	47.54
Al-Turkait and Ohaeri (2012)	624	20.80	Mixed	General	Arabic	.83	PA	Promax	4	41.80
Arnau, Meagher, Norris, and Bramson (2001)	340	36.50	Mixed	Preselected	English	.94	PC	Promax	2	53.50
Beck et al. (1996), psychiatric patients	500	37.20	Mixed	Clinical	English	.92	PA	Promax	2	NA
Beck et al. (1996), college students	120	19.58	Mixed	General	English	.93	PA	Promax	2	NA
Bedi (2001)	390	44.80	Female	Clinical	English	NA	PA	Promax	2	44
Bos et al. (2009), pregnant women	331	29.70	Female	Preselected	Portuguese	.88	PC	Varimax	3	47.90
Bos et al. (2009), new mothers	354	30.60	Female	Preselected	Portuguese	.89	PC	Varimax	3	49.40
Brown et al. (2001)	111	44.10	Mixed	Preselected	English	.89	PA	Promax	2	35.72
Buckley et al. (2001)	416	45	Male	Preselected	English	.91	NA	NA	NA	NA
Byrne et al. (2004)	486	15	Mixed	General	Chinese	NA	EPI	Oblimin	3	NA
Cameron (2008)	783	60.90	Mixed	Preselected	English	NA	NA	Promax	2	NA
Campos and Gonçalves (2011)	547	20.65	Mixed	General	Portuguese	.90	PA	Promax	2	NA
Carmody (2005)	502	20.18	Mixed	General	English	.92	NA	NA	NA	NA
Chang (2005), males	481	19.64	Male	General	Chinese	.86	PC	Oblimin	2	NA
Chang (2005), females	612	19.64	Female	General	Chinese	.86	PC	Oblimin	2	NA
Cole et al. (2003)	101	40.80	Mixed	Clinical	English	.95	ML	Oblimin	2	53.06
Cooper (2010), African American women	70	25.34	Female	General	English	.90	PC	Oblimin	3	52.80
Cooper (2010), non–African American women	117	25.34	Female	General	English	.89	PC	Oblimin	2	43
Dozois et al. (1998), male college students	157	21	Male	General	English	.92	PA	Oblique	2	49
Dozois et al. (1998), female college students	328	21	Female	General	English	.91	PA	Oblique	2	45
Dum, Pickren, Sobell, and Sobell (2008)	108	38	Mixed	Preselected	English	.95	PC	Promax	3	63.13
Gary and Yarandi (2004)	206	54	Female	Preselected	English	.91	PA	Promax	2	89
Grothe et al. (2005)	220	49.26	Mixed	Preselected	English	.90	NA	NA	NA	NA
Jakšić, Ivezić, Jokić-Begić, Surányi, and Stojanović-Špehar (2013)	314	55.01	Mixed	Preselected	Croatian	.94	PA	Promax	2	47.34
Johnson, Neal, Brems, and Fisher (2006)	598	38.20	Mixed	Preselected	English	.92	NA	NA	NA	NA
Kapci et al. (2008), normal adults	203	32.68	Mixed	General	Turkish	.90	PA	Promax	2	42.61
Kapci et al. (2008), outpatients	176	36.96	Mixed	Clinical	Turkish	.89	PA	Promax	2	40.72
Kneipp, Kairalla, Stacciarini, and Pereira (2009)	108	28.90	Female	Preselected	English	.94	PA	Promax	2	NA
Kojima et al. (2002)	766	50	Mixed	General	Japanese	.87	PC	Promax	2	NA
Lindsay and Skene (2007)	108	34.70	Mixed	Preselected	English	.90	PC	Quartimax	3	61.16
Lofton (2011)	214	46	Mixed	Preselected	English	NA	PC	Oblimin	5	60.48
Lopez (2011)	387	19.50	Mixed	General	English	NA	PC	Promax	2	35.14
Lopez, Pierce, Gardner, and Hanson (2013)	345	53.26	Male	Preselected	English	.91	PA	Promax	3	NA
Manian, Schmidt, Bornstein, and Martinez (2013)	478	31.82	Female	Preselected	English	NA	ML	Varimax	3	49.10
Mahmud, Awang, Herman, and Mohamed (2004)	354	29.56	Female	Preselected	Malay	.89	PC	Varimax	3	43.56
Osman, Barrios, Gutierrez, Williams, and Bailey (2008)	414	15.77	Mixed	General	English	.92	NA	NA	NA	NA
Osman, Kopper, Barrios, Gutierrez, and Bagge (2004), males	205	15.08	Male	Clinical	English	.91	ML	Promax	2	37.93
Osman et al. (2004), females	203	15.09	Female	Clinical	English	.93	ML	Promax	2	45.32
Palmer and Binks (2008)	117	19.08	Male	Preselected	English	.90	PC	Varimax	2	35.56
Patterson et al. (2011)	341	52.80	Mixed	Preselected	English	NA	ML	Promax	2	NA
Penley (2001), hemodialysis patients	122	62.50	Mixed	Preselected	Spanish	.92	PA	Promax	2	45.16
Penley (2001), students completing Spanish version	276	19.90	Mixed	General	Spanish	.91	PA	Promax	2	42.82
Penley (2001), students completing English version	634	19.90	Mixed	General	English	.90	PA	Promax	2	39.58
Poole, Bramwell, and Murphy (2006)	540	46.73	Mixed	Preselected	English	.92	PA	Oblimin	2	NA
Rodríguez-Gómez, Dávila-Martínez, and Collazo-Rodríguez (2006)	410	75.60	Mixed	General	Spanish	.89	PC	Varimax	4	52.83
Rowland, Lam, and Leahy (2005)	51	40.30	Mixed	Preselected	English	NA	PC	Varimax	3	50.53
Segal et al. (2008), young adults	229	19.60	Mixed	General	English	.92	PC	NA	1	38.48
Segal et al. (2008), old adults	147	70.30	Mixed	General	English	.86	PC	NA	1	30.24
Siegert, Walkey, and Turner-Stokes (2009)	353	43.80	Mixed	Preselected	English	.89	PC	Varimax	2	NA
Spencer (2004)	85	35.40	Male	Preselected	English	.89	PC	Varimax	6	65.30
Steele and Edwards (2008)	122	35.47	Mixed	Preselected	Xhosa	.93	PC	Oblique	2	NA
Steer and Clark (1997)	160	18.76	Mixed	General	English	.89	PA	Promax	2	NA
Steer et al. (1999)	210	41.29	Mixed	Clinical	English	.90	ML	Promax	2	NA
Steer et al. (1998)	210	15.17	Mixed	Clinical	English	.92	PA	Promax	3	NA
Uslu, Kapci, Oncu, Ugurlu, and Turkcapar (2008), general adolescents	300	15.72	Mixed	General	Turkish	.90	PA	Promax	2	NA
Uslu et al. (2008), clinical adolescents	166	16.10	Mixed	Clinical	Turkish	.90	PA	Promax	2	NA
VanVoorhis and Blumentritt (2007)	131	15.50	Mixed	Preselected	English	.90	PA	Promax	2	42.7
Viljoen, Iverson, Griffiths, and Woodward (2003)	127	36.50	Mixed	Preselected	English	NA	NA	Oblimin	2	53
Whisman, Perez, and Ramel (2000)	576	18.64	Mixed	General	English	.89	NA	NA	NA	NA
Wu and Chang (2008)	2,095	16.76	Mixed	General	Chinese	NA	PC	Unknown	2	NA
Zenger (2009)	96	37	Mixed	Preselected	English	.90	NA	Oblimin	3	52

Note. NA = not available/applicable. Extraction method, PA = principal axis; PC = principal component; EPI = equal prior instant communalities extraction; ML = maximum likelihood.

Six samples (Buckley et al., 2001; Carmody, 2005; Grothe et al., 2005; Johnson et al., 2006; Osman et al., 2008; Whisman et al., 2000) did not conduct exploratory factor analyses but provided correlation matrices. Three samples did not report the extraction method. Of the remaining 53 samples, 24 used principal axis analysis, 22 used principal components analysis, 6 used maximum likelihood, and 1 used equal prior instant communalities.

