Single Higher-order Factor Research Articles

Confirmatory factor analysis of the Nicotine Dependence Syndrome Scale in an American college sample of light smokers

The five-factor structure of the Nicotine Dependence Syndrome Scale (NDSS) obtained with samples of mature, heavy smokers has not been replicated in samples of younger, less experienced smokers. Furthermore, the idea that the interrelationships among the NDSS factors are best explained by a single higher-order factor (nicotine dependence) has not been evaluated empirically. This study examined the first- and second-order factor structure of the NDSS in a first-year college sample of light smokers (N = 154). NDSS measures completed at the end of the first semester of college were analyzed using confirmatory factor analysis. The hypothesized five-factor model provided adequate fit to the data. The second-order factor model did not fit the data as well as the correlated first-order factor model, suggesting that an overarching dependence factor may not account for the interrelationships among the five first-order factors. This study provided support for the multidimensional structure of the NDSS among a first-year college sample of light smokers.

Introduction to Special Feature on the Structure of Psychopathy

The subject of this special feature—the structure of psychopathy—is currently a topic of considerable debate. Perhaps surprisingly, papers on the structure of psychopathy have generated more controversy than any other Journal publications. The debate which centers on the optimal way to represent the factor structure of psychopathy echoes some aspects of the more general debate on the factor structure of normal personality, a debate that has become somewhat muted of late although the issue appears far from settled. The structure of psychopathy is important given the prevalence of the disorder and the implications of the diagnosis. For this reason, most major contributors to the debate were asked to present their views for this special feature although, unfortunately, some of those invited could not meet the journal’s publication schedule. Much of the controversy is not actually about the structure of the disorder but rather about the best way to represent the factor structure of a specific measure, Robert Hare’s Psychopathy Checklist-Revised (PCL-R). This is understandable given the significance of the PCL-R although, as some contributors note, problems occur when the measure is confused with the disorder. Nevertheless, the substantial progress that has been made over the last decade or so in understanding the consequences and correlates of psychopathy owes much to the robustness of this measure. For this reason, the papers in this feature cover both the structure of the PCL-R and the structure of psychopathy as conceptualized from other vantage points. The first two papers deal specifically with the structure of the checklist. The first paper by Craig Neumann, Robert Hare, and Joseph Newman describes their four-factor model of Interpersonal, Affective, Lifestyle, and Antisocial components. This structure contrasts with the three-factor structure described by David Cooke and colleagues who argued that the antisocial component should not be included as a primary feature when defining the construct. The four factors are discussed in detail and further empirical evidence is offered to support a model of four first-order factors being explained by a single higher-order factor. The second paper by Christopher Patrick, Brian Hicks, Penny Nichol,

The K-factor, Covitality, and personality : A Psychometric Test of Life History Theory.

We present a psychometric test of life history theory as applied to human individual differences using MIDUS survey data (Brim et al. 2000). Twenty scales measuring cognitive and behavioral dimensions theoretically related to life history strategy were constructed using items from the MIDUS survey. These scales were used to construct a single common factor, the K-factor, which accounted for 70% of the reliable variance. The scales used included measures of personal, familial, and social function. A second common factor, Covitality, was constructed from scales for physical and mental health. Finally, a single general factor, Personality, was constructed from scales for the "Big Five" factors of personality. The K-factor, covitality factor, and general personality factor correlated significantly with each other, supporting the prediction that high K predicts high somatic effort and also manifests in behavioral display. Thus, a single higher-order common factor, the Super-K factor, was constructed that consisted of the K-factor, covitality factor, and personality factor.

Higher-Order Factors in a Five-Factor Personality Inventory and its Relation to Social Desirability

Abstract. This paper compared two higher-order factor models using a 100-item, five-factor personality inventory originating from the IPIP database. The sample consisted of 2,019 subjects tested on the Internet. The two models were compared using confirmatory factor analysis. The two-factor model showed a similar fit to the data. The criteria for parsimony favored a hierarchical model with one higher-order factor at the top and five personality factors beneath. The single higher-order factor was found to be related to social desirability in a subsample of 196 subjects.

