Validation of the Adult Manifest Anxiety Scale–College Version Scores in a Sample of U.S. College Students

Abstract

French

Two studies examined the psychometric properties of the Adult Manifest Anxiety Scale–College Version (AMAS-C) scores among U.S. college students. In Study 1, 300 college students were administered the AMAS-C. Confirmatory factor analyses (CFAs) indicated that the five-factor model (four anxiety factors and one lie factor) with a higher order factor provided the best fit to the data. In Study 2, 177 college students were administered the AMAS-C and external measures and correlational analyses indicated support for the convergent and discriminant validity of the AMAS-C scores. Implications of the findings of the studies for mental health professionals who work with college students are discussed.

Keywords

Adult Manifest Anxiety Scale–College Version (AMAS-C)anxiety construct validity reliability college students

Mental health problems are a major concern found among students in higher education institutions (Eisenberg, Gollust, Golberstein, & Hefner, 2007). Cross-sectional and longitudinal studies indicate that the prevalence rate of mental health problems among college students is high (Zivin, Eisenberg, Gollust, & Golberstein, 2009), and the prevalence and severity of these problems continue to increase among the college student population (Gallagher, 2007). Zivin and colleagues (2009) conducted a longitudinal study and examined mental health problems, including anxiety, depression, eating disorders, self-injurious behaviors, and suicidal thoughts, among college students. The authors found that approximately 33% of the college students had a mental health problem at baseline and that these problems persisted for 60% of the students over time. Mental health problems have a profound impact on college students’ general well-being (Burris, Brechting, Salsman, & Carlson, 2009), as these problems affect students’ emotional, cognitive, academic, interpersonal, and physical functioning (Tosevski, Milovancevic, & Gajic, 2010).

Anxiety, Prevalence, and Related Problems

Anxiety is one of the most common mental health problems experienced by students on college campuses today (Baez, 2005; Benton, Robertson, Tseng, Newton, & Benton, 2003). Twenge et al. (2010) conducted a cross-temporal meta-analysis among 117 samples of 63,706 college students between the years of 1938 and 2007 on the Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1942) and the Minnesota Multiphasic Personality Inventory–Second Edition (MMPI-2; Butcher, Dahlstrom, Graham, Tellegen, & Kraemmer, 1989) and found a large generational increase in mental health problems, including an increase in anxiety symptoms, among college students over time. These results are similar to Twenge’s (2000) findings of an increased linear trend in anxiety symptoms reported since the 1950s in the college student population. High levels of anxiety are associated with cigarette smoking, heavy alcohol use, and frequent binge drinking (Cranford, Eisenberg, & Serras, 2009), and have a detrimental impact on college students’ educational achievement (Van Ameringen, Mancini, & Farvolden, 2003), school attendance, career selection, interpersonal relationships, and physical health (Baez, 2005). Students who experience high levels of anxiety are at an increased risk for depression and suicidal thoughts (Olfson, Marcus, Wan, & Geissler, 2004). Therefore, assessment of current anxiety symptoms among college students is needed, so effective interventions can then be implemented to reduce students’ anxiety levels.

Anxiety Symptoms Across the Life Span

Researchers have suggested that the expression of anxiety symptoms may vary somewhat across the life span (Beidel & Stanley, 1992; Lipzin, 1991; Lowe, 2001; Lowe & Reynolds, 2005), with a common set of anxiety symptoms found among individuals of all ages and a unique set of anxiety symptoms found among individuals of different age groups (Lowe & Reynolds, 2005). In Lowe’s (2001) study, the author found a common set of anxiety symptoms across the child, adolescent, college student, young and middle-aged adult, and older adult populations as well as a unique set of anxiety symptoms for these different age groups using exploratory factor analysis (EFA). With a slight difference in the expression of anxiety symptoms for different age groups, different measures of anxiety with age-appropriate items would need to be developed for specific age groups, including college students (Lowe & Reynolds, 2005).

Adult Manifest Anxiety Scale–College Version (AMAS-C)

The AMAS-C (C. R. Reynolds, Richmond, & Lowe, 2003a) is a 49-item, multidimensional measure of general or manifest anxiety developed specifically for college students. The concept of manifest anxiety was derived from a trait theory of anxiety (C. R. Reynolds, 1985). The AMAS-C is an upward extension of the Revised Children’s Manifest Anxiety Scale (RCMAS; C. R. Reynolds & Richmond, 1978). The original trial version of the RCMAS served as the model for initial item generation of the AMAS-C items. Each item on the RCMAS was rewritten for the AMAS-C to be consistent with life situations of college students. In addition, test anxiety items that reflect the test anxiety construct were written for the AMAS-C based on the test anxiety literature and the extensive clinical experience of the authors. The initial draft of the AMAS-C consisted of 120 items. Forty-nine of these items were then selected for the final version of the measure based on college students’ responses to the items found on the original draft of the AMAS-C (C. R. Reynolds, Richmond, & Lowe, 2003b).

To the author’s knowledge, the AMAS-C is the only multidimensional measure of general or manifest anxiety developed specifically for college students. The items selected for the AMAS-C were based on college students’ responses to the items found on the original draft of the measure. The factor structure of the AMAS-C was determined based on the results of an EFA of college students’ responses to the items on the measure, and the norms for the AMAS-C consist of college students only (Lowe & Reynolds, 2005; C. R. Reynolds et al., 2003b).

