The Life Attitudes Schedule Short Form: An Abbreviated Measure of Life-Enhancing and Life-Threatening Behaviors in Adolescents
The authors would like to thank the students, high school staff, and parents for their participation in this project. A special thanks to Richard Langford for his participation in this project. We are also grateful to the research assistants form Humboldt State University, the University of Oregon, and Oregon Research Institute for their efforts in the collection of these data.
Abstract
An abbreviated version of the Life Attitudes Schedule (LAS) was developed, consisting of 24 items, each representing one cell of the original LAS theoretical matrix (4 content categories × 2 behavior types × 2 valence). Items were retained on the basis of high correlations with LAS total score and low correlations with gender. Psychometric properties of the LAS Short Form were robust and the Short Form total score correlated .93 with the original LAS total score. As with the original LAS, boys reported more injury-related behaviors than girls. Future research and clinical directions are suggested.
It has been known for some time that problematic behaviors tend to cluster within a subset of adolescents (e.g., Jessor & Jessor, 1977). These behaviors include cigarette, alcohol, and illicit drug use; general deviant behaviors (e.g., theft, vandalism, shoplifting, aggression, lying); and early sexual behavior (e.g., Donovan & Jessor, 1985). More recently, these adolescents have been shown to be at increased risk for psychological difficulties, such as depression (e.g., Jessor, 1991). Many of these problematic behaviors are potentially life-threatening. In fact, each of the three leading causes of death among adolescents 15 to 24 years of age (i.e., accidents, homicide, and suicide—National Center for Education Statistics, 1993) can be presumed to be associated with problematic behaviors, especially risk-taking and substance use (Clark, Sommerfeldt, Schwartz, Hedeker, & Watel, 1990; Garrison, McKeown, Valois, & Vincent, 1993). As another indication of the high levels of engagement in risk-taking behaviors among adolescents, the rate of AIDS, a disease that is often a consequence of the individual's sexual or drug use behavior, is thought to be doubling each year in this age group (e.g., Biglan et al., 1990). Given that the occurrence of risk-taking and potentially life-threatening behaviors tend to cluster, better methods of identifying the youth who are especially prone to engage in these behaviors are needed for prevention and intervention strategies.
Given its particularly negative potential, more attention has been focused on the assessment of suicidal behavior than other risk-taking and life-threatening behaviors of adolescence. Although considerable strides have been made in this area, several critical issues have limited progress (e.g., Garrison, Lewinsohn, Marsteller, Langhinrichsen, & Lann, 1991; Lewinsohn, Garrison, Langhinrichsen, & Marsteller, 1989), including insufficient attention to issues of validity, lack of specification regarding the purpose of the instrument and the theoretical model underlying the measure, and lack of prospective research. In addition, most instruments focus primarily, often exclusively, on suicidal ideation and aim to assess behaviors at only a very high threshold of severity. This narrow focus often results in low levels of item endorsement, and more importantly, may lead to the instrument missing important relationships between suicidal behavior and other risk-taking and life-threatening behaviors. Given that a clear need remained for a psychometrically sound measure of a broad array of life-threatening behaviors (including suicidal behaviors), the 96-item Life Attitudes Schedule (LAS—Lewinsohn, Langhinrichsen-Rohling et al., 1996) was developed. The goal of the current study was to develop an abbreviated version of this measure that could be easily administered in a variety of clinical and research settings.
