An 8-Item Short Form of the Inventory of Dimensions of Emerging Adulthood (IDEA) Among Young Swiss Men

Abstract

Emerging adulthood is a period of life transition, in which youths are no longer adolescents but have not yet reached full adulthood. Measuring emerging adulthood is crucial because of its association with psychopathology and risky behaviors such as substance use. Unfortunately, the only validated scale for such measurement has a long format (Inventory of Dimensions of Emerging Adulthood [IDEA]—31 items). This study aimed to test whether a shorter form yields satisfactory results without substantial loss of information among a sample of young Swiss men. Data from the longitudinal Cohort Study on Substance Use Risk Factors were used (N = 5,049). IDEA, adulthood markers (e.g., parenthood or financial independence), and risk factors (i.e., substance use and mental health issues) were assessed. The results showed that an 8-item, short-form scale (IDEA-8) with four factors (experimentation, negativity, identity exploration, and feeling in between) returned satisfactory results, including good psychometric properties, high convergence with the initial scale, and strong empirical validity. This study was a step toward downsizing a measure of emerging adulthood. Indeed, this 8-item short form is a good alternative to the 31-item long form and could be more convenient for surveys with constraints on questionnaire length. Moreover, it should help health care practitioners in identifying at-risk populations to prevent and treat risky behaviors.

Keywords

psychometrics risk factors short-form survey youth

Introduction

Emerging adulthood is a distinct step toward mature adulthood characterized by various changes in education, work, love, and worldview (Arnett, 2000). This developmental period is a major public health concern, as it constitutes a risk phase for young adults, including increase in risky behaviors such as drug use and abuse (Arnett, 2000; Huh, Huang, Liao, Pentz, & Chou, 2013) and increase in psychopathology (Rohde, Lewinsohn, Klein, Seeley, & Gau, 2013). Therefore, identifying when youths are undergoing the process of emerging adulthood is crucial. Emerging adulthood has been measured in various ways including demographic and psychosocial assessments. To our knowledge, the only measure of psychological issues associated with emerging adulthood is the Inventory of Dimensions of Emerging Adulthood (IDEA, Lisha et al., 2012; Reifman, Colwell, & Arnett, 2007), a 31-item scale designed to assess whether young people were undergoing the transition to adulthood and to investigate related psychological changes.

Despite how useful such a scale is, a 31-item scale is too long for inclusion in some large-scale health surveys. This study aimed to test whether a shorter form of the IDEA yields—without substantial loss of information—satisfactory results, including psychometric properties and empirical validity, assessed with adulthood markers and risk factors.

Method

Participants

This study analyzed data collected in the Cohort Study on Substance Use Risk Factors among young Swiss men (N = 5,223). Participants were enrolled during conscription in three Swiss national military recruitment centers (French and German speaking). Assessment was carried out outside the army environment and independently of eligibility for military service (January 2012–April 2013). Missing values were listwise deleted on the IDEA scale (N = 5,049, 96.7%). We assessed construct validity considering risky behaviors and adulthood markers. Missing values were also deleted listwise on these variables (N = 4,816, 95.4% of the participants who completed the IDEA). Lausanne University Medical School’s Clinical Research Ethics Committee approved the study protocol (Protocol No. 15/07).

Measures

IDEA

We used the original IDEA with 31 items to assess psychological issues associated with emerging adulthood (Reifman et al., 2007). We asked participants to think of a roughly 5-year period, with the present in the middle, and gave answers on a 4-point scale of strongly disagree, somewhat disagree, somewhat agree, and strongly agree. The items investigated the six dimensions defined by Arnett (2004) and Reifman, Colwell, and Arnett (2007): identity exploration, experimentation/possibilities, negativity/instability, self-focused, feeling in between, and other-focused.

Adulthood markers

Indicators of adulthood reached included parenthood (coded 1 if participants were a parent or if their partner was pregnant, 0 otherwise), stable relationship (coded 1 if participants were married or lived with their partner, 0 otherwise), financial independence (coded 1 if participants covered their own living expenses by themselves, 0 otherwise), and job situation (coded 1 if participants worked, 0 otherwise [e.g., students], with participants who both studied and worked recorded as students because their education was not yet complete). We also considered age as an indicator of adulthood.

