The State Trait Anger Expression Inventory (STAXI)

 

---

The STAXI was developed with two goals in mind.. The first was to develop a measure of the components of anger in the context of both normal and abnormal personality. The second goal seems rather more specific to a particular research orientation - in this case examining the contribution of anger to the development or exacerbation of medical conditions such as hypertension, coronary heart disease, and cancer.

 

            As the name implies there are two fundamental aspects of anger which are addressed - the experience of anger, and the expression of anger. The experience of anger can be understood in the context of state - subjective feelings that vary from irritability to intense rage, and trait anger which refers to a disposition to perceive situations as annoying and to respond to these situations by more frequent expressions of state anger. Thus state and trait anger are unlikely to actually be independent characteristics or components of anger.

 

            When expressing anger, it may be focused outward on other people or objects (Anger-Out), or directed inward (Anger-In). A third component is the degree to which people attempt to control their expression of anger (Anger Control).

 

            The STAXI is designed to be administered to people aged 13 through adulthood with a minimum fifth grade reading level. The task of administering the test is straightforward, essentially self-administered, but interpretation of test scores requires formal training in assessment. There are no time limits imposed on completing the STAXI, but it is a brief test and most people complete it within 15 minutes.

 

STAXI Scales:

            The STAXI consists of 44 items which are distributed across the five main scales. Consistent with the conceptualisation of anger above there are three main aspects to the STAXI scales: State, Trait, and Anger Expression. Trait contains two subscales that examine different dispositions in trait anger - temperament and reaction. Anger Expression is actually an experimental composite of the three expression constructs -In, Out, and Control.

 

 

S-Anger - This is a 10-item stand-alone scale which measures the respondent's current feelings of anger.

 

T-Anger - This scale also contains 10 items which asks the respondent to answer questions about his or her disposition towards anger.

            T-Anger/T - is a subscale of T-Anger consisting of 4 items that generally address the disposition to express anger without provocation.

            T-Anger/R - is a subscale of T-Anger also consisting of 4 items that ask about the respondent's disposition to express anger when provoked.

 

AX/In - this 8 item scale measures the frequency with which the respondent holds in or suppresses his or her anger.

AX/Out - this 8 item scale measures the frequency with which the respondent expresses her anger to other people or objects.

 

AX/Con - another 8 item scale that attempts to measure the degree to which the respondent attempts to control his or her expression of anger.

 

AX/EX - this is an experimental composite score that is designed to represent the combination of AX/In, AX/Out, and AX/Con and essentially examines the overall frequency of anger expression. AX/EX is computed with the following formula:

            AX/EX = AX/Out + AX/In - AX/Con +16.

The addition of 16 at the end may seem unusual but it is designed to ensure that a negative score cannot be achieved. Since the four possible responses to each item are assigned a number between 1 and 4 and each of the AX scales has 8 questions, the minimum total for each scale is 8, and the maximum score is 32. The lowest possible score would be minimum In and Out and maximum Con which would be 8+8-32 = -16. Adding 16 to this total would give 0. Similarly if Out and In were at maximum and Con was at minimum the total would be 32+32-8+16 = 72. For this reason the raw score range for AX/EX is 0 to 72.

 

All items are rated on a four-point scale and are assigned a score of between 1 and 4.

Raw score totals are converted to percentile ranks and T-scores using normative tables. There are separate normative tables for males and female adolescents, adults, and college students. There are also special population normative tables for .

 

Standardisation:

The normative sample for the STAXI is an impressive 9,000 subjects broken down in the following way:

 

3,769 Adults. Mean age of the combined sample was 40 years. This group was comprised of heterogenous samples that included managerial, technical, clerical, sales, and factory workers (360 males, 328 females); people participating in a stress management program (202 males, 93 females); health care managers (16 males, 131 females); military personnel (1890 males, 104 females) and a small sample of postal workers (33 males, 23 females).  Don't bother adding it up - they're 589 short!

 

2,762 Undergraduate college students the vast majority of whom (2,122 - 777 males, 1,345 females) were enrolled in Introductory Psychology units. The remaining 640 (600 males, 40 females) were cadets at West Point (U.S. Military Academy). Notice that while the combined gender totals are very similar (1377 males, 1385 females) they are the product of two extremely gender biased samples. Median age was 19 years.

 

2,469 adolescents in junior and senior high schools. Median age was 14 years.

 

Influence of Demographic Variables:

Age:

            Age appears to have a profound effect on STAXI scores. Adolescents score highest and adults score lowest on all measures - this is used to justify the use of these three separate normative groups. However, within the adult group there is also a significant effect of age - responses by 18-30 year olds were higher than those of the 31-40, and 41+ age groups. Interestingly, no separate norms are made available for 18-30 year-olds and 31+ adults.

 

Sex:

            Minimal differences were found across the three samples in terms of the gender of respondents. Again this is interesting as the norms provided are separate for males and females. The variable that does not impact on scores (sex) is accounted for and the variable that does impact on scores (age) is not - at least in the adult sample.

 

Interpreting Scores:

            Raw scores are converted into percentile ranks for interpretative purposes. Scores that are commonly found in the middle fifty percent of the distribution - the 25^th to 75^th percentiles are considered to fall within the normal range. Individuals with high anger scores (above the 75^th percentile) are endorsing levels of state, trait, expression, or control that are likely to impair their optimal functioning. Very high scores would fall at or above the 90^th percentile. The distributions for S-Anger and T-Anger/T are so skewed that low scores are unlikely to be interpretable. Individuals who score below the 25^th percentile on the remaining six scales, however, generally experience, express, or control relatively little anger.

 

 

 

Reliability:

 

	Adults		College Students		Adolescents
Scale	Males	Females	Males	Females	Males	Females
S-Anger	.90	.91	.93	.90	.87	.90
T-Anger	.82	.82	.82	.83	.82	.84
T-Anger/T	.89	.88	.88	.89	.85	.85
T-Anger/R	.69	.69	.71	.70	.65	.70
AX/In	.86	.81	.73	.74	.84	.81
AX/Out	.75	.78	.75	.77	.73	.75
AX/Con	.81		.85	.84