The original LAS was designed on the basis of an articulated theoretical model, in which behaviors—separated into actions, thoughts, and feelings—were hypothesized to fall along a continuum from positive to negative. Behaviors were categorized into four content categories thought to be sufficient to encompass the entire domain of potentially life-threatening behaviors: (1) death-related (DR) behaviors, which included traditional suicide and death-related items as well as items regarding life and longevity; (2) health-related (HR) behaviors, which included items assessing illness, lack of self-care, health, and wellness; (3) injury-related (IR) behaviors, which included injury, risk-taking, and safety-related items; and (4) self-related (SR) behaviors, which included behaviors that either enhance or compromise one's self-worth, as well as items describing accomplishments or self-image. Items were generated to complete all of the 24 cells in a balanced matrix (4 content categories × 3 behavior types × 2 valence). Care was taken to create and select items that would share only moderate variance with self-reports of depression, hopelessness, and social desirability, as these variables have been shown to be related to suicidal behavior but neither necessary nor sufficient factors in predicting actual suicide attempts (Cole, 1988; Holden, Mendonca, & Serin, 1989).
Extensive previous research (Lewinsohn, Clarke et al., 1996) suggests that the 96-item LAS is a psychometrically sound instrument. One-month test–retest correlations and measures of internal consistency (coefficient alpha) for the main scales were quite high (e.g., alphas ranged from .72 to .92; 1-month test–retest correlations ranged from .71 to .90). Moreover, the scale correlations with LAS total score remained robust after controlling for depression, hopelessness, and social desirability.
Support was also evident for the theoretical model underlying the LAS. As predicted, a high degree of association was obtained between latent constructs assessing the positive and negative poles, suggesting that this domain can be conceptualized as a bipolar continuum of life-enhancing and life-threatening behaviors. Structural equation analyses found that all three behavior types (i.e., actions, thoughts, and feelings) loaded about equally on the higher-order construct of proneness for life-shortening behavior. Compared to the three behavior types, latent constructs representing the four content categories (DR, HR, IR, ST) were less highly interrelated, and none of their path coefficients were as high as those for the three behavior types. The death-related (DR) construct was most strongly related to the other content categories, and the injury-related (IR) construct was least. It is also noteworthy that a subsequent analysis of the LAS data (Langhinrichsen-Rohling, Lewinsohn, Rohde, Seeley, & Langford, 1995) suggested that boys and girls have significantly different LAS content category profiles (e.g., boys report much higher rates of injury-related behaviors; girls report higher rates of self-related behaviors), which argues for the potential importance of differentiating the four content categories.
Although the initial findings from studies conducted with adolescents confirmed that the LAS is a psychometrically sound instrument, its length was seen as a potentially serious restriction to its use, particularly in clinical settings. To be practical, a screener would need to be considerably shorter. Thus, the goal of this study was to develop an abbreviated version of the LAS that retained adequate psychometric properties.
METHOD
Subjects and Procedures
Sample 1. In 1993, 638 high school students from a high school in northern California participated in a study to determine the psychometric properties of three versions of the LAS (the main form and two alternate versions). Subjects were recruited through passive consent from the parents and active consent from the adolescents (the participation rate was approximately 80%). Slightly less than half of the subjects (48.1%) were female. The sample was predominantly Caucasian (80.4%); the remaining subjects were 1.1% African American, 5.6% Hispanic, 1.8% Asian, 5.3% Native American, and 5.8% other. Although data on age were not available, approximately equal numbers of participants were freshmen (32.1%), sophomores (27.9%), juniors (20.2%), and seniors (19.7%).
A 30-day test–retest of the LAS and other measures was conducted with 412 of the participants. The LAS readministration (T2) was designed to obtain both test–retest data and correlational data across the three forms; the majority of subjects who repeated the questionnaires (n = 238) received a different version of the LAS, to provide data regarding alternate form comparability; all combinations of the repeated forms were administered. The remaining 174 subjects completed the same LAS form at T2, providing test–retest reliability.
All items for the LAS Short Form were selected from the main LAS form, which was completed by 206 adolescents at the first administration. Complete LAS data (i.e., no missing items) were available for 188 of the adolescents. Of these, 55 subjects repeated the same form at the readministration.
Assessments were coordinated by trained graduate research assistants with teachers present to assist with distribution and collection of the paperwork. Subjects at both assessments had 50 minutes to complete the questionnaires, which was sufficient for 97.9% of the adolescents to complete the assessment battery. Payment consisted of a donation of $ 1 to the high school for every student who completed the questionnaires.