Risk factors

Eight variables assessed substance use. We assessed (1) previous 12 months’ hazardous use of alcohol (i.e., 21 or more standard drinks per week [a drink = 10-12 grams of alcohol], coded 1, 0 for no hazardous use), (2) previous 12 months’ frequent risky single occasion drinking (RSOD; i.e., monthly ingestion of six or more standard drinks on a single occasion, coded 1, 0 for no RSOD), (3) previous 12 months’ hazardous tobacco use (i.e., daily smoking, coded 1, 0 for less than daily smoking), and (4) previous 12 months’ hazardous use of cannabis (i.e., cannabis use at least twice a week, coded 1, 0 for no hazardous use). We also investigated (1) alcohol dependence as in Diagnostic Statistical Manual of Mental disorders, Fourth Edition (DSM-IV), (2) nicotine dependence with the Fagerström test for nicotine dependence, and (3) cannabis use disorder with the Cannabis Use Disorder Identification Test. For these three variables, we used a sum score and a higher score indicated dependence/disorder. Other use of illicit drugs was also assessed with the use of 15 illicit drugs and coded 1 for illicit drug use and 0 otherwise.

We assessed mental health issues with the Major Depressive Inventory (MDI; International Classification of Diseases–10th revision)—World Health Organization-MDI for depression level, using a sum score from 0 (no depression) to 50 (depression), and the mental component summary of the Short Form Health Survey (SF-12), with a global score ranging from 0 (mental health problems) to 100 (no mental health problems).

Statistical Analyses

IDEA structure and item selection

The 5,049 participants who answered the IDEA were randomly split in two subsamples (N = 2,525 and N = 2,524). The first subsample was used for item selection on the 31 IDEA items (IDEA-31). We performed exploratory factor analysis (EFA) for ordinal data with unweighted least squares estimation with an oblique (geomin) rotation to select items with the highest loadings. Item with loadings ≥.70 were selected. This cutoff refer to items for which variance is at least half explained by the factor and has been described as a good measure of latent construct.

The second subsample was used for analyses of the short-form IDEA, with 8 selected items (loadings ≥ .70; IDEA-8). We performed two confirmatory factor analyses (CFAs) for ordinal data with weighted least squares means and variance adjusted estimation to test the structure of the IDEA-8, one CFA with the structure highlighted in the previous analysis, and one CFA with a single factor. To assess models’ adequacy, we used the root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the weighted root mean square residual (WRMR). Values of .05 or less, .95 or more, and 0.90 or less were, respectively, considered as good fits.

Construct and convergent validity

The correlations of the IDEA-8 and IDEA-31 with risk factors were compared to assess whether the results of the IDEA-8 could be unambiguously compared with those of the IDEA-31 (N = 4,816). We used the whole scales instead of factors to give an overview of the overall correlation of IDEA-31 and IDEA-8 with other variables and because previous latent class analysis showed that being an emerging adult was associated with high scores on all dimensions of the IDEA. We computed Pearson correlations between IDEA-31/IDEA-8 mean scores, mental health issues and substance dependence (we first computed Spearman and Pearson correlations to check is the results were comparable), and point-biserial correlations between IDEA-31/IDEA-8 mean scores and substance use. We used point-biserial correlations instead of the whole scales because the type of the variables or their distribution prevented us to use Pearson correlations and to compare correlations with Fischer r to z transformations. We also assessed convergent validity using correlations of the IDEA with adulthood markers (point-biserial correlations, except age: Pearson correlation). Differences between correlations for IDEA-31 and IDEA-8 were assessed with Fischer transformation r to z. We also computed the correlation between IDEA-31 and IDEA-8 mean scores. We performed the analyses with Statistical Package for Social Science version 21.

Results

Overall, participants were 21.3 ± 1.23 years old and 54.5% were French speaking. More demographics about adulthood markers are available at the bottom of Table 1.

Table 1.

Correlations Between IDEA-31 and IDEA-8 With Substance Use, Mental Health, and Adulthood Markers.