Of 56 samples providing information about exploratory analyses, 2 samples retained one factor, 37 samples retained two factors, 13 samples retained three factors, 2 samples retained four factors, 1 sample retained five factors, and 1 sample retained six factors. For the 54 samples that retained more than one factor, 30 used Promax rotation, 10 used Oblimin rotation, 9 used Varimax rotation, 1 used Quartimax rotation, 3 used some oblique rotation, and 1 sample did not report its rotation method. Of the 56 samples providing information about exploratory factor analyses, 20 did not report the percentage of variance accounted for by the retained factors. In the remaining 36 studies, mean variance explained was 47.96% (range = 30.24% to 89%).

As mentioned, pattern matrices were not reported in six samples. After principal axis analyses, Promax rotation was performed to examine the factor structure for these six samples. Based on the criterion proposed by Conway and Huffcutt (2003) and Henson and Roberts (2006), the two-factor solution was supported for studies by Grothe et al. (2005), Johnson et al. (2006), and Osman et al. (2008). The three-factor solution best represented data in the studies by Buckley et al. (2001), Carmody (2005), and Whisman et al. (2000).

Factor Analysis Results for the Full Sample

The upper diagonal of Table 2 is the raw co-occurrence matrix for BDI-II items. The best agreement between studies was for the relation between Items 20 and 15. Items of Tiredness or Fatigue and Loss of Energy were related in 56 of 62 samples. Exploratory factor analysis was performed using the sample-size-weighted mean co-occurrence matrix in the lower diagonal of Table 2.

Table 2.

Raw Co-Occurrence Matrix and Sample-Size-Weighted Standardized Co-Occurrence Matrix.

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
	41	31	28	29	31	37	31	40	26	18	28	27	37	17	11	20	11	12	15	10
.58		43	17	36	38	46	39	45	24	11	17	25	47	9	6	10	4	6	7	7
.46	.67		16	46	45	50	48	39	22	11	14	25	46	4	3	9	4	3	2	5
.39	.21	.19		14	13	19	18	18	15	19	44	27	18	38	27	25	24	29	35	27
.38	.51	.68	.15		43	45	45	35	22	10	14	22	40	5	5	11	5	4	3	4
.47	.60	.72	.15	.61		44	41	37	26	13	14	21	42	5	4	10	4	4	3	5
.53	.71	.80	.23	.66	.70		51	45	23	11	16	25	52	7	5	12	6	5	5	5
.45	.62	.74	.21	.67	.64	.78		38	25	14	15	25	45	4	5	12	5	4	2	4
.56	.65	.62	.21	.52	.60	.68	.59		25	12	15	24	46	8	4	13	5	5	6	6
.37	.32	.27	.19	.26	.35	.30	.28	.33		21	15	23	22	9	11	17	11	9	8	5
.26	.14	.11	.27	.10	.14	.12	.14	.14	.27		24	25	10	22	24	28	24	23	23	10
.37	.19	.15	.67	.13	.14	.18	.17	.16	.18	.34		30	17	37	28	28	28	30	34	21
.37	.33	.33	.41	.29	.32	.35	.33	.33	.31	.35	.44		26	25	21	23	22	26	24	14
.47	.66	.70	.24	.58	.65	.85	.69	.64	.27	.12	.22	.35		9	4	11	5	6	7	8
.21	.08	.03	.54	.04	.04	.06	.03	.07	.12	.33	.54	.39	.11		43	32	41	51	56	28
.14	.06	.02	.43	.03	.03	.04	.03	.04	.16	.35	.45	.35	.06	.71		28	48	43	48	25
.25	.09	.08	.38	.11	.11	.11	.12	.12	.22	.37	.43	.33	.09	.47	.45		27	28	31	16
.14	.04	.03	.38	.03	.03	.05	.03	.05	.16	.32	.43	.33	.07	.67	.82	.40		43	46	21
.15	.06	.02	.44	.03	.03	.04	.03	.04	.10	.31	.44	.38	.08	.76	.65	.40	.65		51	26
.19	.69	.02	.52	.02	.03	.05	.02	.05	.12	.34	.52	.38	.09	.84	.84	.45	.80	.76		30
.11	.06	.06	.38	.05	.06	.06	.05	.06	.04	.12	.32	.23	.12	.44	.38	.24	.32	.42	.45

Note. The raw co-occurrence matrix is displayed in the upper diagonal and the sample size-weighted co-occurrence matrix is presented in the lower diagonal.

As most studies retained two or three factors, the two-factor and three-factor solutions were explored. Item 10 (Crying) did not have a salient (≥.4) loading in the two-factor solution, while Items 10 (Crying), 11 (Agitation), 13 (Indecisiveness), 17 (Irritability), and 21 (Loss of Interest in Sex) had no salient loadings in the three-factor solution. As the two-factor solution better approximated a simple structure, the two-factor solution was more acceptable than the three-factor solution. Table 3 presents the Promax-rotated pattern matrix for the two-factor solution.

Table 3.

Promax-Rotated Factor Structure for the BDI-II.

	Full sample		English		Non-English		General		At-risk/clinical		Youths/college students		Adults
	F1	F2	F1	F2	F1	F2	F1	F2	F1	F2	F1	F2	F1	F2
Sadness	.59	.20	.61	.18	.56	.25	.56	.14	.62	.28	.53	.20	.65	.22
Pessimism	.78	.01	.92	−.03	.57	.09	.69	.00	.88	.03	.72	−.02	.84	.04
Past Failure	.87	−.06	.94	−.04	.76	−.08	.81	−.04	.92	−.07	.82	−.04	.91	−.07
Loss of Pleasure	.18	.60	.20	.62	.16	.58	.19	.56	.18	.64	.24	.47	.15	.71
Guilty Feelings	.73	−.03	.84	−.02	.55	−.06	.62	−.02	.83	−.03	.59	−.02	.86	−.04
Punishment Feelings	.80	−.03	.86	.01	.71	−.09	.72	.01	.87	−.07	.64	.02	.93	−.07
Self-Dislike	.92	−.03	.94	−.03	.89	−.03	.92	−.05	.92	−.01	.91	−.04	.93	−.03
Self-Criticalness	.83	−.04	.90	−.03	.72	−.03	.79	−.02	.86	−.04	.77	−.02	.88	−.05
Suicidal Thoughts or Wishes	.76	.00	.79	−.00	.72	.02	.73	−.02	.80	.03	.71	.01	.81	−.00
Crying	.36	.16	.49	.11	.15	.25	.24	.22	.49	.10	.29	.21	.43	.11
Agitation	.12	.42	.15	.40	.07	.45	.07	.45	.17	.39	.08	.49	.17	.35
Loss of Interest	.15	.62	.11	.68	.21	.55	.13	.56	.18	.68	.16	.50	.15	.72
Indecisiveness	.36	.44	.38	.47	.33	.40	.39	.35	.34	.53	.44	.27	.30	.58
Worthlessness	.83	.02	.95	−.02	.65	.09	.73	.10	.94	−.05	.85	−.04	.82	.07
Loss of Energy	−.06	.88	−.05	.95	−.07	.78	−.05	.82	−.06	.94	−.02	.79	−.07	.96
Changes in Sleeping Pattern	−.09	.85	−.08	.89	−.12	.80	−.10	.85	−.07	.87	−.09	.84	−.07	.86
Irritability	.07	.54	.10	.65	.01	.41	.12	.41	.03	.68	.15	.49	.02	.58
Changes in Appetite	−.08	.80	−.06	.84	−.12	.76	−.09	.82	−.07	.80	−.09	.86	−.05	.76
Concentration Difficulty	−.07	.80	−.08	.91	−.06	.63	−.04	.66	−.10	.94	−.04	.62	−.08	.96
Tiredness or Fatigue	−.10	.94	−.09	.97	−.12	.91	−.09	.93	−.11	.96	−.08	.93	−.09	.96
Loss in Interest in Sex	.00	.47	−.04	.56	.08	.33	.08	.23	−.06	.70	.07	.08	−.05	.80

Note. Factor loadings equal to or greater than .40 are in boldface.

In the two-factor solution, the first two eigenvalues were 6.23 and 5.58. The correlation coefficient between these two factors was .17, much lower than that reported in the study by Beck et al. (1996). Factor 1 was related to nine depressive symptoms: Sadness, Pessimism, Past Failure, Guilty Feelings, Punishment Feelings, Self-Dislike, Self-Criticalness, Suicidal Thoughts or Wishes, and Worthlessness. This factor was called the Cognitive factor by Beck et al. (1996). Factor 2 consisted of 11 depressive symptoms. This factor was similar to the Somatic-Affective factor in the study by Beck et al. (1996) except that the Somatic-Affective factor in Beck et al. (1996) was also defined by the symptom of Crying. In summary, the factor structure for the full sample was similar to that for psychiatric outpatients in the study by Beck et al. (1996).

