The structure of affective symptoms in a sample of young adults☆☆☆★

Participants 

Between fall 1998 and spring 1999, schools in the metropolitan area of Stuttgart to Tuebingen were asked to participate in a research project with the official permission of the Department of Education and Culture of the Baden-Wuerttemberg state. The participants were attending high school or college, or vocational schools providing job training (e.g., mail services, hair cutting, banking). The questionnaire consisted of an extended version of the CES-D (see below) as well as other questions about sociodemographic variables, grades, number of friends, and so on.

We received a total of 2,059 questionnaires from young adults. There was a predominance of female subjects in the total sample (61.2 %) The age range was from 18 to 40 years. However, looking at the distribution of age in our sample, we excluded all respondents age 26 or older (n = 42) in order to have a more homogeneous group. The majority of the sample described their nationality as German (n = 1,740, 84.5 %). We found a significant effect of nationality on the CES-D (F [5, 1,843] = 6.07, P < .001). Post-hoc Scheffé test revealed only one significant difference, with Italian individuals scoring higher on the CES-D than Germans. However, we decided to restrict our sample to subjects whose nationality was German. The mean age of the final sample was 19.10 years (SD = 1.49). Sample characteristics are listed in Table 1. All subjects were given written information approved by the ethical committee of the German Psychological Association (Deutsche Gesellschaft für Psychologie, DGPs), but only those who wished to continue with the study completed a written informed consent paper to ensure anonymity for all others.

Table 1. Sample Characteristics (N = 1,639) and Scores on the Extended CES-D Assessing Depressive and Manic Symptoms

				“Depressive”	“Manic”
		%	n	Mean (SD)	Mean (SD)
	Age group (yr)
	18	48.4	793	15.66 (9.07)	7.07 (4.22)
	19	24.6	403	16.52 (9.87)	6.88 (4.15)
	20	12.0	196	15.28 (8.73)	6.64 (4.32)
	21	6.0	99	16.44 (10.24)	6.16 (3.41)
	22	4.3	71	15.34 (9.32)	6.27 (4.11)
	23-25	4.7	77	13.76 (8.15)	6.93 (4.28)
	Sex
	Male	38.6	633	14.10 (8.55)	7.15 (4.46)
	Female	61.3	1003	16.84 (9.58)	6.69 (3.97)
	Missing	0.2	3	-	-
	Education*
	No high school diploma	0.6	10	14.63 (8.60)	10.71 (2.43)
	Low level	7.3	119	19.65 (9.04)	8.48 (4.59)
	Intermediate level	32.9	539	16.14 (9.78)	7.16 (4.37)
	High level	53.8	882	15.03 (8.84)	6.43 (3.94)
	Other or unspecified	5.4	89	-	-
	Socioeconomic level†
	High	0.4	7	12.14 (5.76)	5.57 (4.50)
	Medium	73.0	1197	15.48 (9.17)	6.78 (4.08)
	Low	22.9	375	16.70 (9.63)	7.37 (4.48)
	Missing information	3.7	60	-	-
	Nationality‡
	German	84.5	1740	15.80 (9.29)	6.86 (4.17)
	Italian	1.7	34	22.19 (12.70)	8.43 (4.22)
	Turkish	3.2	66	19.63 (9.60)	7.97 (5.32)
	Yugoslavian	2.7	56	18.71 (11.03)	8.06 (6.05)
	Greek	0.8	17	19.33 (9.82)	7.45 (4.43)
	Other (eg, USA)	3.2	66	-	-
	Missing information	3.9	80	-	-
	*High level of education = 13 years of secondary education (“Abitur”); intermediate level = attending school for 10 years (“Mittlere Reife”); low level = 9 years of education (“Hauptschule”). †The socioeconomic level was determined for this German sample using Kraemer's (1983) financial economic index based on the job of the father and/or mother. ‡The nationality was determined by subjects self-classification. The category “Yugoslavian” is a post-hoc data reduction of all persons coming from former Yugoslavia.