Lowe and Reynolds (2005) examined the factor structure of the AMAS-C with a sample of 943 college students using EFA. The EFA produced a five-factor structure, four anxiety factors (Physiological Anxiety, Social Concerns/Stress, Test Anxiety, and Worry/Oversensitivity), and a lie factor. Evidence also supported the existence of a higher order factor (a general anxiety factor) named the Total Anxiety factor. However, a study has not been conducted to date to validate the factor structure of the AMAS-C using confirmatory factor analysis (CFA).

Dimensionality of Trait Anxiety

Like Lowe and Reynolds (2005), some researchers view trait anxiety as a multidimensional construct (Endler & Edwards, 1985; Endler, Edwards, & Vitelli, 1991; C. R. Reynolds & Richmond, 1978), whereas other researchers view trait anxiety as a unidimensional construct (Cattell & Scheier, 1961; Spielberger, 1983). Although they disagree about the dimensionality of trait anxiety, many of these researchers in both groups have adopted and argued for Cattell and Scheier’s (1961) state-trait distinction in their research and their measures (Endler et al., 1991). Spielberger (1972) defined state anxiety as a reaction consisting of feelings of tension and apprehension and is situational and transitory in nature, whereas trait anxiety is viewed as anxiety proneness and is a relatively stable personality characteristic. According to Spielberger, individuals differ in trait anxiety based on the way they “perceive a wide range of situations as threatening and respond to these situations with differential elevations of state anxiety” (p. 137). Researchers who view trait anxiety as a unidimensional construct do not specify the types of stressful situations individuals experience that lead to expected increases in state anxiety (Endler et al., 1991) and they do not specify the different components (e.g., physiological, behavioral, and cognitive components) of trait anxiety.

In contrast, Endler (1975) viewed trait anxiety as a multidimensional construct. In his interactional model, he specifies the types of situations in which individuals differ in anxiety proneness. These types of situations interact with trait anxiety resulting in appraisal of situational threat and state anxiety. These types of situations are viewed as multiple related dimensions or factors in his model. Based on the research on the dimensionality of trait anxiety, there is support for a unified construct, a construct with multiple related factors, and a construct with a general factor and specific factors. C. R. Reynolds and colleagues (C. R. Reynolds & Richmond, 1978, C. R. Reynolds et al., 2003b) conceptualized the construct of general or manifest anxiety as consisting of a general factor and specific factors and this conceptualization is reflected in their measures, including the AMAS-C.

The purpose of the present study is to examine the psychometric properties of the AMAS-C scores. The AMAS-C is a multidimensional measure of manifest or general anxiety (C. R. Reynolds et al., 2003b). Limited information is available on the psychometric properties of the AMAS-C scores. In Study 1, the factor structure of the AMAS-C was validated using CFA. The target model (four anxiety factors and one lie factor with a higher order factor) was compared with a two-factor model (one anxiety factor and one lie factor) and a five-factor model (four anxiety factors and one lie factor) to determine the best fit model to the data. The author expected that the target model would provide a better fit to the data than the two- and five-factor models. In Study 2, the convergent and discriminant validity of the AMAS-C scores were examined with the scores of measures external to the test. The findings of Study 2 are expected to contribute to the literature as C. R. Reynolds and colleagues (2003b) did not report convergent and discriminant validity results between the AMAS-C scores and scores of measures external to the test in the Adult Manifest Anxiety Scale manual.

Study 1

Method

Participants

The participants consisted of 300 college students, 142 (47.3%) males and 158 (52.7%) females. The students ranged in age from 18 to 26 years (M = 20.62, SD = 1.98). The average number of years of school for the college students was 13.46 (SD = 1.23, range = 12-16 years). Racial/ethnic composition of the sample included Asian Americans (12.4%), African Americans (4.0%), Hispanics (1.7%), Native Americans (1.0%), Whites (78.9%), and Others (2.0%). The college students were recruited from their classes at a university in the midwestern region of the United States.

Instrument

The AMAS-C is a self-report measure of general or manifest anxiety that consists of a Total Anxiety scale (42 items), four anxiety subscales, Physiological Anxiety (8 items), Social Concerns/Stress (7 items), Test Anxiety (15 items), and Worry/Oversensitivity (12 items), and a Lie scale (7 items). The Total Anxiety scale consists of all of the items from the four anxiety subscales and is an overall measure of general or manifest anxiety. The Physiological Anxiety subscale assesses a student’s physical symptoms associated with one’s anxiety and the Social Concerns/Stress subscale measures a student’s concerns about the views of others and social and daily living activities. The Test Anxiety subscale assesses the anxiety or stress associated with taking tests and the Worry/Oversensitivity subscale provides a measure of excessive nervousness and unproductive rumination. The Lie scale is a validity index and is used to detect purposeful distortion or response bias. College students rate their responses on the AMAS-C using a yes/no format (C. R. Reynolds et al., 2003b).

C. R. Reynolds et al. (2003b) reported internal consistency reliability estimates for the AMAS-C scores. Coefficient alphas of .72 to .95 were found for the AMAS-C scores. C. R. Reynolds and colleagues also found evidence supporting the construct validity of the AMAS-C scores. However, the evidence is limited and is based on the internal structure of the measure. More specifically, C. R. Reynolds and colleagues reported on two indexes of construct validity, including the magnitude of the internal consistency reliability estimates of the AMAS-C scale and subscale scores and the magnitude of the intercorrelations among the AMAS-C subscale scores. The strong to very strong internal consistency reliability estimates reported provide some evidence that the AMAS-C scales and subscales represent coherent constructs. In addition, the authors reported moderate intercorrelations of .36 to .48 between the AMAS-C subscale scores. Moderate intercorrelations between subscale scores of a measure are considered consistent with adequate construct validity (C. R. Reynolds et al., 2003b).