Comparison of Factor Analysis Between English and Non-English BDI-II Versions

As the two-factor solution better approximated a simple structure for the full sample, the two-factor solution was explored for each subgroup. Forty-two samples were administered the English BDI-II, involving 11,536 participants. The first two eigenvalues were 7.46 and 6.37. The correlation between these two factors was .15. Factor 1 consisted of 10 items, and was labeled Negative Attitudes by Harris and D’Eon (2008). Factor 2 was similar to the Somatic-Affective factor in the study by Beck et al. (1996).

Twenty samples used non-English versions of the BDI-II, consisting of 8,939 participants. The corresponding eigenvalues for non-English versions were 4.75 and 4.67. The correlation coefficient between the two factors was .19. Factor 1 resembled the Cognitive factor identified for the full sample. Factor 2 was similar to the Somatic-Affective factor identified for the full sample except Factor 2 was not defined by Loss of Interest in Sex. In sum, the factor structure for the non-English version was quite similar to that for the full sample.

Comparison of Factor Analysis Between the General and At-Risk/Clinical Samples

Twenty-five samples involving 10, 841 participants were selected from the general population. The first two eigenvalues were 5.38 and 4.80. The correlation coefficient between these two factors was .16. Factor 1 resembled the Cognitive factor identified in the full sample. Factor 2 was similar to the Somatic-Affective factor identified in the full sample except Factor 2 was not defined by Indecisiveness and Loss of Interest in Sex.

Thirty-seven samples consisting of 9,634 participants were selected from at-risk/clinical populations. The first two eigenvalues were 7.22 and 6.62. The correlation of these two factors was .15. The factor structure for at-risk/clinical samples was similar to that of the English BDI-II except the item of Agitation did not have a salient loading.

Comparison of Factor Analysis Between Youths/College Students and Adults

The main life tasks for youths and college students are preparation for career and marriage, while those for adults were engagement in career and marriage. As the life demands between youths/college students and adults were different, the age groups were categorized into youths/college students versus adults. Twenty-seven samples were selected from youths/college students (N = 10,347). The first two eigenvalues were 5.46 and 4.60. The correlation between these two factors was .14. Factor 1 was similar to the Cognitive factor identified in the full sample except that the Cognitive factor in the full sample was not defined by Indecisiveness. Factor 2 resembled the Somatic-Affective factor identified in the general samples.

Thirty-five samples consisting of 10,128 participants were selected from adults. The first two eigenvalues were 7.09 and 6.85. The correlation between these two factors was .16. The factor structure of the BDI-II for adults was similar to that of the English BDI-II.

Study 2

Method

Literature Search

Using search procedures described in Study 1, 13 studies were identified that reported covariance matrices among the BDI-II items. Beck et al. (1996) had two samples and Penley (2001) had three samples. In total, 16 independent samples involving 5,128 participants were used to examine the factor structure via covariance structure analysis.

Models

The following models were tested using LISREL 8 (Jöreskog & Sörbom, 1993)

Model 1, the two-correlated-factor model, identified in the full sample with a Cognitive factor (defined by Items 1-3, 5-9, and 14) and a Somatic-Affective factor (defined by Items 4, 11-13, and 15-21).

Model 2, the two-correlated-factor model, identified in the clinical samples in the study by Beck et al. (1996), had 12 items loading on a Somatic-Affective factor (Items 4, 10-13, and 15-21), and 9 items loading on a Cognitive factor.

Model 3, the two-correlated-factor model, identified in Dozois et al. (1998), had a Cognitive-Affective factor consisting of 10 items (Items 1-3, 5-9, 13, and 14) and a Somatic-Vegetative factor defined by 11 items (Items 4, 10-12, and 15-21).

Model 4 was the two-correlated-factor model identified by Steer et al. (1999) with a Cognitive factor defined by Items 2, 3, 5 to 9, and 14 and a non-Cognitive factor defined by the remaining items.

Model 5 was the three-correlated-factor model tested in Beck, Steer, Brown, and van der Does (2002), with a Cognitive factor consisting of seven items (Items 3, 5-8, 13, and 14), a Somatic factor consisting of nine items (Items 10, 11, and 15-21), and an Affective factor defined by five items (Items 1, 2, 4, 9, and 12).

Model 6 was the three-correlated-factor model used in the study by Buckley et al. (2001), with a Cognitive factor defined by nine items (Items 1-3, 5-9, and 14), an Affective factor defined by four items (Items 4, 10, 12, and 13), and a Somatic factor consisting of eight items (11 and 15-21).

Model 7 was the three-correlated-factor model tested by Osman et al. (1997), which had 10 items loading on Negative Attitudes (Items 1-3, 5-10, and 14), seven items loading on a Performance Difficulty factor (Items 4, 12, 13, 15, 17, 19, and 20), and four items loading Somatic Elements (Items 11, 16, 18, and 21).

As the general depression factor was proposed in previous research (Brouwer et al., 2013; Ward, 2006), an alternative set of models with a general depression factor was added to these seven models. For example, the alternative model for Model 1 was a model with a general factor defined by all items, Cognitive and Somatic-Affective factors.

To assess model adequacy, χ² was derived to measure the difference between an a priori model and data. A significant χ² value indicates that the tested model deviates from data and that a hypothesized model should be rejected. One limitation associated with the χ² value is dependency on sample size. Hence, the Tucker–Lewis index, comparative fit index (CFI), and root mean square error of approximation (RMSEA) were applied. A value of .90 or greater for TLI indicates an acceptable fit from a practical stand point (Bentler, 1990). This is also the case for the CFI (Bentler, 1990). Conversely, an RMSEA value of .06 or less indicates good model fit (Hu & Bentler, 1999).

Results

Confirmatory factor analyses were performed using the sample-size-weighted mean covariance matrix (Table 4). Table 5 presents fit indices of all models. Although χ² values for all models were statistically significant, comparative fit indices indicate that models fit well. Analytical results also show that the fit of the two-factor models (Models 1-4) was comparable to that of the three-factor models (Models 5-7). As the two-factor model was more parsimonious, it was considered superior. Although the fit of the alternative set of models (Models 1a-7a) to data was good, these models lacked empirical identification (Brunner, Nagy, & Wilhelm, 2012). Specifically, the covariance between the general and one of the depression subfactors was not identified in all seven alternative models; thus, alternative models were not considered.

Table 4.

Sample-Size-Weighted Mean Covariance Matrix.

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
.55
.25	.56
.20	.25	.64
.24	.22	.18	.57
.18	.16	.22	.17	.51
.23	.23	.27	.21	.24	.94
.26	.27	.32	.26	.24	.26	.76
.25	.25	.29	.24	.27	.28	.38	.77
.16	.15	.15	.13	.12	.14	.17	.16	.31
.29	.21	.19	.23	.20	.27	.26	.29	.17	.95
.20	.19	.17	.19	.16	.20	.20	.23	.11	.24	.68
.25	.25	.20	.32	.18	.23	.27	.24	.14	.27	.22	.70
.24	.24	.23	.25	.22	.24	.29	.29	.15	.26	.23	.31	.73
.25	.27	.30	.24	.20	.26	.33	.31	.18	.23	.18	.26	.28	.56
.22	.19	.16	.24	.13	.15	.22	.19	.11	.21	.16	.27	.24	.21	.57
.18	.16	.14	.19	.13	.16	.19	.20	.11	.21	.20	.22	.22	.16	.23	.79
.24	.21	.18	.22	.18	.21	.24	.23	.13	.25	.27	.28	.26	.22	.23	.20	.65
.20	.16	.14	.20	.13	.18	.20	.21	.11	.23	.21	.25	.24	.18	.22	.29	.22	.79
.23	.21	.21	.24	.17	.22	.26	.25	.13	.25	.25	.29	.34	.24	.26	.24	.26	.25	.66
.24	.20	.18	.26	.15	.19	.23	.21	.12	.23	.21	.30	.27	.23	.38	.29	.27	.27	.31	.70
.13	.11	.09	.17	.10	.10	.13	.11	.08	.14	.10	.18	.15	.13	.14	.13	.13	.14	.14	.18	.61

Table 5.

Fit Indices.