	% (N)	IDEA-31	IDEA-8	r to z
Substance use
Hazardous alcohol use^a	6.6 (318)	.013	−.003	0.78
Monthly RSOD^a	44.1 (2,124)	.062***	.058***	0.02
Hazardous tobacco use^a	20.3 (978)	.019	−.014	1.62
Hazardous cannabis use^a	7.5 (359)	.057***	.047***	0.49
Use of illicit drugs^a	10.9 (525)	.078***	.065***	0.64
Alcohol dependence (0–7)^b	0.75 (1.10)	.124***	.152***	1.40
Nicotine dependence (0–10)^b	0.88 (1.65)	.001	−.037*	1.87
Cannabis use disorder (0–40)^b	1.73 (4.29)	.131***	.125***	0.30
Mental health
Depression (0–50)^b	7.82 (7.13)	.125***	.147***	1.10
Mental health (0–100)^b	45.34 (9.51)	−.133***	−.151***	0.90
Adulthood markers
Age^b	21.26 (1.23)	−.069***	−.127***	2.87**
Parenthood (0 = no, 1 = yes/ pregnant partner)^a	2.2 (105)	−.032*	−.047***	0.74
Stable relationship (0 = no, 1 = yes)^a	6.1 (296)	.003	−.038**	2.01*
Financial independence (0 = no, 1 = yes)^a	36.6 (1,763)	−.080***	−.150***	3.48***
Job (0 = no, 1 = yes)^a	25.1 (1,211)	−.043**	−.091***	2.37*

Note. N = 4,816. IDEA = Inventory of Dimensions of Emerging Adulthood; RSOD = risky single occasion drinking.

^aPoint-biserial correlations. Percentages and N are given for those who endorsed the corresponding item.

^bPearson correlations, means (SD) are given.

*p < .05. **p < .01. ***p < .001.

IDEA Structure and Item Selection

In the first subsample (N = 2,525), the EFA for IDEA-31 showed six factors with Eigenvalues higher than 1 (from 9.30 to 1.16), explaining a total of 61.75% of the variance. The scree plot displayed a heavy downward curve after Factor 2 (Eigenvalue = 3.95), but this small number of factors did not explain enough of the variability of the data (48.82% of the total variance). Therefore, we decided to apply the usual cutoff including Eigenvalues > 1 (Factors 3–6 were comprised between 1.16 and 1.89). Eight items showed loadings equal to or higher than .70, which loaded on different factors corresponding to four initial subscales of IDEA-31: “many possibilities” (.80) and “exploration” (.77) for the factor “experimentation/possibilities;” “feeling stressed out” (.86); and “high pressure” (.77) for the factor “negativity/instability;” “defining yourself” (.79); and “deciding on your own beliefs and values” (.80) for the factor “identity exploration;” and “feeling adult in some ways but not others” (.74) and “gradually becoming adult” (.71) for the factor “feeling in between.” The initial factors “self-focus” and “other focus” did not show any items with loading up to .70, and items were mixed up with other subscales in the six-factor model.

The results of the CFA with the IDEA-8 performed with the items presented previously on the second subsample (N = 2,524) showed that each item loaded significantly (p < .001) on the corresponding factor (many possibilities: 0.80, exploration: 0.90, feeling stressed out: 0.78, high pressure: 0.89, defining yourself: 0.82, deciding on your own beliefs and values: 0.74, feeling adult in some ways but not others: 0.78, and gradually becoming adult: 0.75). Correlations between factors were medium or high (.21 ≤ r ≤ .56). Fit indices were good (RMSEA = 0.032, CFI = 0.997, WRMR = 0.656). The model with only one single factor showed poor fit indices (RMSEA = 0.278, CFI = 0.642, WRMR = 11.43).

Construct Validity

The correlation of IDEA-31 and IDEA-8 was .91 (p < .001). Correlations between IDEA-31/IDEA-8 and risk factors showed similar results for both versions of the IDEA (see Table 1). Scoring high on the IDEA was positively associated with monthly RSOD, hazardous cannabis use, use of other illicit drugs, alcohol dependence, and cannabis use disorder. Neither IDEA-31 nor IDEA-8 was associated with hazardous alcohol use, hazardous tobacco use, or nicotine dependence. Mental health issues were associated with emerging adulthood in the same way for IDEA-31 and IDEA-8. For all risk factors, there were no significant differences between correlation coefficients for IDEA-31 and IDEA-8.

Correlations between IDEA-31/IDEA-8 and adulthood markers were quite different (see foot of Table 1). All correlations were negative, which meant that having adult responsibilities was associated with lower emerging adulthood, that is, with achieved adulthood. IDEA-8 performed better, showing significant negative correlations with all adulthood markers, whereas IDEA-31 showed four significant correlations (age, parenthood, financial independence, and job) and three of those four were higher for IDEA-8 than for IDEA-31 (age, financial independence, and job).