Procedures

Students were administered the AMAS-C in their classrooms, as the AMAS-C may be administered in a group setting or on an individual basis (C. R. Reynolds et al., 2003b). Before the students began work on the AMAS-C, the test administrator asked the students to complete a demographic sheet. Information requested on the demographic sheet included the student’s age, gender, years of schooling completed, and race/ethnicity. Once the students completed the demographic sheet, they were instructed to read the instructions on the measure and record their responses.

Data analyses

Coefficient alphas (equivalent to Kuder–Richardson 20, KR₂₀, reliability estimates when the items are dichotomous) for the AMAS-C scores were computed to assess the internal consistency reliability of the scores of the measure. In addition, the 95% confidence interval (CI) around each reliability estimate (see Fan & Thompson, 2001; Feldt, 1990) was computed. CFAs were then performed using Mplus Version 6 (Muthén & Muthén, 1998-2010). The five-factor model (four anxiety factors and one lie factor) with a higher order factor was compared with two competing models, a two-factor model (one anxiety factor and one lie factor) and a five-factor model (four anxiety factors and one lie factor). Because the factor indicators were binary, model parameters were estimated using robust weighted least squares (WLSMV).

Fit indices were computed for the two-factor model, five-factor model, and five-factor model with a higher order factor to evaluate their overall fit. The fit indices computed were chi-square (χ²), Comparative Fit Index (CFI), Tucker–Lewis Index (TLI), and root mean square error of approximation (RMSEA). Guidelines used to support the contention that a good fit was obtained between the model and the observed data were CFI and TLI values close to .95 or higher and a RMSEA value of less than .06 (Hu & Bentler, 1999).

Results

Coefficient alphas and the 95% CI around each reliability estimate for the AMAS-C scores are presented in Table 1. The internal consistency reliability estimate for the AMAS-C Total Anxiety scale scores was .91. Coefficient alphas for the AMAS-C anxiety subscale scores ranged from .69 to .87. For the AMAS-C Lie scale scores, the internal consistency reliability estimate was .82. Overall, these reliability estimates and the 95% CI around each reliability estimate are within acceptable limits for reliability of a psychological measure (Nunnally & Bernstein, 1994), with the exception of the internal consistency reliability estimate and the 95% CI for the Social Concerns/Stress scores. Given the current data set, the Social Concerns/Stress scores should be interpreted with caution.

Table 1.

Coefficient Alphas (α) and the 95% Confidence Intervals (CIs) for the College Student Sample for the Adult Manifest Anxiety Scale–College Version (AMAS-C) Scale and Subscale Scores (N = 300).

Scale/subscale	α	CI	CI width
Physiological anxiety	.82	[.79, .85]	.06
Social concerns/stress	.69	[.64, 74]	.10
Test anxiety	.87	[.85, .89]	.04
Worry/oversensitivity	.81	[.76, .84]	.08
Lie	.82	[.78, .85]	.07
Total anxiety	.91	[.89, .92]	.03

CFAs were then performed. The fit indices for the two-factor model, five-factor model, and the five-factor model with a higher order factor are presented in Table 2. Based on the fit statistics, the five-factor model with a higher order factor had a better fit to the data than the two-factor model and a similar fit when compared with the five-factor model. Overall, the five-factor model with a higher order factor and the five-factor model provide an acceptable fit to the data. However, the five-factor model with a higher order factor is more parsimonious than the five-factor model and appears justifiable based on theoretical grounds. (The covariance matrix of all of the observed variables in the five-factor model with a higher order factor is available from the author.) The standardized coefficients for the five-factor model with a higher order factor are presented in Table 3. Standardized second-order coefficients were .72 (Social Concerns/Stress), .58 (Test Anxiety), .81 (Worry/Oversensitivity), and .78 (Physiological Anxiety).

Table 2.

Fit Indices From the Confirmatory Factor Analyses of the Adult Manifest Anxiety Scale–College Version (AMAS-C) Scores (N = 300).

Model	χ²	df	CFI	TLI	RMSEA
Two factor	2,749.50*	1,126	.80	.79	.07
Five factor	1,768.96*	1,117	.92	.92	.04
Five factor with a higher order factor	1,766.41*	1,122	.92	.92	.04

Note. CFI = Comparative Fit Index; TLI = Tucker–Lewis Index; RMSEA = root mean square error of approximation.

p < .0001.

Table 3.

Standardized Factor Coefficients for the Five-Factor Model With a Higher Order Factor for the Adult Manifest Anxiety Scale–College Version (AMAS-C; N = 300).

	Standardized factor coefficients
Item number	I	II	III	IV	V
18	.92
6	.88
35	.78
29	.74
43	.73
25	.67
49	.62
13	.59
46	.55
22	.64
4	.46
15	.41
22	.64
45		.93
34		.92
40		.90
32		.85
30		.84
36		.84
19		.79
47		.48
41			.93
33			.90
23			.90
39			.89
2			.86
31			.81
44			.75
14			.75
7			.74
37			.72
5			.70
27			.63
11			.61
17			.44
20			.41
42				.97
26				.86
1				.75
3				.72
38				.71
9				.56
12				.54
21					.98
10					.91
8					.88
16					.83
28					.80
48					.80
24					.53

Note. Factor I = Worry/Oversensitivity; Factor II = Physiological Anxiety; Factor III = Test Anxiety; Factor IV = Social Concerns/Stress; Factor V = Lie.