Model	df	χ²	RMSEA	TLI	CFI
Model 1	169	2397.76**	.05	.98	.98
Model 2	188	2679.38**	.05	.98	.98
Model 3	188	2847.71**	.06	.98	.98
Model 4	188	2662.91**	.05	.98	.98
Model 5	186	2791.81**	.06	.98	.98
Model 6	186	2368.30**	.05	.98	.98
Model 7	186	2598.38**	.05	.98	.98
Model 1a	166	1482.59**	.04	.99	.99
Model 2a	165	1478.59**	.04	.99	.99
Model 3a	165	1459.15**	.04	.99	.99
Model 4a	165	1491.23**	.04	.99	.99
Model 5a	162	1378.17**	.04	.99	.99
Model 6a	162	1334.36**	.04	.99	.99
Model 7a	162	1347.47**	.04	.99	.99
One-factor model	189	4256.17**	.07	.96	.97

Note. RMSEA = root mean square error of approximation; TLI = Tucker–Lewis index; CFI = comparative fit index. Models 1a to 7a were alternative models with a general depression factor and depression subfactors.

p < .01.

The correlation between latent constructs was strong in Models 1 to 4, demonstrating the existence of a considerable amount of shared variance (Table 6). To address the issue of highly correlated latent constructs, the one-factor model with a general depression factor was tested. The χ² value was 4,256.17, df = 189. Although this χ² value was statistically significant, the RMSEA, TLI, and CFI values were good at .07, .96, and .97, respectively, providing support for the one general depression factor from a practical standpoint.

Table 6.

Factor Loadings, Reliability Estimates, and Correlation Between Latent Constructs.

	Model 1		Model 2		Model 3		Model 4		Model 5			Model 6			Model 7			1-f
	C	SA	C	SA	CA	SV	NC	C	A	C	S	C	A	S	NA	PD	SE	G
Sadness	.48		.48		.48		.47		.49			.48			.48			.48
Pessimism	.48		.48		.48			.48	.47			.48			.48			.46
Past Failure	.50		.50		.49			.51		.50		.50			.49			.45
Loss of Pleasure		.49		.49		.49	.49		.49				.49			.49		.48
Guilty Feelings	.40		.40		.40			.41		.41		.40			.40			.38
Punishment Feelings	.49		.49		.49			.49		.49		.49			.49			.46
Self-Dislike	.59		.59		.59			.60		.60		.59			.59			.55
Self-Criticalness	.57		.57		.56			.57		.58		.57			.57			.53
Suicidal Thoughts and Wishes	.31		.31		.31			.31	.30			.31			.31			.29
Crying				.49		.49	.50				.50		.50		.50			.50
Agitation		.42		.43		.43	.43				.43			.43			.44	.42
Loss of Interest		.56		.56		.56	.56		.55				.56			.57		.54
Indecisiveness		.55		.55	.53		.54			.53			.55			.55		.54
Worthlessness	.55		.55		.55			.56		.55		.55			.55			.51
Loss of Energy		.49		.49		.49	.48				.50			.50		.49		.45
Changes in Sleeping Pattern		.44		.44		.44	.43				.45			.46			.47	.41
Irritability		.49		.49		.49	.49				.50			.49		.49		.48
Changes in Appetite		.46		.45		.46	.45				.46			.47			.49	.43
Concentration Difficulty		.54		.54		.53	.53				.54			.54		.54		.51
Tiredness or Fatigue		.57		.56		.57	.55				.58			.59		.56		.51
Loss in Interest in Sex		.29		.29		.29	.29				.29			.29			.30	.28
Reliability (ω)	.70	.77	.68	.79	.73	.75	.81	.63	.37	.60	.67	.68	.33	.62	.73	.60	.24	.91
r between latent constructs	.81		.83		.83		.83		r_C,S = .78; r_C,A = .90; r_S,A = .89			r_C,S = .76; r_C,A = .88; r_S,A = .93			r_NA,PD = .83; r_NA,SE = .76; r_PD,SE= .93

Note. C = a Cognitive factor; SA = a Somatic-Affective factor; SA = a Somatic-Affective factor; CA = a Cognitive-Affective factor; SV = a Somatic-Vegetative factor; NC = a non-Cognitive factor; A = an Affective factor; S = a Somatic factor; ND = a Negative Attitude factor; PD = a Performance Difficulty factor; SE = a Somatic Element factor; 1-f = the one-factor model; G = a General Depression factor.

All factor loadings (Table 6) were statistically significant, indicating that each latent construct was well defined and each item was influenced by its corresponding latent construct. Item 21, Loss of Interest in Sex, had the lowest factor loading. The latent construct accounts for roughly 12% to15% of variance in this item. How well the scale score represented the latent construct, indicated by ω, ranged from .24 to .79. That is, the proportion of variance accounted for by a latent construct was 24% to 79%. Thus, reliability of the depression subfactors varied markedly.

Discussion

The BDI-II is one of most popular inventories of depression. Given the widespread use of the BDI-II, examining its factor structure is important. This study examined the dimensionality of the BDI-II. Meta-analyses were used to combine findings of pattern and covariance matrices. On the basis of studies that conducted exploratory factor analysis, the BDI-II consists of Cognitive and Somatic-Affective factors. The two-factor structure identified in this study was similar to that obtained in the clinical study by Beck et al. (1996). This finding is inconsistent with the three-factor structure found by Shafer (2006). The factors of Performance Impairment and Somatic Concerns in Shafer (2006) combined to form the Somatic-Affective factor. The factor of Negative Attitudes was similar to the Cognitive factor.

Based on confirmatory factor analysis results of combined covariance matrices, the two-correlated-factor model was the best representation of data. All four two-correlated-factor models had a Cognitive or Cognitive-Affective factor defined by the core item Self-Dislike. The other factor in the four two-factor models was a Somatic-Affective, Somatic-Vegetative, or non-Cognitive factor defined by the core item Tiredness or Fatigue. The Somatic factor was also supported in the study by Shafer (2006).

Given the importance of the general depression factor proposed in the literature and the high correlation between latent constructs, models with a general depression factor and depression subfactors were explored. These models did not achieve empirical identification, likely due to factor overextraction (Chen, West, & Sousa, 2006; Rindskopf, 1984). Hence, the bifactor models with general and specific depression factors were not considered. To test the possible existence of a common depression factor, the one-factor model assuming all correlations between depression subfactors are equal to 1 was tested. The good fit of the one-factor model supported the existence of a general depression factor (Brouwer et al., 2013). The finding of a general depression factor does not necessarily invalidate the differentiation of specific depression factors as a different depression subfactor was reportedly related to different etiologies and external criteria (Abramson, Metalsky, & Alloy, 1989; Alloy, Just, & Panzarella, 1997; Payne, Palmer, & Joffe, 2009).

Based on exploratory factor analyses, refinement of the items Crying, Agitation, Indecisiveness, and Loss of Interest in Sex should be considered as these items did not have salient loadings in the full sample or some study subgroups. As these items were not a good depression marker, they should be revised in future BDI-II versions. On the basis of confirmatory factor analysis, refinement of the item Loss of Interest in Sex was further supported due to it having the lowest reliability among the 21 items.

The effect of language version on the factor structure was noted by Kim et al. (2013); therefore, the factor structures between English and non-English versions of the BDI-II were compared. Although the two-factor solutions were supported for these versions, factor patterns differed. Hence, comparison of mean depression across language versions revealed the true group differences among measured constructs and constructs with irrelevant variance. Culturally specific item adaptations of BDI-II items are recommended in future versions. For example, Bravo, Canino, Rubio-Stipec, and Woodbury-Fariña (1991) suggest a comprehensive model of translation and cultural adaption to assure sematic, content, technical, criterion, and conceptual equivalence. As such, the item wordings can be relevant to different cultures. The moderating effects of population type and age group were also supported. Nevertheless, invariance of a factor structure across subgroups of studies using confirmatory factor analysis was not tested in this study due to the small number of data points.

Although the weighted mean co-occurrence matrices were computed respectively for males and females, these matrices were not positive definite. The gender effect was thus not examined. Future research should address this issue. The effect of ethnicity on the factor structure of depression was explored in a previous meta-analysis (Kim, DeCoster, Huang, & Chiriboga, 2011). Examining the ethnicity effect with the few data available to this meta-analysis is not feasible. Future research should address this issue as mean depression scores were frequently compared across ethnicities in primary research (Cooper, 2010; Penley, 2001).

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by National Science Council of the Republic of China, Taiwan under Grant No. NSC 101-2410-H-018-017.