The CES-D and its extension 

The CES-D originally consists of 20 items assessing depressive symptoms.26, 31 Responses to the items are rated using a four-point scale: 0 (rarely or none of the time), 1 (some or a little of the time), 2 (occasionally or a moderate amount of the time), and 3 (most of the time). For these four points, anchors are described specifying what it means to say “some or a little of the time.” Participants are asked how often they have experienced each of the symptoms during the previous week. The psychometric properties of the German and American CES-D are comparable.32, 33

In the present sample the mean score of the CES-D was 15.75 (SD = 9.28). The mean item difficulty was .78 with a range from .30 to 1.65, and the reliability was Cronbach's α = .87. Table 1 displays the means and standard deviations depending on several sociodemographic variables.

For the assessment of hypomanic or manic symptoms the criteria of the DSM-IV8 were used and transformed into questions. The format of the question was the same as the one generally used in the CES-D. Two additional principles were important to us for item development: (1) each DSM-IV criteria had to be covered by at least one question; (2) the instrument should stay as short as possible. The questions were presented in the same response format as the original items of the CES-D. The items were worded as follows: (a) “I was unusually happy, excited or elated”; (b) “My thoughts raced through my mind”; (c) “I was very irritable”; (d) “I was extremely active and occupied myself with several things”; (e) “I was easily distracted and constantly lost my flow of thoughts”; (f) “I hardly needed any sleep nor did I have the desire to sleep”; (g) “I spoke a lot more or faster”; (h) “I thought I possessed special abilities or powers”; and (i) “I couldn't sit still and felt compelled to do something.” The possible range of scores is 0 to 27. In a sample of adolescents the reliability of these nine items proved to be sufficient with α = .65 with a mean score of 7.0. There were some significant associations to sociodemographic variables like sex or age, but they proved to be practically irrelevant according to effect size measures. Depressive and (hypo-) manic symptoms correlated with each other (r = .17; P < .001, n = 3,157).33

In the present sample of young adults the scale mean was 6.88 (SD = 4.17). The item difficulty was .76 with a range from .45 (for item 7) to 1.26 (for item 4), and reliability was Cronbach's α = .68.

CES-D 

For comparison with other studies using the original CES-D, we first present results for the depression scale only. As expected women had higher CES-D scores than men (t = 5.67, df = 1,406.9, P < .001; d = .29) (see Table 1), while age was not significantly related to depressive symptoms in this sample (F [5, 1,595] = 1.30, difference not significant). High, medium, or low education levels were significantly associated with depression (F [2, 1,461] = 12.66, P < .001). Although the effect size was small (f = 0.13), people having a lower education reported more depressive symptoms for the previous week than the other two groups, which did not differ from each other. Using the recommended clinical cutoff for German samples of 23, about 23.4 % (n = 384) of our sample would be considered as “possibly depressed.” In general, these results resemble the ones obtained in other studies.29, 30, 32, 33

Manic/hypomanic symptoms 

As stated, we extended the CES-D and also included items for hypomanic and manic episodes. Although our primary goal was to investigate the factorial structure of the affective symptoms all together, we also wanted information about the nine “manic items.” As with depression, we did not find any age effect for the manic symptoms (F [5, 1,595] = 1.30). However, there was an effect of the education level (F [2, 1,503] = 14.74, P < .001) with people having a high education reporting the least (hypo-) manic symptoms, and those with a low education reporting most (hypo-) manic symptoms. However, the effect size was negligible with f = 0.02. Men reported significantly higher scores than women (t = 2.09, df = 1,196.9, P < .05), but the effect size also was negligible (d = 0.07). The sum score of the manic items correlated significantly with the depression score (r = .17, P < .001), which is in accordance with other data and studies.24, 33, 34, 35

Factorial structure of the affective symptoms 

The difference in rates of depression and depression scores is well documented36, 37, 38; therefore we decided to run the principal component factor analysis separately by gender. To determine the number of factors we used the scree test.

The factor analyses for men and for women both resulted in six factors with an eigenvalue greater than 1.00. The variance explained by these factors was 51.1% in the male sample and 53.0% in the female sample. The eigenvalues and variance explained were in the male sample 6.60 (22.8%), 3.32 (11.5%), 1.49 (5.1%), 1.23 (4.2%), 1.12 (3.8%), and 1.07 (3.7%). The equivalent values in the female sample were 7.58 (26.2%), 2.90 (10.0%), 1.41 (4.9%), 1.32 (4.5%), 1.13 (3.9%), and 1.0 (3.5%). In both cases, however, the scree test suggested a two-factor solution. Table 2 lists the factor loadings. A varimax-rotation led for men and women to a similar factor solution as the unrotated ones. The amount of variance explained by the first two rotated factors was, however, smaller at 24% for men and 29.2% for women. Additionally, the unrotated factor solutions made more sense pertaining to interpretation of the factors.

