The Trauma Symptom Inventory (TSI) was created by Briere, J. in 1995, and appears in the “Trauma Symptom Inventory Professional Manual.” The test is also featured in “Psychological Assessment Resources, Briere, et al., (1995). The test exists in form of manuals, ten item booklets, and twenty-five scorable answer sheets, with 25 each for female and male profile. The manuals can be obtained online from PAR, location http://www.parinc.com at $182 (Strauss et al., 2006). The alternative TSI-A test, is an eighty-six item test different from the TSI test for it does not contain sexual concern scale, critical items for sexual content, or dysfunctional sexual behavior.
Background and History of TSI
The TSI test under assessment in this review is designed for individuals of age 18 and above. The test was designed by Briere to assess psychological sequelae of traumatic events like sexual assault, physical violence, automobile accidents, combat, and natural disasters (Strauss et al., 2006). TSI has 100 items that measure four broad areas of distress including posttraumatic stress, self-dysfunction, dysphoric mood, and sexual difficulties.
The TSI was designed by Briere to assess chronic and acute traumatic symptomatology in adults. The first measurement with the scale was a survey of traumatic stress professionals published in Briere (1995). In the first clinical trial, Briere (1995) carried out a discriminant function analysis to test the relationship of the normative sample of TSI-T scores and the four modes of traumatic experiences (Purves & Erwin, 2004). The four types of traumatic experiences are childhood interpersonal violence, adult disaster, adult interpersonal violence, and childhood disaster.
Purpose and Content
The purpose of the TSI test is to assess chronic and acute traumatic symptomatology. The test has 100 items based on four broad distress areas including self-dysfunction, sexual difficulties, posttraumatic stress, and dysphoric (Purves & Erwin, 2004). Respondents to the test rate the symptom item based on the frequency of occurrence over the preceding six months. The rating uses a four-point likert scale from 0-never to 3-often (Ghetti et al., 2006). The 100 scale items are arranged into ten clinical scales, and three validity scales, and twelve critical items like self-mutilation. The test interpretations begin with assessment of validity scales and TSI scale examination as T scores65 (Strauss et al., 2006). The TSI scales are interpreted according to primary scales for trauma symptoms as DA, IE, D, TRB, DIS, ISR, and AA, and self-dysfunction symptoms as ISR, SC, DSB, AI, and TRB.
The test is administered individually or in groups, in which the respondents fill in the questionnaire on the given answer sheet. The questionnaire is timed and administered in twenty minutes time interval.
The total score of the test is in a measure that represents the effect of traumatic experiences, which can also be decomposed into three reliability scales and ten clinical subscales. The TSI’s ten subscales are depression, defensive avoidance, anger/irritability, anxious arousal, impaired self-reference dysfunction sexual behavior, sexual concerns, dissociation, tension reduction behavior, and intrusive experiences (Strauss et al., 2006). The three reliability scales entail response levels, within-scale inconsistent answers, and number of atypical answers. The TSI test has a very high internal consistency, convergent, test-retest reliability, and discriminate validity. This implies that the test applied to clinical samples, U.S. navy population, public, and students gave Cronbach alpha of .86 (Ghetti et al., 2006). Briere also created the test to have an internal consistency on clinical scales. In this scale, clinical, military, university, and standardization samples have a mean alpha of .87, .84, .84, and .86 respectively (Purves & Erwin, 2004). These scales also have incremental validity, predictive, and reasonable convergent. The test was tried on a psychiatric inpatient sample, and identified 89% of those the respondents to have a diagnosis of borderline personality disorder.
TSI scoring entails the identification of each item on a scale, with the raw scale of the main scales making up the total of the scale. This is applied to all scores DA, IE, D, TRB, DIS, ISR, ISR, SC, DSB, AI, TRB, and AA, except INC and RL. Scores in the TSI test reflect ten symptoms including Depression (D), Anxious Arousal (AA), Dissociation (DIS), and Intrusive Experiences (IE) (Ghetti et al., 2006). Other are Anger/Irritability (AI), Dysfunctional Sexual Behavior (DSB), Sexual Concerns (SC), Tension Reduction Behavior (TRB), and Impaired Self-Reference (ISR). In the Response Level (RL) scale, the number of items endorsed with zeros is recorded (Strauss et al., 2006). However, in the inconsistent responses INC scale, the absolute values of specific item pairs’ differences is summed. These raw scores are taken to the age-based profile that gives the T-score, which has a standard deviation of 10 and mean of 50. In the TSI test, all scores that are 65 and above are taken as clinically significant (Purves & Erwin, 2004). However, a significant different with other tests of dysfunctional behavior is the separation of female and male profiles.
Standardization and Normative Sample
Briere designed the test and standardized it a random sample of women and men from a population of N=828. The sample population was 18 years and older, and had different norms for female and male navy recruits creating N=3659 (Purves & Erwin, 2004). The TSI score also provides different norms for different combinations of age and sex, which include 18-54 and 55 or older, making it appropriate for all adult sexes. Notably, the test scores slightly vary on the race function at 2-3%, leading Briere (1995) to make suggestions racial groups (Strauss et al., 2006). To complete the TSI test, respondents are required to complete successfully a 5th-7th grade reading skill.