References

Abramson

L. Y.

Metalsky

G. I.

Alloy

L. B.

(1989). Hopelessness depression: A theory-based subtype of depression. Psychological Review, 96, 358-372. doi:10.1037/0033-295X.96.2.358

Alloy

L. B.

Just

Panzarella

(1997). Attributional style, daily life events, and hopelessness depression: Subtype validation by prospective variability and specificity of symptoms. Cognitive Therapy and Research, 21, 321-344. doi:10.1023/A:1021878516875

*Al-Musawi

N. M.

(2001). Psychometric properties of the Beck Depression Inventory-II with university students in Bahrain. Journal of Personality Assessment, 77, 568-579. doi:10.1207/S15327752JPA7703_13

Al-Turkait

Ohaeri

J. U.

(2012). Exploratory factor analytical study of depressive symptomatology: An Arab experience with a sample comprising college students, using the revised Beck Depression Inventory. Annals of Saudi Medicine, 32, 19-26.

Al-Turkait

F. A.

Ohaeri

J. U.

(2010). Dimensional and hierarchical models of depression using the Beck Depression Inventory-II in an Arab college student sample. BMC Psychiatry, 10, 60. doi:10.1186/1471-244X-10-60

American Psychiatric Association. (1994). The diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.

Arnarson

Þ. Ö.

Ólason

D. Þ.

Smári

Sigurđsson

J. F.

(2008). The Beck Depression Inventory Second Edition (BDI-II): Psychometric properties in Icelandic student and patient populations. Nordic Journal of Psychiatry, 62, 360-365. doi:10.1080/08039480801962681

*Arnau

R. C.

Meagher

M. W.

Norris

M. P.

Bramson

(2001). Psychometric evaluation of the Beck Depression Inventory-II with primary care medical patients. Health Psychology, 20, 112-119. doi:10.1037/0278-6133.20.2.112

*Beck

A. T.

Steer

R. A.

Brown

G. K.

(1996). Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation.

10.

Beck

A. T.

Steer

R. A.

Brown

G. K.

van der Does

A. J. W.

(2002). BDI-II-Dutch manual. Lisse, Netherlands: Psychological Corporation.

11.

*Bedi

R. P.

(2001). The dimensionality of the Beck Depression Inventory-II in a sample of women diagnosed with major depressive disorder (Master’s thesis). Available from ProQuest Dissertations and Theses database. (UMI No. MQ61531)

12.

Bentler

P. M.

(1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238-246. doi:10.1037/0033-2909.107.2.238

13.

Beretvas

S. N.

Furlow

C. F.

(2006). Evaluations of an approximate method for synthesizing covariance matrices for use in meta-analytic SEM. Structural Equation Modeling, 13, 153-185. doi:10.1207/s15328007sem1302_1

14.

*Bos

S. C.

Pereira

A. T.

Marques

Maia

Soares

M. J.

Valente

. . . Azevedo

M. H.

(2009). The BDI-II factor structure in pregnancy and postpartum: Two or three factors? European Psychiatry, 24, 334-340.

15.

Bravo

Canino

G. J.

Rubio-Stipec

Woodbury-Fariña

(1991). A cross-cultural adaptation of a psychiatric epidemiologic instrument: The diagnostic interview schedule’s adaptation in Puerto Rico. Culture, Medicine and Psychiatry, 15, 1-18. doi:10.1007/BF00050825

16.

Brouwer

Meijer

R. R.

Zevalkink

(2013). On the factor structure of the Beck Depression Inventory-II: G is the key. Psychological Assessment, 25, 136-145. doi:10.1037/a0029228

17.

*Brown

Kaplan

Jason

(2011). Factor analysis of the Beck Depression Inventory-II with patients with chronic fatigue syndrome. Journal of Health Psychology, 17, 799-808. doi:10.1177/1359105311424470

18.

Brunner

Nagy

Wilhelm

(2012). A tutorial on hierarchically structured constructs. Journal of Personality, 80, 796-846. doi:10.1111/j.1467-6494.2011.00749.x

19.

*Buckley

T. C.

Parker

J. D.

Heggie

(2001). A psychometric evaluation of the BDI-II in treatment-seeking substance abusers. Journal of Substance Abuse Treatment, 20, 197-204. doi:10.1016/S0740-5472(00)00169-0

20.

*Byrne

B. M.

Stewart

S. M.

Lee

P. W. H.

(2004). Validating the Beck Depression Inventory-II for Hong Kong community adolescents. International Journal of Testing, 4, 199-216. doi:10.1207/s15327574ijt0403_1

21.

*Cameron

J. J.

(2008). Depressive symptoms, depressive symptom types, and coping strategies among cardiac rehabilitation participants (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3316456)

22.

*Campos

R. C.

Gonçalves

(2011). The Portuguese version of the Beck Depression Inventory-II (BDI-II): Preliminary psychometric data with two nonclinical samples. European Journal of Psychological Assessment, 27, 258-264. doi:10.1027/1015-5759/a000072

23.

*Carmody

D. P.

(2005). Psychometric characteristics of the Beck Depression Inventory-II with college students of diverse ethnicity. International Journal of Psychiatry in Clinical Practice, 9, 22-28. doi:10.1080/13651500510014800

24.

*Chang

(2005). Dimensions of the Chinese Beck Depression Inventory-II in a university sample. Individual Differences Research, 3, 193-199.

25.

Chen

F. F.

West

S. G.

Sousa

K. H.

(2006). A comparison of bifactor and second-order models of quality of life. Multivariate Behavioral Research, 41, 189-225. doi:10.1207/s15327906mbr4102_5

26.

*Cole

J. C.

Grossman

Prilliman

Hunsaker

(2003). Multimethod validation of the Beck Depression Inventory-II and Grossman-Cole Depression Inventory with an inpatient sample. Psychological Reports, 93, 1115-1129. doi:10.2466/PR0.93.8.1115-1129

27.

Conway

J. M.

Huffcutt

A. I.

(2003). A review and evaluation of exploratory factor analysis practices in organizational research. Organizational Research Methods, 6, 147-168. doi:10.1177/1094428103251541

28.

*Cooper

(2010). Depression among African American female college students: Exploratory factor analysis of the Beck Depression Inventory-II (Master’s thesis). Available from ProQuest Dissertations and Theses database. (UMI No. 1484858)

29.

Corbière

Bonneville-Roussy

Franche

R. L.

Coutu

M. F.

Choinière

Durand

M. J.

Boulanger

(2011). Further validation of the BDI-II among people with chronic pain originating from musculoskeletal disorders. Clinical Journal of Pain, 27, 62-69. doi:10.1097/AJP.0b013e3181eef903

30.

*Dozois

D. J. A.

Dobson

K. S.

Ahnberg

J. L.

(1998). A psychometric evaluation of the Beck Depression Inventory-II. Psychological Assessment, 10, 83-89. doi:10.1037/1040-3590.10.2.83

31.

*Dum

Pickren

Sobell

L. C.

Sobell

M. B.

(2008). Comparing the BDI-II and the PHQ-9 with outpatient substance abusers. Addictive Behaviors, 33, 381-387. doi:10.1016/j.addbeh.2007.09.017

32.

Gahr

C. S.

(2001). Evaluation of the psychometric properties of the Spanish version of the Beck Depression Inventory-II (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3029174)

33.

*Gary

F. A.

Yarandi

H. N.

(2004). Depression among southern rural African American women: A factor analysis of the Beck Depression Inventory-II. Nursing Research, 53, 251-259. doi:10.1097/00006199-200407000-00008

34.

*Grothe

K. B.

Dutton

G. R.

Jones

G. N.

Bodenlos

Ancona

Brantley

P. J.

(2005). Validation of the Beck Depression Inventory-II in a low-income African American sample of medical outpatients. Psychological Assessment, 17, 110-114. doi:10.1037/1040-3590.17.1.110

35.

Harris

C. A.

D’Eon

J. L.

(2008). Psychometric properties of the Beck Depression Inventory-Second Edition (BDI-II) in individuals with chronic pain. Pain, 137, 609-622. doi:10.1016/j.pain.2007.10.022

36.

Henson

R. K.

Roberts

J. K.

(2006). Use of exploratory factor analysis in published research: Common errors and some comment on improved practice. Educational and Psychological Measurement, 66, 393-416. doi:10.1177/0013164405282485

37.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55. doi:10.1080/10705519909540118

38.

*Jakšić

Ivezić

Jokić-Begić

Surányi

Stojanović-Špehar

(2013). Factorial and diagnostic validity of the Beck Depression Inventory-II (BDI-II) in Croatian primary health care. Journal of Clinical Psychology in Medical Settings, 20, 311-322. doi:10.1007/s10880-013-9363-2

39.