Study 2

In Study 2, the convergent and discriminant validity of the AMAS-C scores were examined. The AMAS-C scores were compared and contrasted with the scores of the Brief Symptom Inventory 18 (BSI 18; Derogatis, 2000a), Multidimensional Anxiety Questionnaire (MAQ; W. M. Reynolds, 1999a), Test Anxiety Inventory (TAI; Spielberger, 1980a), and college grade point average (GPA). The MAQ and BSI 18 Anxiety scales were selected to assess the convergent validity of the AMAS-C anxiety scores because these three measures assess general anxiety. Researchers have found strong correlations between scores of general anxiety measures at the total level (W. M. Reynolds, 1999b; Spielberger, 1983). Based on these research findings, the author expected the AMAS-C Total Anxiety scores to have their highest correlations with the MAQ Total and BSI 18 Anxiety scores. The MAQ subscales and TAI scale were selected for the present study because these scale and subscales assess similar dimensions found on the AMAS-C. The author expected higher correlations between the AMAS-C anxiety subscale scores and scores of similar dimensions than dissimilar dimensions found on other measures. Conversely, the BSI 18 Depression scores and college GPA were selected to assess the discriminant validity of the AMAS-C anxiety scores. The author expected higher correlations between the AMAS-C anxiety scores and the scores of other anxiety measures (MAQ, BSI 18 Anxiety, and TAI) than the AMAS-C scores and the BSI 18 Depression scores. Equivocal findings have been reported in the literature about the dynamic and complex relationship between anxiety and academic achievement. For example, Ma (1999) stated that math anxiety can facilitate, debilitate, or be unrelated to mathematics performance and is dependent upon the social and academic characteristics of students. In Ma’s meta-analysis, the author found a negative and small correlation between anxiety and math achievement. In contrast, Spielberger (1983) reported a negligible relationship between state and trait anxiety, as measured by the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1997) scores, and student GPAs. Based on these findings, Spielberger suggested that state and trait anxiety are unrelated to achievement. Like the STAI, the AMAS-C is a measure of general (or trait) anxiety, therefore the author of the current study expected negligible to small correlations between the AMAS-C anxiety scores and GPAs. Finally, the author also expected negligible to small correlations between the AMAS-C Lie scores and the MAQ, BSI 18, and TAI scores and college GPAs. The AMAS-C Lie scale scores assess a theoretically distinct construct from the constructs measured by the scores of other instruments and college GPAs included in the current study.

Method

Participants

The participants consisted of 177 college students, 65 males (36.7%) and 112 females (63.3%). The mean age of the college students was 20.13 years (SD = 1.67, range = 18-25). The average number of years of schooling for the undergraduates was 14.32 years (SD = 1.28, range 12-16). Racial/ethnic composition of the sample consisted of 2.8% Asian Americans, 5.1% African Americans, 3.4% Hispanics, .6% Native Americans, 85.3% Whites, and 2.8% Others. The undergraduates were recruited from their classes at a university in the midwestern region of the United States.

Instruments

The AMAS-C, BSI 18, MAQ, TAI, and college GPA were used. The AMAS-C is described in Study 1. A description of the BSI 18, MAQ, and TAI is provided below.

BSI 18

The BSI 18 is an 18-item self-report measure used to screen for psychological distress in adults. The BSI 18 has an Anxiety, Depression, and Somatization scale. Respondents rate each item on the BSI 18 on a 5-point Likert-type scale, ranging from 0 (not at all) to 4 (extremely; Derogatis, 2000b).

Derogatis (2000b) reported internal consistency reliability estimates of .74 to .84 for the BSI 18 scores. Evidence supporting the construct validity of the BSI 18 scores has been found (Derogatis, 2000b; Lowe, 2012).

MAQ

The MAQ is a self-report measure used to assess anxiety symptoms in adults. The MAQ consists of four anxiety subscales (Physiological-Panic, Social Phobia, Worry-Fears, and Negative Affectivity) and a Total scale. Individuals respond to items on the MAQ on 4-point Likert-type scale, ranging from 1 (almost never) to 4 (almost all of the time; W. M. Reynolds, 1999b).

W. M. Reynolds (1999b) reported internal consistency reliability estimates of .80 to .93 for the MAQ scores in a sample of college students. Evidence supporting the construct validity of the MAQ scores has been found (W. M. Reynolds, 1999b).

TAI

The TAI consists of 20 items that assesses test anxiety. Individuals respond to the TAI items on a 4-point Likert-type scale, ranging from 1 (almost never) to 4 (almost always; Spielberger, 1980b).

Spielberger (1980b) reported internal consistency reliability estimates of .92 to .96 in samples of college students. Evidence supporting the construct validity of the TAI scores has been found (Spielberger, 1980b).

Procedures

Students were given a packet of measures, including the AMAS-C, BSI 18, MAQ, and TAI, to complete in their classrooms. The measures were counterbalanced to prevent an order effect. Prior to beginning work on the packet of measures, the test administrator asked the students to complete a demographic sheet. The information requested on the demographic sheet included the student’s age, gender, years of schooling completed, race/ethnicity, and college GPA. Then the students were directed to read the instructions printed on the measures and record their responses.