Discussion

This study aimed to provide a shorter form of the IDEA without substantial loss of information. An 8-item version (IDEA-8) showed satisfactory results for both psychometric properties and empirical validity. The IDEA-8 selected the 8 items with loadings of at least .70 (i.e., half of the variance of the item is explained by the factor) and presented suitable fit indices. Its four factors included “experimentation/possibilities,” “negativity/instability,” “identity exploration,” and “feeling in between.” Thus, the factors “self-focus” and “other-focus” were not included. The subscale related to “other focus” was not part of the original conceptualization of Arnett (2000), so the exclusion of this factor was relevant to the framework used for this study. Moreover, previous studies (Atak & Cok, 2008; Lisha et al., 2012) did not report the existence of a factor of “self-focus” when measuring emerging adulthood.

IDEA-31 and IDEA-8 were then compared. Overall, IDEA-8 showed an ability similar to that of IDEA-31 to discriminate among risky behaviors associated with emerging adulthood. Scoring high on the IDEA should be synonymous with significant substance use and mental health issues, as reported in previous studies (Arnett, 2000; Huh et al., 2013; Rohde et al., 2013), except for hazardous alcohol use and tobacco use.

Finally, we assessed convergent validity with comparisons of IDEA with adulthood markers, expecting that emerging adulthood should be negatively associated with adopting roles and responsibilities of adulthood. The results showed that the convergent validity of IDEA-8 was better than IDEA-31. Indeed, the negative correlations of IDEA-8 with age, stable relationship, financial independence, and job were stronger than those of IDEA-31. Despite its shorter format, IDEA-8 seemed to be reliable in detecting emerging adulthood.

This study carried some limitations; the main one being that no women could be included. The study is largely representative of men, and studies including women are needed in order to establish whether these findings are consistent for both genders. A second shortcoming was the cross-sectional design of the study that did not allow an evaluation either of test–retest reliability or of temporality between risk factors and emerging adulthood and thus of IDEA’s predictive validity. Another limitation was that conclusions about IDEA-8 were based on analyses of items interspersed within IDEA-31. Applications of IDEA-8 alone are needed. A final limitation dealt with the factor “other focus.” Previous findings showed its importance in predicting substance use among Hispanics (i.e., a more collectivistic group; Allem, Lisha, Soto, Baezconde-Garbanati, & Unger, 2013). This factor may be considered in other samples in order to confirm its relevance.

To conclude, this study was a step toward downsizing a measure of emerging adulthood. This format could be more convenient for surveys with constraints on questionnaire length and for health care practitioners. Indeed, emerging adulthood is an at-risk period for youths, and IDEA-8 should be relevant for health care practitioners in identifying developmental transitions and at-risk youths, and to prevent and treat risky behaviors. Therefore, this instrument should be helpful to enhance youth health behaviors.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Received grant from Swiss National Science Foundation --FN 33CS30_139467.

References

Allem

J.-P.

Lisha

N. E.

Soto

D. W.

Baezconde-Garbanati

Unger

J. B.

(2013). Emerging adulthood themes, role transitions and substance use among Hispanics in Southern California. Addictive Behaviors, 38, 2797–2800.

Arnett

J. J.

(2000). Emerging adulthood. A theory of development from the late teens through the twenties. The American Psychologist, 55, 469–480.

Arnett

J. J.

(2004). Emerging adulthood: The winding road from the late teens through the twenties. New York, NY: Oxford University Press.

Atak

Cok

(2008). The Turkish version of inventory of the dimensions of emerging adulthood. International Journal of Human and Social Sciences, 3, 451–457.

Huh

Huang

Liao

Pentz

Chou

C.-P.

(2013). Transitional life events and trajectories of cigarette and alcohol use during emerging adulthood: Latent class analysis and growth mixture modeling. Journal of Studies on Alcohol and Drugs, 74, 727–735.

Lisha

N. E.

Grana

Sun

Rohrbach

Spruijt-Metz

Reifman

Sussman

(2012). Evaluation of the psychometric properties of the revised inventory of the dimensions of emerging adulthood (IDEA-R) in a sample of continuation high school students. Evaluation & the Health Professions, 37, 156–177.

Reifman

Colwell

M. J.

Arnett

J. J.

(2007). Emerging adulthood: Theory, assessment, and application. Journal of Youth Development, 2, 39–50.

Rohde

Lewinsohn

P. M.

Klein

D. N.

Seeley

J. R.

Gau

J. M.

(2013). Key Characteristics of major depressive disorder occurring in childhood, adolescence, emerging adulthood, and adulthood. Clinical Psychological Science, 1, 41–53.