*Johnson

M. E.

Neal

D. B.

Brems

Fisher

D. G.

(2006). Depression as measured by the Beck Depression Inventory-II among injecting drug users. Assessment, 13, 168-177. doi:10.1177/1073191106286951

40.

Jöreskog

K. G.

Sörbom

(1993). LISREL VIII: A guide to the program and applications. Mooresville, IN: Scientific Software.

41.

*Kapci

E. G.

Uslu

Turkcapar

Karoglan

(2008). Beck Depression Inventory II: Evaluation of the psychometric properties and cut-off points in a Turkish adult population. Depression and Anxiety, 25, 104-110. doi:10.1002/da.20371

42.

Kim

DeCoster

Huang

C. H.

Bryant

A. N.

(2013). A meta-analysis of the factor structure of the Geriatric Depression Scale (GDS): The effects of language. International Psychogeriatrics, 25, 71-81. doi:10.1017/S1041610212001421

43.

Kim

DeCoster

Huang

C. H.

Chiriboga

D. A.

(2011). Race/ethnicity and the factor structure of the Center for Epidemiologic Studies Depression Scale: A meta-analysis. Cultural Diversity and Ethnic Minority Psychology, 17, 381-396. doi:10.1037/a0025434

44.

*Kneipp

S. M.

Kairalla

J. A.

Stacciarini

Pereira

(2009). The Beck Depression Inventory II factor structure among low-income women. Nursing Research, 58, 400-409. doi:10.1097/NNR.0b013e3181bee5aa

45.

*Kojima

Furukawa

T. A.

Takahashi

Kawai

Nagaya

Tokudome

(2002). Cross-cultural validation of the Beck Depression Inventory-II in Japan. Psychiatry Research, 110, 291-299. doi:10.1016/S0165-1781(02)00106-3

46.

*Lindsay

W. R.

Skene

D. D.

(2007). The Beck Depression Inventory II and the Beck Anxiety Inventory in people with intellectual disabilities: Factor analyses and group data. Journal of Applied Research in Intellectual Disabilities, 20, 401-408. doi:10.1111/j.1468-3148.2007.00380.x

47.

*Lofton

M. C. J. E.

(2011). Inflammation, altruism, religiousness, and the Beck Depression Inventory-II in urban African Americans (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3494867)

48.

*Lopez

M. D.

(2011). The associations between individualism, collectivism, acculturation, and depressive symptoms (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3495017)

49.

*Lopez

M. N.

Pierce

R. S.

Gardner

R. D.

Hanson

R. W.

(2013). Standardized Beck Depression Inventory-II scores for male veterans coping with chronic pain. Psychological Services, 10, 257-263. doi:10.1037/a0027920

50.

*Mahmud

W. M. R. M.

Awang

Herman

Mohamed

M. N.

(2004). Analysis of the psychometric properties of the Malay version of Beck Depression Inventory II (BDI-II) among postpartum women in Kedah, North West of Peninsular Malaysia. Malaysian Journal of Medical Sciences, 11, 19-25.

51.

Manian

Schmidt

Bornstein

M. H.

Martinez

(2013). Factor structure and clinical utility of BDI-II factor scores in postpartum women. Journal of Affective Disorders, 149, 259-268. doi:10.1016/j.jad.2013.01.039

52.

Ohayon

M. M.

Schatzberg

A. F.

(2010). Social phobia and depression: Prevalence and comorbidity. Journal of Psychosomatic Research, 68, 235-243. doi:10.1016/j.jpsy-chores.2009.07.018

53.

*Osman

Barrios

F. X.

Gutierrez

P. M.

Williams

J. E.

Bailey

(2008). Psychometric properties of the Beck Depression Inventory-II in nonclinical adolescent samples. Journal of Clinical Psychology, 64, 83-102. doi:10.1002/jclp.20433

54.

Osman

Downs

W. R.

Barrios

F. X.

Kopper

B. A.

Gutierrez

P. M.

Chiros

C. E.

(1997). Factor structure and psychometric characteristics of the Beck Depression Inventory-II. Journal of Psychopathology and Behavioral Assessment, 19, 359-376. doi:10.1007/BF02229026

55.

*Osman

Kopper

B. A.

Barrios

Gutierrez

P. M.

Bagge

C. L.

(2004). Reliability and validity of the Beck Depression Inventory-II. Psychological Assessment, 16, 120-132. doi:10.1037/1040-3590.16.2.120

56.

*Palmer

E. J.

Binks

(2008). Psychometric properties of the Beck Depression Inventory-II with incarcerated male offenders aged 18-21 years. Criminal Behaviour and Mental Health, 18, 232-242. doi:10.1002/cbm.701

57.

*Patterson

A. L.

Morasco

B. J.

Fuller

B. E.

Indest

D. W.

Loftis

J. M.

Hauser

(2011). Screening for depression in patients with hepatitis C using the Beck Depression Inventory-II: Do somatic symptoms compromise validity? General Hospital Psychiatry, 33, 354-362. doi:10.1016/j.genhosppsych.2011.04.005

58.

Payne

J. L.

Palmer

J. T.

Joffe

(2009). A reproductive subtype of depression: Conceptualizing models and moving toward etiology. Harvard Review of Psychiatry, 17, 72-86. doi:10.1080/10673220902899706

59.

*Penley

J. A.

(2001). Psychometric properties of the Spanish Beck Depression Inventory-II in a medical and a student sample (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3035100)

60.

*Poole

Bramwell

Murphy

(2006). Factor structure of the Beck Depression Inventory-II patients with chronic pain. Clinical Journal of Pain, 22, 790-798. doi:10.1097/01.ajp.0000210930.20322.93

61.

Quilty

L. C.

Zhang

K. A.

Bagby

R. M.

(2010). The latent symptom structure of the Beck Depression Inventory-II in outpatients with major depression. Psychological Assessment, 22, 603-608. doi:10.1037/a0019698

62.

Rindskopf

(1984). Structural equation models: Empirical identification, Heywood cases, and related problems. Sociological Methods and Research, 13, 109-119. doi:10.1177/0049124184013001004

63.

*Rodríguez-Gómez

J. R.

Dávila-Martínez

M. G.

Collazo-Rodríguez

L. C.

(2006). Factor structure of the Beck Depression Inventory-Second Edition (BDI-II) with Puerto Rican elderly. Puerto Rico Health Sciences Journal, 25, 127-132.

64.

*Rowland

S. M.

Lam

C. S.

Leahy

(2005). Use of the Beck Depression Inventory-II (BDI-II) with persons with traumatic brain injury: Analysis of factorial structure. Brain Injury, 19, 77-83. doi:10.1080/02699050410001719988

65.

*Segal

D. L.

Coolidge

F. L.

Cahill

B. S.

O’Riley

A. A.

(2008). Psychometric properties of the Beck Depression Inventory-II (BDI-II) among community-dwelling older adults. Behavior Modification, 32, 3-20. doi:10.1177/0145445507303833

66.

Shafer

(2005). Meta-analysis of the Brief Psychiatric Rating Scale factor structure. Psychological Assessment, 17, 324-335. doi:10.1037/1040-3590.17.3.324

67.

Shafer

A. B.

(2006). Meta-analysis of the factor structures of four depression questionnaires: Beck, CES-D, Hamilton, and Zung. Journal of Clinical Psychology, 62, 123-146. doi:10.1002/jclp.20213

68.

Shulz

Beach

Ives

Martire

Ariyo

Kop

(2000). Association between depression and mortality in older adults: The Cardiovascular Health Study. Archives of Internal Medicine, 160, 1761-1768. doi:10.1001/archinte.160.12.1761

69.

*Siegert

R. J.

Walkey

F. H.

Turner-Stokes

(2009). An examination of the factor structure of the Beck Depression Inventory-II in a neurorehabilitation inpatient sample. Journal of the International Neuropsychological Society, 15, 142-147. doi:10.1017/S1355617708090048

70.

*Spencer

P. S.

(2004). A psychometric study of the Beck Depression Inventory Second Edition used with incarcerated adult males (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3123491)

71.

*Steele

G. I.

Edwards

D. J. A.

(2008). Development and validation of the Xhosa translations of the Beck Inventories: 2. Item analysis, internal consistency, and factor analysis. Journal of Psychology in Africa, 18, 217-226.

72.