To derive this population, Briere (1995) used a general population selected using a random selection that mailed out questionnaires to 836 Americans. The mean age of the sample population was 47.3 years with a range of 18-88 and standard deviation of 16.6%. In the sample, 57.1% were married persons, 50.8% male, 77.5% Caucasian, 16.5% single, 16.6% separated, 6.1% Hispanics, 10.3% African Americans, 2.3% Native Americans, and 2.9% Asian (Purves & Erwin, 2004). Normative data for the TSI test scales were obtained from raw score data of standardized samples. The test had separate norms for a sample population of 3659 female and male navy recruits.
Briere (1995) identified the internal reliability correlation of the scales against a standardized for a clinical sample N-836 for each number of items on the test scale. On the validity scale, a response level (RL) has an alpha of .80 for ten items, atypical response (ATR) has .75 for ten items, and inconsistent response (INC) has .51 alpha coefficients for twenty items (Strauss et al., 2006). On the clinical scale, anxious arousal has an alpha coefficient of .86 for eight items, depression has .91 coefficients for eight items, and anger/irritability has .90 coefficients for nine items. Intrusive experience has an alpha .89 for eight items, defensive avoidance with .90 for eight items, and dissociation with .82 for nine items (Ghetti et al., 2006). Dysfunctional sexual behavior has an alpha of .85 for nine items, impaired self-reference with .88 for nine items, and tension reduction behavior with .74 for eight items (Strauss et al., 2006). Briere (1995) identifies that the mean inter-correlation of the reliability scores for the 10 clinical scale are internally consistent with an alpha mean coefficient of .86 for N=836, .84 for university sample of N=279, .87 for clinical sample of N=370 and .85 for military samples of N=3659.
To test the validity of the TSI test, Norris & Raid (1997) did a discriminant function analysis with the TSI’s standardized clinical scales to predict PTSD. The results were compared with the Brief Symptom Inventory subscales and Impact of Event Scale. The results of the comparison indicated that TSI scales are correlated with PTSD. The prediction for TSI was 91% for negative cases of PTSD, where the test identified 385 out of 423 negative cases (Strauss et al., 2006). This validity test proves that the TSI is valid for use on a sample of borderline personality disorder. The validity test also conceptualized the TSI test on three constructs. These are traumatic stress DIS, DA, IE, dysphoria, AA, AI, D, and self-references ISR, AI, TRB, and SC.
The test requires equipment and materials including an item booklet, test manual, answer sheet, computer scoring program, and scoring/profile sheet. These are present in form of Windows XP, Vista, 7, and 8 versions for Mac and Windows users. The advantage of the TSI test is its manuals, booklets, kits, and scoring equipment and software is readily available online at a cost effective price. The score and profile equipment and software are easy to use and create score reports and graphic profiles of the individual’s results. This is vital since it indicates the relative extent of the person’s experiences posttraumatic stress and other traumatic experiences. Moreover, the equipment has clear graphic images that indicate a change in the score report. This change in report helps clinicians to determine the significant changes that are occurring in the patient’s behavior over time.
Ease of Administration, Scoring, and Interpretation
The instrument is ideal for administering to individuals and groups. It is easy for the clinician and patient for the response item list on the 4-point scale is entered on the top page of the booklet. The booklet is carbonless, such that item responses are transferred to the scoring sheet underneath it automatically. This allows for easy administration and scoring by hand. It is also easy to interpret the scores since the graphic profiles convert raw scores to sex- and age-appropriate T scores and graphical representations.
The advantage of the Trauma Symptom Inventory test is its ability to assess a broad range of symptoms including those related to acute stress disorder and posttraumatic disorder. This is because the test has a ten clinical scale and three-validity scale. The second advantage is that the test is self-administered, easy to read for an individual with and above a 5th grade reading level.
Summary Evaluation and Critique
The TSI test is readily available in form of manuals, booklets, score sheets, and profile from both vendors and online sources, and is relatively affordable for it costs between $182-$250. The second strength of the test is that though it was designed for persons over the age of 18, it can assess trauma and trauma symptoms in children. In addition, it can be used to assess a wide range of traumatic events from sexual assault, physical violence, automobile accidents, combat, and natural disasters. Fourthly, the test is easy to administer, easy for the respondent to follow and read, easy for the clinician to score.
However, a closer examination of the ATR scale of the TSI test reveals that it is flawed for Briere (1995) confounded it with experiences of complex PTSD. The result of this confounding is it makes it difficult for the clinician to evaluate and discriminating feigned and genuine responses for severely traumatized patients.
To eliminate the weakness future studies and revised editions of the test should:
Create scales by considering several detection strategies for feigned symptoms in severely traumatized patients. This can be through assessment of rate symptoms, symptom selectivity, symptom combinations, and symptom severity.
Future revised editions can check trends in feigned and genuine symptoms through the detection of comorbidity in the scale and embedding feigning scales.