Article citationsSpielberger, C. Test Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.has been cited by the following article:.TITLE:AUTHORS:,KEYWORDS:,JOURNAL NAME:,June29,2011ABSTRACT: The present study examined the psychometric properties of the Greek version of Spielberger (1980) self-report measure of test anxiety, the Test Anxiety Inventory (TAI). The total sample consisted of 231 undergraduate stu-dents (124 male, 107 female). The results verified the well established two-factor structure for the TAI. The two factors represented the Worry (TAI-W) and Emotionality (TAI-E) subscales, respectively. Furthermore, on the bases of the confirmatory factor analyses, using either the set of 20 items or the set of 16 items, we found con-vincing support for the existing relationship between the two subscales of the Test Anxiety Inventory.
The in-ternal consistency of the twenty-item TAI-T scale and for the eight-item Worry and Emotionality subscales ranged from Cronbach’s α =.81 to.94. The G-TAI and its subscales showed differential statistically significant relationships with a self-report measure of cognitive interference.
Internal consistency coefficients for the scale have ranged from.86 to.95; test-retest reliability coefficients have ranged from.65 to.75 over a 2-month interval (Spielberger et al., 1983). Test-retest coefficients for this measure in the present study ranged from.69 to.89. ABSTRACT: The present study examined the psychometric properties of the Greek version of Spielberger (1980) self-report measure of test anxiety, the Test Anxiety Inventory (TAI). The total sample consisted of 231 undergraduate stu-dents (124 male, 107 female). The results verified the well established two-factor structure for the TAI.
IntroductionAnxiety has negative effects on the cognitive performance and psychosocial adjustment of elderly people. Given the high prevalence of anxiety symptoms in patients suffering from cognitive impairment, it has been suggested that these symptoms may be an early marker of dementia. The State-Trait Anxiety Inventory (STAI) is one of the widely-used used scales for evaluating anxiety in elderly people. However, inasmuch as the STAI may be difficult to apply to older people, having a short form of it would be desirable. MethodsThe participants comprised 489 community-dwelling individuals aged 68 years and over. All of them were volunteers in a longitudinal study for early detection of Alzheimer’ Disease (Proyecto Vallecas). The full sample was divided in two homogeneous subgroups: Group A, used to reduce the number of items and response options, and Group B, the group used to determine the psychometric properties of the new short form (STAIr).
ResultsA dichotomous Rasch model was used to obtain the STAIr. No statistically significant differences for STAIr scores were found with respect to sociodemographic variables. Psychometric properties and normative data were obtained for the new short-version.
ConclusionsThe STAIr is composed of 13 items and data fits the model well. Since it is short and easy to apply to elderly people, STAIr will be very useful in clinical and research settings. The strong correlation between the 3 factors identified in the exploratory factor analysis ( r = 0.68) suggested that all items revolve around the same one-dimensional concept of anxiety. As a result, no items had to be deleted before the IRT could be applied to the STAI subscales.The dichotomous Rasch model was applied to the new item coding (1,0).
The state anxiety STAI subscale was analysed in 2 consecutive stages. Only 6 items were shown to fit the model; consequently, the new state anxiety subscale included items 1, 5, 8, 12, 18, and 19 ( χ 25 = 370.59; P. Once the short STAI subscales had been created, we studied their properties with group B. Results showed an adequate correlation between STAI and STAIr for both the state anxiety and the trait anxiety subscales ( r = 0.81 and r = 0.80, respectively). The reliability coefficient was higher for the STAI state anxiety subscales ( λ 4 = 0.86) and trait anxiety subscales ( λ 4 = 0.81) than for the STAIr subscales ( λ 4 = 0.58 and λ 4 = 0.60, respectively).We subsequently studied the potential influence of sociodemographic variables on both STAIr subscales since these variables are so important when obtaining normative data for the inventory. No significant differences were found in the state anxiety STAIr subscale regarding age (Kruskal–Wallis χ 22 = 2.75; P =.253), sex (W = 2.75; P =.050), or educational level (Kruskal–Wallis χ 22 = 1.51; P =.471). The trait anxiety STAIr subscale showed no significant differences for age (Kruskal–Wallis χ 22 = 0.05; P =.975), sex (W = 963; P =.762), or educational level (Kruskal–Wallis χ 22 = 0.61; P =.738).