Callous-Unemotional Traits in a Community Sample of Adolescents

This study examined the structure, distribution, and correlates of a new measure of self-reported callous-unemotional (CU) traits in 1,443 adolescents (774 boys, 669 girls) between the ages of 13 to 18 years. The Inventory of Callous-Unemotional Traits was subjected to exploratory factor analysis and confirmatory factor analysis. Exploratory factor analysis produced three factors: callousness, uncaring, and unemotional. Fit indexes suggested that the three-factor model, with a single higher-order factor, represented a satisfactory solution for the data. This factor structure fits well for both boys and girls. CU traits correlated significantly with measures of conduct problems and psychosocial impairment. Furthermore, the traits showed predicted associations with sensation seeking and the Big Five personality dimensions, supporting the construct validity of the measure of CU traits.

Structural Validity of a Self-Report Concussion-Related Symptom Scale

This study evaluated the factorial validity of a self-report measure of concussion-related symptom severity among a large sample of male, high-school athletes. Participants (N = 1089) were nonconcussed, male, high-school football players. All participants completed a single baseline self-report measure of concussion-related symptom severity, namely the graded symptom checklist (GSC). We tested the factorial validity of the measure with confirmatory factor analysis using LISREL 8.50. The analysis indicated that a theoretically derived, three-factor model provided a good, but not excellent, fit for the 16-item GSC. Excellent model-data fit was demonstrated for the three-factor model for a 9-item version of the GSC. In both instances, the three factors were best described by a single second-order factor, namely concussion symptomatology. This study provides additional evidence for the factorial validity of a summative self-reported measure of concussion-related symptoms. The factor structure represents a cohesive group of nine symptoms that can be explained by three underlying latent variables, namely somatic symptoms, neurobehavioral symptoms, and "cognitive" symptoms, subsumed under a single higher-order factor, namely concussion symptoms.

Construct Validity of the Offer Self-Image Questionnaire and Its Relationship with Self-Esteem, Depression, and Ego Development

Construct validity of the Offer Self-Image Questionnaire (OSIQ) was studied in a sample of 194 normal Finnish adolescents from 14 to 16 years of age. Confirmatory factor analysis provided support for the hierarchical structure of adolescents’ self-image with 5 lower-order factors loading on a single higher-order factor. Lower-order factors were Personally Anxious Self, Social Relationships, Sexual Attitudes, Family Relationships, and Social Conscience. The relationships between the self-image areas of the OSIQ, global self-esteem, depression, and ego development were also examined. The results support the construct, convergent, and discriminant validity of the presented structure of the OSIQ.

Brief Report: Factor Structure of the Childhood Illness Attitude Scales (CIAS)

To examine the factor structure of the Childhood Illness Attitude Scales (CIAS). The CIAS is a 35-item self-report measure based on the Illness Attitudes Scales, designed for use with school-age children. The CIAS measures fears, beliefs, and attitudes associated with health anxiety and abnormal illness behavior in childhood. CIAS item responses for 201 school-age children were subjected to principal-components analysis with oblique rotation. The CIAS was best conceptualized as comprising four factors: fears, help seeking, treatment experience, and symptom effects. Further factor analysis supported the notion that the CIAS can also be conceptualized as having a hierarchical structure, with four lower-order factors loading onto a single higher-order factor of health anxiety. Results suggest that the CIAS possesses good psychometric properties, including factorial validity and internal consistency, and appears to be a psychometrically sound instrument for measuring children's health anxiety.

The heritability of life history strategy: The k‐factor, covitality, and personality

Archival data from the MIDUS survey (Brim et al., 2000), a nationally representative sample, on 309 MZ and 333 DZ twin pairs aged 25-74 years were used to test the psychometrics and behavioral genetics of life history strategy. We organized 253 of the originally administered 2,000 questions into 30 scales measuring life history traits (e.g., quality of family relationships and altruism towards kin), medical symptoms (e.g., thyroid problems), personality traits (e.g., neuroticism, extraversion, conscientiousness), and social background (e.g., financial security). A single higher-order factor, indicating a general life history strategy, composed of three lower-order factors, was replicated. Factor analyses were then performed on the genetic variance-covariance matrices. We found that (a) a single higher-order factor explained the preponderance of the genetic correlations among the scales and (b) this higher-order factor was itself 68 percent heritable and accounted for 82 percent of the genetic variance among the three component lower-order factors.

Relationships Among Social Anxiety Measures and Their Invariance1

Summary: Social phobia is becoming increasingly recognized as an important disorder among adolescents. The body of research on assessment measures in adolescents with social phobia has grown considerably. Unfortunately, little is known about the relationship among these measures and its invariance across clinical and community samples. The objective of the present study is to examine this issue. Results show that all of these measures are invariant among samples and assess a single higher-order factor, labeled as “social anxiety,” although each measure appears to tap a specific symptom (cognitive, behavioral, and somatic). Further, results do support the Social Phobia and Anxiety Inventory (SPAI) and the Social Anxiety Scale for Adolescents (SAS-A) as first-line assessment measures for adolescents' social anxiety.