Data analyses

Coefficient alphas for the AMAS-C, BSI 18, MAQ, and TAI scores were computed to assess the internal consistency reliability of the scores of these measures. In addition, the 95% CI around each reliability estimate (see Fan & Thompson, 2001; Feldt, 1990) was computed. Correlational analyses were performed between the AMAS-C scores and scores external to the test. Pearson Product–Moment correlation coefficients were computed between the AMAS-C scores and the BSI 18, MAQ, and TAI scores and college GPAs to assess the convergent and discriminant validity of the AMAS-C scores.

Results

Internal consistency reliability estimates and the 95% CI around each reliability estimate for the AMAS-C, BSI 18, MAQ, and TAI scores are presented in Table 4. Coefficient alphas ranged from .72 to .91for the AMAS-C scores. For the BSI 18 and MAQ scores, internal consistency reliability estimates ranged from .80 to .85 and .79 to .95, respectively. Coefficient alpha for the TAI scores was .95. The internal consistency reliability estimates and the 95% CIs around each reliability estimate reported for the AMAS-C, BSI 18, MAQ, and TAI scores are within the accepted limits for the reliability of a psychological measure (Nunnally & Bernstein, 1994), with the exception of the 95% CI for the AMAS-C Social Concerns/Stress scores, suggesting the AMAS-C Social Concerns/Stress scores should be interpreted with caution given the current data set.

Table 4.

Coefficient Alphas (α) and the 95% Confidence Intervals (CIs) for the College Student Sample for the Adult Manifest Anxiety Scale–College Version (AMAS-C), the Brief Symptom Inventory 18 (BSI 18), the Multidimensional Anxiety Questionnaire (MAQ), and the Test Anxiety Inventory (TAI) Scale and Subscale Scores (N = 177).

Scale/subscale	α	CI	CI width
AMAS-C physiological anxiety	.82	[.77, .85]	.08
AMAS-C social concerns/stress	.72	[.65, .78]	.13
AMAS-C test anxiety	.87	[.84, .89]	.05
AMAS-C worry/oversensitivity	.82	[.77, .85]	.08
AMAS-C lie	.79	[.74, .84]	.10
AMAS-C total anxiety	.91	[.89, .93]	.04
BSI 18 anxiety	.85	[.81, .88]	.07
BSI 18 depression	.80	[.76, .85]	.09
BSI 18 somatization	.83	[.79, .86]	.05
MAQ negative affectivity	.82	[.78, .86]	.08
MAQ physiological-panic	.89	[.86, .91]	.05
MAQ social phobia	.79	[.74, .84]	.10
MAQ worry-fears	.83	[.79, .86]	.07
MAQ Total Scale	.95	[.93, .96]	.03
TAI total	.95	[.94, .96]	.02

Validity coefficients between the AMAS-C scores and scores external to the test are presented in Table 5. Overall, the AMAS-C Total Anxiety scores had their highest correlations with the MAQ Total and the BSI 18 Anxiety scores. The MAQ Total and the BSI 18 Anxiety scores accounted for 57.76% and 51.84% of the variance in the AMAS-C Total Anxiety scores, respectively. These results were expected as the AMAS-C Total, the MAQ Total, and the BSI 18 Anxiety scores assess a similar construct.

Table 5.

Correlations Between the Adult Manifest Anxiety Scale–College Version (AMAS-C) Scores and the Brief Symptom Inventory 18 (BSI 18), the Multidimensional Anxiety Questionnaire (MAQ), and the Test Anxiety Inventory (TAI), and College Grade Point Averages (GPAs; N = 177).

	AMAS-C
Scales/Subscales	Total anxiety	Physiological anxiety	Social concerns/stress	Test anxiety	Worry/oversensitivity	Lie
MAQ Total Scale	.76**	.67**	.70**	.44**	.66**	−.06
BSI 18 anxiety	.72**	.66**	.55**	.52**	.55**	−.07
MAQ social phobia	.72**	.56**	.69**	.43**	.63**	−.05
MAQ negative affectivity	.69**	.62**	.60**	.42**	.59**	−.12
TAI total	.69**	.48**	.35**	.80**	.33**	.02
MAQ worry-fears	.66**	.60**	.62**	.37**	.61**	−.02
MAQ physiological-panic	.63**	.60**	.58**	.36**	.55**	−.03
BSI 18 depression	.58**	.51**	.60**	.32**	.50**	−.13
BSI 18 somatization	.52**	.49**	.44**	.35**	.44**	.00
College GPA	−.05	.01	−.03	−.16	.12	.11

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

Two of the AMAS-C (Social Concerns/Stress and Test Anxiety) subscale scores had their highest correlations with the MAQ Total scores and/or scores of a similar dimension found on a different measure. On the other hand, the AMAS-C Worry/Oversensitivity subscale scores had their highest correlations with the MAQ Total scores, MAQ Social Phobia scores, and the MAQ Worry-Fears scores. This finding was not too surprising as the MAQ Social Phobia subscale assesses a number of worries individuals may have and, like the AMAS-C Worry/Oversensitivity subscale, taps into the cognitive component of anxiety. In contrast, the AMAS-C Physiological Anxiety scores had their highest correlations with the MAQ Total scores, BSI 18 Anxiety scores, and the MAQ Negative Affectivity scores and a lower correlation with the MAQ Physiological-Panic scores. The AMAS-C Physiological Anxiety scores assess physical symptoms, including somatic complaints and sleeping difficulties, associated with anxiety. Likewise, the MAQ Negative Affectivity scores assess these same symptoms and this may account for the higher correlation reported between the scores of these two subscales. In addition, the MAQ Physiological-Panic subscale consists of items that deal with panic attacks and agoraphobia (W. M. Reynolds, 1999b), whereas the AMAS-C Physiological Anxiety subscale does not assess panic attacks and agoraphobia. This may explain the lower correlation between the scores of the AMAS-C Physiological Anxiety subscale and the MAQ Physiological-Panic subscale.