*Steer

R. A.

Ball

Ranieri

W. F.

Beck

A. T.

(1999). Dimensions of the Beck Depression Inventory-II in clinically depressed outpatients. Journal of Clinical Psychology, 55, 117-128.

73.

*Steer

R. A.

Clark

D. A.

(1997). Psychometric characteristics of the Beck Depression Inventory-II with college students. Measurement and Evaluation in Counseling and Development, 30, 128-136.

74.

*Steer

R. A.

Kumar

Ranieri

W. F.

Beck

A. T.

(1998). Use of the Beck Depression Inventory-II with adolescent psychiatric outpatients. Journal of Psychopathology and Behavioral Assessment, 20, 127-137. doi:10.1023/A:1023091529735

75.

Titov

Dear

B. F.

McMillan

Anderson

Zou

Sunderland

(2011). Psychometric comparison of the PHQ-9 and BDI-II for measuring response during treatment of depression. Cognitive Behaviour Therapy, 40, 126-136. doi:10.1080/16506073.2010.550059

76.

*Uslu

R. I.

Kapci

E. G.

Oncu

Ugurlu

Turkcapar

(2008). Psychometric properties and cut-off scores of the Beck Depression Inventory-II in Turkish adolescents. Journal of Clinical Psychology in Medical Settings, 15, 225-233. doi:10.1007/s10880-008-9122-y

77.

Vanheule

Desmet

Groenvynck

Rosseel

Fontaine

(2008). The factor structure of the Beck Depression Inventory-II: An evaluation. Assessment, 15, 177-187. doi:10.1177/1073191107311261

78.

*VanVoorhis

C. R. W.

Blumentritt

T. L.

(2007). Psychometric properties of the Beck Depression Inventory-II in a clinically-identified sample of Mexican American adolescents. Journal of Child and Family Studies, 16, 789-798. doi:10.1007/s10826-006-9125-y

79.

*Viljoen

J. L.

Iverson

G. L.

Griffiths

Woodward

T. S.

(2003). Factor structure of the Beck Depression Inventory-II in a medical outpatient sample. Journal of Clinical Psychology in Medical Settings, 10, 289-291. doi:10.1023/A:1026353404839

80.

Ward

L. C.

(2006). Comparison of factor structure models for the Beck Depression Inventory-II. Psychological Assessment, 18, 81-88. doi:10.1037/1040-3590.18.1.81

81.

*Whisman

M. A.

Perez

J. E.

Ramel

(2000). Factor structure of the Beck Depression Inventory-Second edition (BDI-II) in a student sample. Journal of Clinical Psychology, 56, 545-551.

82.

*Wu

Chang

(2008). Psychometric properties of the Chinese version of the Beck Depression Inventory-II using the Rasch model. Measurement and Evaluation in Counseling and Development, 41, 13-31.

83.

*Zenger

(2009). Malingering of depression in a prison population: Evaluating the utility of the Beck Depression Inventory-II (Doctoral dissertation). Available from ProQuest Dissertations and Theses database. (UMI No. 3376142)

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
	41	31	28	29	31	37	31	40	26	18	28	27	37	17	11	20	11	12	15	10
.58		43	17	36	38	46	39	45	24	11	17	25	47	9	6	10	4	6	7	7
.46	.67		16	46	45	50	48	39	22	11	14	25	46	4	3	9	4	3	2	5
.39	.21	.19		14	13	19	18	18	15	19	44	27	18	38	27	25	24	29	35	27
.38	.51	.68	.15		43	45	45	35	22	10	14	22	40	5	5	11	5	4	3	4
.47	.60	.72	.15	.61		44	41	37	26	13	14	21	42	5	4	10	4	4	3	5
.53	.71	.80	.23	.66	.70		51	45	23	11	16	25	52	7	5	12	6	5	5	5
.45	.62	.74	.21	.67	.64	.78		38	25	14	15	25	45	4	5	12	5	4	2	4
.56	.65	.62	.21	.52	.60	.68	.59		25	12	15	24	46	8	4	13	5	5	6	6
.37	.32	.27	.19	.26	.35	.30	.28	.33		21	15	23	22	9	11	17	11	9	8	5
.26	.14	.11	.27	.10	.14	.12	.14	.14	.27		24	25	10	22	24	28	24	23	23	10
.37	.19	.15	.67	.13	.14	.18	.17	.16	.18	.34		30	17	37	28	28	28	30	34	21
.37	.33	.33	.41	.29	.32	.35	.33	.33	.31	.35	.44		26	25	21	23	22	26	24	14
.47	.66	.70	.24	.58	.65	.85	.69	.64	.27	.12	.22	.35		9	4	11	5	6	7	8
.21	.08	.03	.54	.04	.04	.06	.03	.07	.12	.33	.54	.39	.11		43	32	41	51	56	28
.14	.06	.02	.43	.03	.03	.04	.03	.04	.16	.35	.45	.35	.06	.71		28	48	43	48	25
.25	.09	.08	.38	.11	.11	.11	.12	.12	.22	.37	.43	.33	.09	.47	.45		27	28	31	16
.14	.04	.03	.38	.03	.03	.05	.03	.05	.16	.32	.43	.33	.07	.67	.82	.40		43	46	21
.15	.06	.02	.44	.03	.03	.04	.03	.04	.10	.31	.44	.38	.08	.76	.65	.40	.65		51	26
.19	.69	.02	.52	.02	.03	.05	.02	.05	.12	.34	.52	.38	.09	.84	.84	.45	.80	.76		30
.11	.06	.06	.38	.05	.06	.06	.05	.06	.04	.12	.32	.23	.12	.44	.38	.24	.32	.42	.45

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
.55
.25	.56
.20	.25	.64
.24	.22	.18	.57
.18	.16	.22	.17	.51
.23	.23	.27	.21	.24	.94
.26	.27	.32	.26	.24	.26	.76
.25	.25	.29	.24	.27	.28	.38	.77
.16	.15	.15	.13	.12	.14	.17	.16	.31
.29	.21	.19	.23	.20	.27	.26	.29	.17	.95
.20	.19	.17	.19	.16	.20	.20	.23	.11	.24	.68
.25	.25	.20	.32	.18	.23	.27	.24	.14	.27	.22	.70
.24	.24	.23	.25	.22	.24	.29	.29	.15	.26	.23	.31	.73
.25	.27	.30	.24	.20	.26	.33	.31	.18	.23	.18	.26	.28	.56
.22	.19	.16	.24	.13	.15	.22	.19	.11	.21	.16	.27	.24	.21	.57
.18	.16	.14	.19	.13	.16	.19	.20	.11	.21	.20	.22	.22	.16	.23	.79
.24	.21	.18	.22	.18	.21	.24	.23	.13	.25	.27	.28	.26	.22	.23	.20	.65
.20	.16	.14	.20	.13	.18	.20	.21	.11	.23	.21	.25	.24	.18	.22	.29	.22	.79
.23	.21	.21	.24	.17	.22	.26	.25	.13	.25	.25	.29	.34	.24	.26	.24	.26	.25	.66
.24	.20	.18	.26	.15	.19	.23	.21	.12	.23	.21	.30	.27	.23	.38	.29	.27	.27	.31	.70
.13	.11	.09	.17	.10	.10	.13	.11	.08	.14	.10	.18	.15	.13	.14	.13	.13	.14	.14	.18	.61