Therefore, none of these variables was used to stratify STAIr normative data. Shows the equivalence between raw scores on both subscales, together with percentile ranks and the corresponding scaled scores. Raw scores equal to or higher than 3 or 4 were used as the cut-off point for state anxiety and trait anxiety STAIr subscales, respectively. DiscussionAnxiety symptoms are highly prevalent among the elderly. Presence of these symptoms has been associated with decreased processing speed and poorer cognitive performance in elderly patients with normal cognitive function.
Likewise, patients diagnosed with cognitive impairment display higher anxiety levels than cognitively healthy subjects. This finding has led researchers to consider anxiety as a potential marker of cognitive decline and even as a predictor of cognitive impairment conversion to dementia. In light of the above, assessing level of anxiety in cognitive impairment units may be particularly relevant for the diagnosis and prognosis of cognitive decline.The STAI is used internationally to evaluate anxiety symptoms. However, it may be too long and difficult to complete in some clinical settings. Some researchers have therefore been particularly interested in developing a short version. Several versions have been created to measure anxiety levels in pregnant women, patients with mechanical ventilatory support, or patients scheduled for surgery with general anaesthesia, and these versions are based on different statistical techniques, including IRT.Assessing elderly patients with STAI and other such subjective scales with polytomous items has several disadvantages: patients may be confused and unable to understand the question, and they may even find it difficult to choose a response option since they may not understand ordinal scales.
In contrast, dichotomous items are much more suitable for the elderly population since they reduce confusion and uncertainty about answers. The purpose of the present study was to design a short questionnaire that can be completed in a short time and which evaluates the presence of anxiety symptoms in elderly patients. To this end, we administered the STAI to a sample of elderly subjects with normal cognitive function and applied the IRT to items on both STAI subscales.Psychometric analyses determined the structure of the new short versions of the state anxiety and trait anxiety subscales, which were made up of 6 and 7 dichotomous items, respectively.
According to the bivariate analysis, STAIr scores were not influenced by either age, sex, or educational level. These findings reveal that anxiety levels in elderly patients are less influenced by sociodemographic variables in the new short version of STAI. However, this hypothesis should be interpreted with caution; further studies are necessary to prove its validity.Furthermore, we found striking disparities between anxiety scores in our sample and those obtained in other studies including young adults. Some studies have shown that elderly people are eager to present themselves as socially desirable and therefore score higher on social desirability scales than younger individuals. This may lead anxiety levels in elderly subjects to appear lower than they actually are.Reducing the number of items is associated with decreases in scale reliability, which may explain why the STAIr subscales showed decreased reliability. On the other hand, reducing the number of response options is also problematic: decreased variance of total scores leads to a decreased reliability index. Both of these factors could explain the reduced reliability of the STAIr.
However, reliability indices for the STAIr are moderate and there is a strong correlation between the STAIr and the STAI subscales. We can therefore state that the STAIr has acceptable psychometric properties.In conclusion, the negative impact of anxiety disorders on psychosocial adjustment and cognitive function in the elderly has led many authors to suggest that anxiety is an early predictor of cognitive impairment and of conversion to dementia. Although anxiety is not usually evaluated in clinical and research settings, tools for measuring the level of anxiety in the elderly are necessary, especially in cognitive impairment research. This study describes modifications to the STAI, one of the most widely-used scales for anxiety, to create a short version, the STAIr.
This scale demonstrates appropriate psychometric properties and can be applied quickly and easily to elderly subjects. Conflicts of interestThe authors have no conflicts of interest to declare.