Moderate correlations of .32 to .60 were found between the AMAS-C anxiety scores and the BSI 18 Depression scores. Overall, these correlations tended to be somewhat lower than the moderate to strong correlations (rs = .36 to .76) reported between the AMAS-C anxiety scores and the MAQ and BSI 18 anxiety scores. However, moderate relations were expected between the scores of the AMAS-C anxiety scale and subscales and the BSI 18 Depression scale as anxiety and depression tend to co-occur and are related constructs (Endler & Summerfeldt, 1995; C. R. Reynolds, 2001). Likewise, moderate correlations of .35 to .52 between the AMAS-C anxiety scores and the BSI 18 Somatization scores were found. Moderate relations were also expected between the AMAS-C anxiety scores and the BSI 18 Somatization scores as individuals with anxiety often present with somatic symptoms (Derogatis, 2000b).

Correlations between the AMAS-C anxiety scores and college GPAs (rs = −.16 to .12) were negligible to small. In addition, the correlations between the AMAS-C Lie scale scores and the BSI 18 scores (rs = −.13 to .00), MAQ scores (rs = −.12 to −.02), TAI scores (r = .02), and college GPAs (r = .11) were negligible to small. Overall, these findings provide support for the convergent and discriminant validity of the AMAS-C scores.

Discussion

Overall, the findings of the present studies provide support for the construct validity of the AMAS-C scores. The results of Study 1 indicated that the five-factor model with a higher order factor provided the best fit to the data. These results reported were similar to Lowe and Reynolds’ (2005) findings of a five-factor structure (Physiological Anxiety, Social Concerns/Stress, Test Anxiety, Worry/Oversensitivity, and Lie) with evidence suggesting the presence of a higher order factor for the AMAS-C using EFA with the AMAS-C’s normative sample of college students. The current study is the first study to validate the AMAS-C’s factor structure using CFA. In Study 2, the AMAS-C scores were compared and contrasted with scores of measures external to the test. Overall, the AMAS-C scores correlated higher with the scores of instruments measuring similar constructs and correlated lower with the scores of measures assessing different constructs. The findings of Study 2 provide support for the convergent and discriminant validity of the AMAS-C scores. However, in both studies, the internal consistency reliability estimate and/or the 95% CI around the reliability estimate for the Social Concerns/Stress scores were lower than expected given the current data set. Although Lowe and Reynolds (2005) and C. R. Reynolds et al. (2003b) reported acceptable internal consistency reliability estimates for the AMAS-C Social Concerns/Stress scores, perhaps it is possible given the results of the present studies that the items on this subscale may reflect broader content than the items on the other AMAS-C subscales and thus, mental health professionals working with college students may want to interpret the Social Concerns/Stress subscale scores with caution. Furthermore, negligible to small correlations were found between the AMAS-C anxiety scores and college GPAs. Overall, the results reported in the present study are in agreement with Spielberger’s (1983) correlational findings, but the correlations in the present study are lower than the correlations reported in Ma’s (1999) meta-analysis. In Van Ameringen et al.’s (2003) study, the authors found an association between individuals with anxiety disorders and premature withdrawal from school. It is possible that the correlations reported in the present study may be smaller in magnitude because some or many individuals with anxiety disorders may have already left the university setting or never enrolled. Further research is needed to explore this issue in the college student population.

Mental health problems, including anxiety problems, among college students represent a growing public health concern (Baez, 2005; Eisenberg et al., 2007). Zivin and colleagues (2009) found that over 50% of college students in their longitudinal study experienced one or more mental health problems at baseline or at follow-up 2 years later. Anxiety concerns are among the most frequently reported mental health problems at counseling centers on college campuses (Baez, 2005; Benton et al., 2003) and there is fair consensus among researchers and clinicians of an increase in anxiety symptoms over time (Twenge et al., 2010). Twenge and colleagues (2010) suggest that our consumer culture and individualism is partially responsible for the rise in mental health problems, including the rise in anxiety problems, over time. Our consumer culture and individualism places high expectations on college students and mental health suffers as a result of these high expectations. According to Twenge et al., today’s college students pursue extrinsic goals (e.g., money, status, and looks) at the expense of intrinsic goals (e.g., competence, affiliation, and autonomy), leading to poor interpersonal relationships, anxiety, and depression. With the rise in anxiety symptoms among the college student population (Twenge, 2000; Twenge et al., 2010) and some specific stressors, worries, and burdens unique to the college student population (Tosevski et al., 2010), there appears to be a need for a measure of anxiety, such as the AMAS-C, designed specifically for college-age students.