Personality and the predisposition to engage in risky or problem behaviors during adolescence.

Longitudinal data from a representative sample of 1,978 Black and White adolescents were used to examine the role of personality in multiple risk or problem behaviors. Results indicate that covariation among diverse behaviors (educational underachievement, delinquent behavior, substance use, sexual behavior) can be adequately modeled by a single higher order factor, and that impulsivity and avoidance coping serve as generalized risk factors for involvement in these behaviors. Whereas none of the personality variables examined explained change in problem behaviors once established, avoidance coping prospectively predicted involvement among a subgroup of adolescents with little or no prior involvement. Results suggest that dysfunctional styles of regulating emotions and emotionally driven behaviors are core features of risky or problem behaviors during adolescence.

An expanded childhood anxiety sensitivity index: its factor structure, reliability, and validity in a non-clinical adolescent sample

Anxiety sensitivity refers to the fear of anxiety-related bodily sensations that are interpreted as having potentially harmful somatic, psychological, or social consequences. The current study examined the factor analytic structure of anxiety sensitivity in a large sample of normal adolescents ( N=518) using the revised childhood anxiety sensitivity index (CASI-R). Confirmatory factor analysis indicated that anxiety sensitivity as measured by the CASI-R can best be conceptualised as a hierarchical construct with four lower-order factors loading on a single higher-order factor. The lower-order factors were ‘fear of cardiovascular symptoms’, ‘fear of publicly observable anxiety reactions’, ‘fear of cognitive dyscontrol’, and ‘fear of respiratory symptoms’. An additional aim of the present study was to investigate the psychometric properties of the CASI-R. Results showed the CASI-R to be a reliable scale in terms of internal consistency. Furthermore, CASI-R scores were substantially related to levels of anxiety sensitivity as measured by the original index, trait anxiety, symptoms of anxiety disorders, in particular ‘panic disorder and agoraphobia’, and depression. Finally, some evidence was found for the validity of the CASI-R factor scores. That is, all factors convincingly loaded on symptoms of ‘panic disorder and agoraphobia’, whereas the factor ‘fear of publicly observable anxiety reactions’ was also strongly associated with symptoms of ‘social phobia’.

Assessing individual differences in attention to pain: psychometric properties of the Pain Vigilance and Awareness Questionnaire modified for a non-clinical pain sample

Abstract Recently proposed models [e.g. Eccleston, C., & Crombez, G. (1999). Pain demands attention: a cognitive-affective model of the interruptive function of pain. Psychological Bulletin, 125, 356–366; McCracken, L.M. (1997). Attention to pain in persons with chronic pain: a behavioural approach. Behaviour Therapy, 28, 271–284)] implicating attention to pain as an important factor in the development of chronic pain highlight the need for a self-report measure that assesses the tendency to attend to pain. The Pain Vigilance and Awareness Questionnaire (PVAQ; McCracken, 1997) is a potentially useful measure of pain-related attention. However, its present wording precludes its use with non-clinical samples and its psychometric properties have not been examined in samples other than chronic low back pain. In the present study, the factor structure and psychometric properties of the PVAQ modified for use with a non-chronic pain sample were evaluated. A series of principal-components analyses supported a hierarchical solution comprised of three lower-order factors (viz. Awareness of Change, Intrusion, and Monitoring) and a single higher-order pain vigilance and attention factor. Cronbach's alphas indicated that the PVAQ total scale and the three subscales identified in the principal-components analysis had excellent to adequate internal consistency. Supportive of the validity of the PVAQ, significant positive associations between PVAQ scores and important pain variables (viz., pain severity ratings and change/restriction in lifestyle) were found. Potential applications of the modified PVAQ as well as directions for further psychometric evaluation are discussed.