Table 2. Factor Loadings of Affective Symptoms Using the Extended Version of the CES-D for Male and Female Young Adults

		Males			Females
	Item	h2	I	II	h2	I	II
	1. I was bothered by things	.23	.48	.01	.25	.48	.14
	2. I did not feel like eating	.10	.31	.00	.09	.23	.19
	3. I felt that I could not shake off the blues	.44	.66	−.01	.50	.69	−.15
	4. I felt that I was just as good as other people (−)*	.24	.45	−.18	.27	.48	−.19
	5. I had trouble keeping my mind on what I was doing	.25	.48	.13	.27	.46	.25
	6. I felt depressed	.61	.76	−.17	.67	.82	−.01
	7. I felt everything I did was an effort	.29	.54	−.00	.34	.58	.00
	8. I felt hopeful about the future (−)†	.14	−.00	−.38	.10	−.01	−.31
	9. I though my life had been a failure	.44	.66	−.01	.45	.67	.00
	10. I felt fearful	.34	.58	.00	.44	.66	.00
	11. My sleep was restless	.19	.43	.00	.21	.44	.11
	12. I was happy (−)†	.63	.60	−.52	.61	.70	−.35
	13. I talked less than usual	.36	.60	−.00	.33	.57	−.00
	14. I felt lonely	.40	.63	−.01	.45	.67	−.00
	15. People were unfriendly	.18	.38	.19	.13	.33	.13
	16. I enjoyed life (−)†	.52	.53	−.49	.62	.69	−.38
	17. I had crying spells	.19	.44	.00	.40	.62	.11
	18. I felt sad	.58	.76	−.00	.62	.79	.00
	19. I felt that people disliked me	.30	.55	.00	.25	.47	.17
	20. I could not get “going”	.22	.47	.00	.23	.48	.00
	21. *I was unusually happy, excited or elated	.49	−.00	.70	.54	−.32	.66
	22. *My thoughts raced through my mind	.44	.35	.56	.40	.39	.50
	23. *I was very irritable	.35	.56	.20	.41	.64	.01
	24. *I was extremely active and occupied myself with several things	.27	−.01	.52	.31	−.26	.49
	25. *I was easily distracted and constantly lost my flow of thoughts	.31	.46	.22	.35	.46	.37
	26. *I hardly needed any sleep nor did I have the desire to sleep	.23	.11	.47	.20	−.00	.45
	27. *I spoke a lot more or faster	.39	.12	.61	.34	−.00	.58
	28. *I thought I possessed special abilities or powers	.35	.00	.59	.28	−.01	.53
	29. *I couldn't sit still and felt compelled to do something	.48	.27	.64	.40	.14	.62
	Eigenwert		6.60	3.32		7.58	2.90
	Explained variance		22.8 %	11.5 %		26.2 %	10.0 %
	*Items to assess hypomanic or manic symptoms; †(−) These items are inversely scored for the original CES-D, and this is the reason why the factor loadings with the factor I “dysphoria” are positive, while the factor loadings with factor II have negative signs.

NOTE. Factor loadings ≥ .40 are printed bold for easier identification; h2 = communality.

The factor loadings in Table 2 show a pattern that suggests that factor I taps “dysphoria/depression” while factor II assesses typical “(hypo-) mania.” The pattern of factor loadings was essentially the same for both sex. We defined a factor loading of at least .40 on one factor as sufficiently high to contribute meaningfully to that factor. Some items, however, did not fulfill this criterion. These were “loss of appetite” (item 2) and “unfriendliness of others” (item 15). The item 8 (“hopefulness about the future”) almost met our criteria to be indicative of mania or hypomania. Interestingly, two of the self-rated DSM symptoms that would be considered indicative of a hypomanic or manic episode loaded higher on the depression factor than on the mania factor: for both men and women, these were “irritability” (item 23) and “distractibility” (item 25).