Several limitations are associated with the present studies. In Studies 1 and 2, the ethnic and regional diversity of the samples were limited. The samples used in Studies 1 and 2 were recruited from a university in the midwestern region of the United States. As a result, the generalizability of the findings of the studies may be limited. Replication of the present studies with more representative samples of ethnically diverse (i.e., African American, Hispanic, and Native American) college students from all of the major geographic regions of the United States yielding similar results would provide support for the findings reported in the present studies. In addition, replication of the current studies with representative samples of Canadian college and university students is needed to determine the generalizability of the present findings to Canadian college student populations, as Canadian college students’ demographics and anxieties may differ from U.S. college students and more specifically the current sample. In addition, future studies should examine cross-cultural differences between U.S. college students and college students who reside outside North America and group differences between nonreferred college students and college students with learning disabilities, general anxiety, and test anxiety on the AMAS-C. Another limitation associated with Study 2 was the use of self-report only to examine the convergent and discriminant validity of the AMAS-C scores. The use of self-report only may introduce common method bias. In future studies with the AMAS-C and college students, a multitrait multimethod analysis should be conducted to examine the convergent and discriminant validity of the AMAS-C scores. Finally, the internal consistency reliability estimate and/or the 95% CI around the reliability estimate for the Social Concerns/Stress scores were lower than expected in Studies 1 and 2. It is not clear why the reliability estimate and the 95% CI in Study 1 and the 95% CI in Study 2 were lower than expected. Future studies are needed to examine the psychometric properties of the Social Concerns/Stress subscale in more detail with diverse samples of college students from different geographic regions of the country. In these future studies with diverse samples of college students, internal consistency reliability estimates for the AMAS-C Social Concerns/Stress scores should be computed and examined to determine whether these reliability estimates are within accepted limits for a psychological measure. In addition, structural equation modeling with the responses of these diverse samples of college students could be performed to determine whether the elimination of the Social Concerns/Stress dimension results in a better model for the AMAS-C scores. As for now, based on the findings of these two studies, the Social Concerns/Stress subscale should be interpreted with caution when using the AMAS-C with college students.

Although additional research is needed, the AMAS-C appears to be a promising multidimensional measure of anxiety designed specifically for college-age students. College students have specific concerns, stressors, burdens, and worries that differ somewhat from individuals of other age groups (Tosevski et al., 2010) and measures need to be developed to assess the specific life situations of these different age groups. With a continued increase in anxiety symptoms reported over time by college students (Twenge et al., 2010) and the significant impact anxiety symptoms have on the educational, economic, and social outcomes of college students (Andrews, Hejdenberg, & Wilding, 2006; Andrews & Wilding, 2004), a measure such as the AMAS-C may prove invaluable to mental health experts to assess those college students who experience high levels of anxiety.

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The author is a coauthor of the Adult Manifest Anxiety Scale–College Version.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biography

Patricia A. Lowe, PhD, is an associate professor and director of training of the doctoral and educational specialist programs in school psychology at the University of Kansas. Her research interests are in the areas of general and test anxiety, measurement, and prevention.

References

Andrews

Hejdenberg

Wilding

(2006). Student anxiety and depression: Comparison of questionnaire and interview assessments. Journal of Affective Disorders, 95, 29-34.

Andrews

Wilding

(2004). The relation of depression and anxiety to life-stress and achievement in students. British Journal of Psychology, 95, 509-521.

Baez

(2005). Evidenced-based practice for anxiety disorders in college mental health. Journal of College Student Psychotherapy, 20, 33-48.

Beidel

D. C.

Stanley

M. A.

(1992). Developmental issues in the measurement of anxiety. In Last

G. C.

(Ed.), Anxiety across the life span: A developmental perspective (pp. 167-203). New York, NY: Springer.

Benton

S. A.

Robertson

J. M.

Tseng

Newton

F. B.

Benton

S. L.

(2003). Changes in counseling center client problems across 13 years. Professional Psychology: Research and Practice, 34, 66-72.

Burris

J. L.

Brechting

E. H.

Salsman

Carlson

C. R.

(2009). Factors associated with the psychological well-being and distress of university students. Journal of American College Health, 57, 536-543.

Butcher

J. N.

Dahlstrom

W. G.

Graham

J. R.

Tellegen

Kraemmer

(1989). Minnesota Multiphasic Personality Inventory—Second edition. Minneapolis: University of Minnesota.

Cattell

R. B.

Scheier

I. H.

(1961). The meaning and measurement of neuroticism and anxiety. New York, NY: Guilford Press.

Cranford

J. A.

Eisenberg

Serras

A. M.

(2009). Substance use behaviors, mental health problems, and use of mental health services in a probability sample of college students. Addictive Behavior, 34, 134-145.

10.

Derogatis

L. R.

(2000a). BSI 18: Brief Symptom Inventory 18. San Antonio, TX: Pearson.

11.

Derogatis

L. R.

(2000b). BSI 18: Brief Symptom Inventory 18 manual. San Antonio, TX: Pearson.

12.

Eisenberg

Gollust

S. E.

Golberstein

Hefner

J. L.

(2007). Prevalence and correlates of depression, anxiety, and suicidality among university students. American Journal of Orthopsychiatry, 77, 534-542.

13.

Endler

N. S.

(1975). A person-situation interaction model of anxiety. In Spielberger

C. D.

Sarason

I. G.

(Eds.), Stress and anxiety (Vol. 1, pp. 145-164). Washington, DC: Hemisphere.

14.

Endler

N. S.

Edwards

J. M.

(1985). Evaluation of the state-trait distinction within an interactional model of personality. Southern Psychologists, 2, 63-71.

15.

Endler

N. S.

Edwards

J. M.

Vitelli

(1991). Endler Multidimensional Anxiety Scales (EMAS) manual. Los Angeles, CA: Western Psychological Services.

16.