Assessment of anxiety sensitivity in young American Indians and Alaska Natives

In the present study, the Anxiety Sensitivity Index [ASI; Behav. Res. Ther. 24 (1986) 1] was administered to 282 American Indian and Alaska Native college students in a preliminary effort to: (a) evaluate the factor structure and internal consistency of the ASI in a sample of Native Americans; (b) examine whether this group would report greater levels of anxiety sensitivity and gender and age-matched college students from the majority (Caucasian) culture lesser such levels; and (c) explore whether gender differences in anxiety sensitivity dimensions varied by cultural group (Native American vs. Caucasian). Consistent with existing research, results of this investigation indicated that, among Native peoples, the ASI and its subscales had high levels of internal consistency, and a factor structure consisting of three lower-order factors (i.e. Physical, Psychological, and Social Concerns) that all loaded on a single higher-order (global Anxiety Sensitivity) factor. We also found that these Native American college students reported significantly greater overall ASI scores as well as greater levels of Psychological and Social Concerns relative to counterparts from the majority (Caucasian) culture. There were no significant differences detected for ASI physical threat concerns. In regard to gender, we found significant differences between males and females in terms of total and Physical Threat ASI scores, with females reporting greater levels, and males lesser levels, of overall anxiety sensitivity and greater fear of physical sensations; no significant differences emerged between genders for the ASI Psychological and Social Concerns dimensions. These gender differences did not vary by cultural group, indicating they were evident among Caucasian and Native Americans alike. We discuss the results of this investigation in relation to the assessment of anxiety sensitivity in American Indians and Alaska Natives, and offer directions for future research with the ASI in Native peoples.

Does the social concerns component of the Anxiety Sensitivity Index belong to the domain of anxiety sensitivity or the domain of negative evaluation sensitivity?

The present study utilized an exploratory factor-analytic approach (i.e. principal-components analysis; PCA) to investigate whether the Social Concerns component of the Anxiety Sensitivity Index (ASI [Peterson, R. A., & Reiss, S. (1992). Anxiety Sensitivity Index manual (2nd ed.). Worthington, OH: International Diagnostic Systems.]) is best conceptualized as belonging to the domain of anxiety sensitivity (AS) and/or the domain of negative evaluation sensitivity (NES). A sample of university students ( N=216) was administered measures of both NES (i.e. Brief Fear of Negative Evaluation scale; Leary, 1983) and AS (i.e. ASI). Participants' responses to the items comprising these measures were subjected to a PCA with oblique rotation. Factors representing the NES construct and the three lower-order AS constructs (i.e. AS Physical, Psychological and Social Concerns) were obtained. Subscales derived from these four factors were positively and significantly correlated with one another and loaded on a single higher-order factor labeled Threat Sensitivity. Thus, the present findings suggest that the AS Social Concerns factor is distinct from NES and the other lower-order components of AS. However, correlational analyses and higher-order PCA indicated that the AS Social Concerns factor taps a blend of AS and NES as well as something unique and distinct from both global AS and NES.

Differentiating Anxious and depressive self-statements : Combined factor structure of the anxious self-statements questionnaire and the Automatic Thoughts Questionnaire-Revised

Cognitive models of negative emotion suggest that depression and anxiety are associated with different cognitive features. However, distinguishing anxious from depressive self-talk is difficult because of the overlap between anxiety and depression. The Automatic Thoughts Questionnaire-Revised was developed to assess self-statements related to depression and the Anxious Self-Statement Questionnaire to assess self-statements related to anxiety. However, confirmatory factor analyses of the pooled items from both measures suggested that this implicit two-factor model did not fit the data. Instead, an exploratory common factor analysis yielded four orthogonal factors: self-statements reflecting depression/hopelessness, self-statements reflecting one's inability to cope, self-statements reflecting anxiety/uncertainty about the future, and positive affect self-statements. In an exploratory hierarchical factor analysis, the first three factors loaded onto a single higher order factor, while positive affect self-statements did not. Attempts to predict depression and trait anxiety on the basis of these factor scores produced complex results, at least potentially due to the relative impurity of the criterion measures. These results provide evidence for the differentiation of anxious and depressive self-talk as well as for the common ground shared by these aspects of internal dialogue. They also support the future study of the factors from the ATQ-R and ASSQ in relation to more construct-pure measures of anxiety and depression.