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
	41	31	28	29	31	37	31	40	26	18	28	27	37	17	11	20	11	12	15	10
.58		43	17	36	38	46	39	45	24	11	17	25	47	9	6	10	4	6	7	7
.46	.67		16	46	45	50	48	39	22	11	14	25	46	4	3	9	4	3	2	5
.39	.21	.19		14	13	19	18	18	15	19	44	27	18	38	27	25	24	29	35	27
.38	.51	.68	.15		43	45	45	35	22	10	14	22	40	5	5	11	5	4	3	4
.47	.60	.72	.15	.61		44	41	37	26	13	14	21	42	5	4	10	4	4	3	5
.53	.71	.80	.23	.66	.70		51	45	23	11	16	25	52	7	5	12	6	5	5	5
.45	.62	.74	.21	.67	.64	.78		38	25	14	15	25	45	4	5	12	5	4	2	4
.56	.65	.62	.21	.52	.60	.68	.59		25	12	15	24	46	8	4	13	5	5	6	6
.37	.32	.27	.19	.26	.35	.30	.28	.33		21	15	23	22	9	11	17	11	9	8	5
.26	.14	.11	.27	.10	.14	.12	.14	.14	.27		24	25	10	22	24	28	24	23	23	10
.37	.19	.15	.67	.13	.14	.18	.17	.16	.18	.34		30	17	37	28	28	28	30	34	21
.37	.33	.33	.41	.29	.32	.35	.33	.33	.31	.35	.44		26	25	21	23	22	26	24	14
.47	.66	.70	.24	.58	.65	.85	.69	.64	.27	.12	.22	.35		9	4	11	5	6	7	8
.21	.08	.03	.54	.04	.04	.06	.03	.07	.12	.33	.54	.39	.11		43	32	41	51	56	28
.14	.06	.02	.43	.03	.03	.04	.03	.04	.16	.35	.45	.35	.06	.71		28	48	43	48	25
.25	.09	.08	.38	.11	.11	.11	.12	.12	.22	.37	.43	.33	.09	.47	.45		27	28	31	16
.14	.04	.03	.38	.03	.03	.05	.03	.05	.16	.32	.43	.33	.07	.67	.82	.40		43	46	21
.15	.06	.02	.44	.03	.03	.04	.03	.04	.10	.31	.44	.38	.08	.76	.65	.40	.65		51	26
.19	.69	.02	.52	.02	.03	.05	.02	.05	.12	.34	.52	.38	.09	.84	.84	.45	.80	.76		30
.11	.06	.06	.38	.05	.06	.06	.05	.06	.04	.12	.32	.23	.12	.44	.38	.24	.32	.42	.45

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
.55
.25	.56
.20	.25	.64
.24	.22	.18	.57
.18	.16	.22	.17	.51
.23	.23	.27	.21	.24	.94
.26	.27	.32	.26	.24	.26	.76
.25	.25	.29	.24	.27	.28	.38	.77
.16	.15	.15	.13	.12	.14	.17	.16	.31
.29	.21	.19	.23	.20	.27	.26	.29	.17	.95
.20	.19	.17	.19	.16	.20	.20	.23	.11	.24	.68
.25	.25	.20	.32	.18	.23	.27	.24	.14	.27	.22	.70
.24	.24	.23	.25	.22	.24	.29	.29	.15	.26	.23	.31	.73
.25	.27	.30	.24	.20	.26	.33	.31	.18	.23	.18	.26	.28	.56
.22	.19	.16	.24	.13	.15	.22	.19	.11	.21	.16	.27	.24	.21	.57
.18	.16	.14	.19	.13	.16	.19	.20	.11	.21	.20	.22	.22	.16	.23	.79
.24	.21	.18	.22	.18	.21	.24	.23	.13	.25	.27	.28	.26	.22	.23	.20	.65
.20	.16	.14	.20	.13	.18	.20	.21	.11	.23	.21	.25	.24	.18	.22	.29	.22	.79
.23	.21	.21	.24	.17	.22	.26	.25	.13	.25	.25	.29	.34	.24	.26	.24	.26	.25	.66
.24	.20	.18	.26	.15	.19	.23	.21	.12	.23	.21	.30	.27	.23	.38	.29	.27	.27	.31	.70
.13	.11	.09	.17	.10	.10	.13	.11	.08	.14	.10	.18	.15	.13	.14	.13	.13	.14	.14	.18	.61

Meta-Analysis of the Factor Structures of the Beck Depression Inventory–II

Abstract

Keywords

The Beck Depression Inventory–Second Edition

Previous Meta-Analyses

Effect Sizes in Meta-Analysis of Exploratory Factor Analysis

Moderators

Language Version

Risk Level

Age

Confirmatory Factor Analysis of the BDI-II

Study 1

Method

Literature Search

Analysis

Results

Included Studies

Factor Analysis Results for the Full Sample

Comparison of Factor Analysis Between English and Non-English BDI-II Versions

Comparison of Factor Analysis Between the General and At-Risk/Clinical Samples

Comparison of Factor Analysis Between Youths/College Students and Adults

Study 2

Method

Literature Search

Models

Results

Discussion

Footnotes

Declaration of Conflicting Interests

Funding

References

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
	41	31	28	29	31	37	31	40	26	18	28	27	37	17	11	20	11	12	15	10
.58		43	17	36	38	46	39	45	24	11	17	25	47	9	6	10	4	6	7	7
.46	.67		16	46	45	50	48	39	22	11	14	25	46	4	3	9	4	3	2	5
.39	.21	.19		14	13	19	18	18	15	19	44	27	18	38	27	25	24	29	35	27
.38	.51	.68	.15		43	45	45	35	22	10	14	22	40	5	5	11	5	4	3	4
.47	.60	.72	.15	.61		44	41	37	26	13	14	21	42	5	4	10	4	4	3	5
.53	.71	.80	.23	.66	.70		51	45	23	11	16	25	52	7	5	12	6	5	5	5
.45	.62	.74	.21	.67	.64	.78		38	25	14	15	25	45	4	5	12	5	4	2	4
.56	.65	.62	.21	.52	.60	.68	.59		25	12	15	24	46	8	4	13	5	5	6	6
.37	.32	.27	.19	.26	.35	.30	.28	.33		21	15	23	22	9	11	17	11	9	8	5
.26	.14	.11	.27	.10	.14	.12	.14	.14	.27		24	25	10	22	24	28	24	23	23	10
.37	.19	.15	.67	.13	.14	.18	.17	.16	.18	.34		30	17	37	28	28	28	30	34	21
.37	.33	.33	.41	.29	.32	.35	.33	.33	.31	.35	.44		26	25	21	23	22	26	24	14
.47	.66	.70	.24	.58	.65	.85	.69	.64	.27	.12	.22	.35		9	4	11	5	6	7	8
.21	.08	.03	.54	.04	.04	.06	.03	.07	.12	.33	.54	.39	.11		43	32	41	51	56	28
.14	.06	.02	.43	.03	.03	.04	.03	.04	.16	.35	.45	.35	.06	.71		28	48	43	48	25
.25	.09	.08	.38	.11	.11	.11	.12	.12	.22	.37	.43	.33	.09	.47	.45		27	28	31	16
.14	.04	.03	.38	.03	.03	.05	.03	.05	.16	.32	.43	.33	.07	.67	.82	.40		43	46	21
.15	.06	.02	.44	.03	.03	.04	.03	.04	.10	.31	.44	.38	.08	.76	.65	.40	.65		51	26
.19	.69	.02	.52	.02	.03	.05	.02	.05	.12	.34	.52	.38	.09	.84	.84	.45	.80	.76		30
.11	.06	.06	.38	.05	.06	.06	.05	.06	.04	.12	.32	.23	.12	.44	.38	.24	.32	.42	.45

i1	i2	i3	i4	i5	i6	i7	i8	i9	i10	i11	i12	i13	i14	i15	i16	i17	i18	i19	i20	i21
.55
.25	.56
.20	.25	.64
.24	.22	.18	.57
.18	.16	.22	.17	.51
.23	.23	.27	.21	.24	.94
.26	.27	.32	.26	.24	.26	.76
.25	.25	.29	.24	.27	.28	.38	.77
.16	.15	.15	.13	.12	.14	.17	.16	.31
.29	.21	.19	.23	.20	.27	.26	.29	.17	.95
.20	.19	.17	.19	.16	.20	.20	.23	.11	.24	.68
.25	.25	.20	.32	.18	.23	.27	.24	.14	.27	.22	.70
.24	.24	.23	.25	.22	.24	.29	.29	.15	.26	.23	.31	.73
.25	.27	.30	.24	.20	.26	.33	.31	.18	.23	.18	.26	.28	.56
.22	.19	.16	.24	.13	.15	.22	.19	.11	.21	.16	.27	.24	.21	.57
.18	.16	.14	.19	.13	.16	.19	.20	.11	.21	.20	.22	.22	.16	.23	.79
.24	.21	.18	.22	.18	.21	.24	.23	.13	.25	.27	.28	.26	.22	.23	.20	.65
.20	.16	.14	.20	.13	.18	.20	.21	.11	.23	.21	.25	.24	.18	.22	.29	.22	.79
.23	.21	.21	.24	.17	.22	.26	.25	.13	.25	.25	.29	.34	.24	.26	.24	.26	.25	.66
.24	.20	.18	.26	.15	.19	.23	.21	.12	.23	.21	.30	.27	.23	.38	.29	.27	.27	.31	.70
.13	.11	.09	.17	.10	.10	.13	.11	.08	.14	.10	.18	.15	.13	.14	.13	.13	.14	.14	.18	.61