Endler

N. S.

Summerfeldt

L. J.

(1995). Intelligence, personality, psychopathology, and adjustment. In Saklofske

Zeidner

(Eds.), International handbook of personality and intelligence (pp. 249-284). New York, NY: Plenum Press.

17.

Fan

Thompson

(2001). Confidence intervals about score reliability coefficients, please: An EPM guidelines editorial. Educational and Psychological Measurement, 61, 517-531.

18.

Feldt

L. S.

(1990). The sampling theory for the intraclass reliability coefficient. Applied Measurement in Education, 3, 361-367.

19.

Gallagher

(2007). National Survey of Counseling Center Directors, 2006 (Monograph Series No. 8P). Alexandria, VA: International Association of Counseling Services.

20.

Hathaway

S. R.

McKinley

J. C.

(1942). The Minnesota Multiphasic Personality Inventory. Minneapolis: University of Minnesota Press.

21.

Bentler

P. M.

(1999). Cutoff criteria for fit indices in covariance structural analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

22.

Lipzin

(1991). Masked anxiety-alcohol and drug use. In Salzman

Lebowitz

B. D.

(Eds.), Anxiety in the elderly: Treatment and research (pp. 87-101). New York, NY: Springer.

23.

Lowe

P. A.

(2001). Common and uncommon manifestations of anxiety across the life span (Doctoral dissertation, Texas A&M University, 2000). Dissertation Abstracts International, 61, 3896.

24.

Lowe

P. A.

(2012). The Adult Manifest Anxiety Scale–Elderly Version (AMAS-E): Validation of a measure of manifest anxiety in samples of older adults. Manuscript in preparation.

25.

Lowe

P. A.

Reynolds

C. R.

(2005). Factor structure of AMAS-C scores across gender among students in collegiate settings. Educational and Psychological Measurement, 65, 687-708.

26.

(1999). A meta-analysis of the relationship between anxiety toward mathematics and achievement in mathematics. Journal for Research in Mathematics Education, 30, 520-540.

27.

Muthén

L. K.

Muthén

B. O.

(1998-2010). Mplus Version 6. Los Angeles, CA: Author.

28.

Nunnally

J. C.

Bernstein

I. H.

(1994). Psychometric theory (3rd ed.). New York, NY: McGraw-Hill.

29.

Olfson

Marcus

S. C.

Wan

G. J.

Geissler

E. C.

(2004). National trends in the outpatient treatment of anxiety disorders. Journal of Clinical Psychiatry, 65, 1166-1173.

30.

Reynolds

C. R.

(1985). Multitrait validation of the Revised Children’s Manifest Anxiety Scale for children of high intelligence. Psychological Reports, 56, 402.

31.

Reynolds

C. R.

(2001). Professional manual for the Clinical Assessment Scales for Elderly. Odessa, FL: Psychological Assessment Resources.

32.

Reynolds

C. R.

Richmond

B. O.

(1978). What I think and feel: A measure of children’s manifest anxiety. Journal of Abnormal Child Psychology, 25, 15-20.

33.

Reynolds

C. R.

Richmond

B. O.

Lowe

P. A.

(2003a). The Adult Manifest Anxiety Scale–College Version. Los Angeles, CA: Western Psychological Services.

34.

Reynolds

C. R.

Richmond

B. O.

Lowe

P. A.

(2003b). The Adult Manifest Anxiety Scale–College Version manual. Los Angeles, CA: Western Psychological Services.

35.

Reynolds

W. M.

(1999a). Multidimensional Anxiety Questionnaire. Odessa, FL: Psychological Assessment Resources.

36.

Reynolds

W. M.

(1999b). Multidimensional Anxiety Questionnaire: Professional manual. Odessa, FL: Psychological Assessment Resources.

37.

Spielberger

C. D.

(1972). Anxiety as an emotional state. In Spielberger

C. D.

(Ed.), Anxiety: Current trends in theory and research (Vol. 1). New York, NY: Academic Press.

38.

Spielberger

C. D.

(1980a). Test Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.

39.

Spielberger

C. D.

(1980b). Test Anxiety Inventory manual. Palo Alto, CA: Consulting Psychologists Press.

40.

Spielberger

C. D.

(1983). State-Trait Anxiety Inventory manual. Redwood, CA: Mind Garden.

41.

Spielberger

C. D.

Gorsuch

R. L.

Lushene

(1997). State-Trait Anxiety Inventory for Adults. Palo Alto, CA: Consulting Psychologists Press.

42.

Tosevski

D. L.

Milovancevic

M. P.

Gajic

S. D.

(2010). Personality and psychopathology of university students. Current Opinion in Psychiatry, 23, 48-52.

43.

Twenge

J. M.

(2000). The age of anxiety? Birth cohort change in anxiety and neuroticism, 1952-1993. Journal of Personality and Social Psychology, 79, 1007-1021.

44.

Twenge

J. M.

Gentile

DeWall

C. N.

Lacefield

Schurtz

D. R.

(2010). Birth cohort increases in psychopathology among young Americans, 1938-2007: A cross-temporal meta-analysis of the MMPI. Clinical Psychology Review, 30, 145-154.

45.

Van Ameringen

M. V.

Mancini

Farvolden

(2003). The impact of anxiety disorders on educational achievement. Journal of Anxiety Disorders, 17, 561-571.

46.

Zivin

Eisenberg

Gollust

S. E.

Golberstein

(2009). Persistence of mental health problems and needs in a college student population. Journal of Affective Disorders, 117, 180-185.