Exploratory and confirmatory factor analytic investigations of the Illness Attitudes Scale in a nonclinical sample

The Illness Attitudes Scale (IAS) assesses fears, beliefs and attitudes associated with hypochondriasis [Kellner, R. (1986). Somatization and hypochondriasis. New York: Praeger Publishers.]. Recent factor analytic investigations of the IAS in non-clinical samples have suggested a number of different factor solutions. In study 1, we used principal components analysis with both orthogonal and oblique rotation to better explore the structure of this measure. Using a random selection of 390 participants from a larger pool of 780, a five-factor solution was identified: (1) fear of illness, death, disease and pain, (2) effects of symptoms, (3) treatment experiences, (4) disease conviction and (5) health habits. In study 2, confirmatory factor analysis (CFA) of responses from the remaining 390 students evaluated: (a) a single-factor model, (b) Kellner's original nine-factor model, (c) a four-factor model proposed by Ferguson and Daniel [Ferguson, E. & Daniel, E. (1995). The Illness Attitudes Scale (IAS): a psychometric evaluation on a nonclinical population. Personality and Individual Differences, 18, 463–469.], (d) a different four-factor model proposed by Stewart and Watt [Stewart, S. H. & Watt, M. C. (1998). A psychometric investigation of the Illness Attitudes Scale (IAS) in a nonclinical young adult sample. Submitted for publication.] and (e) the five-factor model derived in study 1. Of these models, greatest support was obtained for our five-factor model. However, it was also clear that this model could be improved. Based on the results of the CFA, as well as previous research and theoretical considerations, we tested a revised model in which the health habits factor was deleted. Analysis of the revised model showed that it received the greatest support and could be conceptualized as either four distinct factors or as hierarchical in nature, with four lower-order factors loading on a single higher-order factor. Future directions for research as well as suggestions for scoring and using the IAS with nonclinical samples are discussed.

An Expanded Anxiety Sensitivity Index: Evidence for a Hierarchic Structure in a Clinical Sample

Anxiety sensitivity (AS) is the fear of anxiety-related sensations. According to Reiss’s (e.g., Reiss, 1991) expectancy theory, AS amplifies fear and anxiety reactions, and plays an important role in the etiology and maintenance of anxiety disorders, particularly panic disorder. Recent evidence suggests that AS has a hierarchical structure, consisting of multiple lower order factors, loading on a single higher order factor. If each factor corresponds to a discrete mechanism (Cattell, 1978), then the results suggest that AS arises from a hierarchic arrangement of mechanisms. A problem with previous studies is that they were based on the 16-item Anxiety Sensitivity Index, which may not contain enough items to reveal the type and number of lower order factors. Also, some of the original ASI items are too general to assess specific, lower order factors. Accordingly, we developed an expanded measure of AS—the ASI-R—which consists of 36 items with subscales assessing each of the major domains of AS suggested by previous studies. The ASI-R was completed by 155 psychiatric outpatients. Factor analyses indicated a four-factor hierarchical solution, consisting of four lower order factors, loading on a single higher factor. The lower order factors were: (1) fear of respiratory symptoms, (2) fear of publicly observable anxiety reactions, (3) fear of cardiovascular symptoms, and (4) fear of cognitive dyscontrol. Each factor was correlated with measures of anxiety and depression, and fear of cognitive dyscontrol was most highly correlated with depression, which is broadly consistent with previous research. At pretreatment, patients with panic disorder tended to scored highest on each of the factors, compared to patients with other anxiety disorders and those with nonanxiety disorders. These findings offer further evidence that Reiss’s expectancy theory would benefit from revision, to incorporate the notion of a hierarchic structure of AS.

Further Validation of the Multidimensional Students' Life Satisfaction Scale

The psychometric properties of a children's life satisfaction measure (MSLSS; Huebner, 1994) were assessed with a middle school sample of 291 students. Similar to findings with elementary school students, the MSLSS demonstrated preliminary evidence of reliability and validity for research purposes. The results of confirmatory factor analyses supported the five satisfaction domains identified in previous research with younger students, along with a single higher-order general life satisfaction factor. Relationships among life satisfaction domains, demographic variables, and social desirability responses were discussed. Potential uses of the MSLSS and future research needs were also discussed.

Childbearing Adolescents and Problem Behavior Theory

ABSTRACT Understanding problem behavior among adolescent mothers is critical, as their behaviors affect not only their own lives but the lives of their children. Using confirmatory factor analysis with an ethnically diverse sample of 233 pregnant adolescents, we test a hypothesis derived from problem behavior theory that diverse problem behaviors are manifestations of a single underlying construct Three measurement models of problem behaviors are compared; a single factor model, a four factor model, and a second order factor model in which a single higher order factor is posited to account for the covariation among the four factors. The results show support for the higher order factor model, suggesting a constellation of related problem behaviors in this population.