Sample 2. In 1990, 32 adolescents were recruited from two ongoing research studies; 16 from an outcome study evaluating group treatments for adolescent depression (Lewinsohn, Clarke, Rohde, Hops, & Seeley, 1996) and 16 from a community epidemiologic study of depression and other psychopathology in adolescents (Lewinsohn, Hops, Roberts, Seeley, & Andrews, 1993). Participants in the treatment outcome study met DSM-III-R (American Psychiatric Association, 1987) criteria for major depression (n = 11) or dysthymia (n = 5). They were matched on age and gender with 16 participants from the community study who had no past or current psychiatric diagnoses.
Data collection occurred on one occasion, in which subjects completed the LAS and participated in an interview assessing current and past suicidal ideation and past suicide attempts. All subjects were paid for their participation. Data from Sample 2 provided an opportunity to evaluate the associations between LAS Short Form scores with actual suicidal behavior.
Measures
The Life Attitudes Schedule. Half of the LAS items were designed to assess negative (life-threatening) behaviors and the other half assessed positive (life-enhancing) behaviors. LAS items were developed for each of the four content categories: death-related (DR), health-related (HR), injury-related (IR), and self-related (SR); and included an equal number of items representing actions, thoughts, and feelings. Thus, items on the LAS can be categorized into a 4 (content area) × 3 (behavior type) × 2 (valence) matrix consisting of 24 cells. Four items were chosen for each cell in the original LAS for a total of 96 items per form. In scoring the LAS, the positive items are reversed. Therefore, the maximum score of 96 indicates the endorsement of all 48 negative items and denial of all 48 positive items. Similarly, the minimum score of zero occurs when a subject endorses none of the negative items and all 48 of the positive items.
Items were chosen for the LAS based on extensive preliminary psychometric data on earlier versions of the form. Items were selected on the basis of low correlations with social desirability, gender, age, depression, and hopelessness; and high correlations with the LAS. In addition, retained items had to be endorsed by at least 5% but not more than 95% of the sample.
The Center for Epidemiologic Studies Depression Scale (CES-D). This is a 20-item self-report measure of depressive symptomatology, which has been shown to have good psychometric properties (Radloff, 1977), and has been successfully used to assess depression in adolescent populations (Roberts, Andrews, Lewinsohn, & Hops, 1990).
The Hopelessness Scale (Beck, Weissman, Lester, & Trexler, 1974). This is a 20-item true–false inventory designed to measure lack of hope about the future. Beck et al. (1974) report internal consistency ratings of .93 for this measure. This scale has been found to be a better predictor of future suicide in a sample of hospitalized adult patients than measures of depression (Beck, Steer, Kovacs, & Garrison, 1985).
The Marlowe-Crowne Social Desirability Scale (Crowne & Marlowe, 1960). Subjects were given an abbreviated version of this scale, which measures the degree to which subjects are responding in a socially desirable way. The 6-item true–false short form has been shown previously to have good reliability and validity properties in an adolescent population (Andrews, Lewinsohn, Hops, & Roberts, 1993).
The Infrequency Scale of the Personality Research Form (Jackson, 1976). This scale consists of items that the vast majority of the population have been found to answer in a certain direction (e.g., “There have been times when I have dialed a telephone number only to find that the line was busy”). Five items from this scale were included to identify subjects who were not legitimately reading and responding to the test questions. Subjects who endorsed three or more items in the infrequent direction (1.9%) were excluded from all analyses.
RESULTS
Item Selection for the LAS Short Form
It was considered important to maintain the theoretical matrix underlying the original LAS. In addition, we were interested in constructing a single LAS Short Form that could be used equally well with both males and females. Therefore, the LAS Short Form was constructed by selecting the single item in each of the 24 cells (4 content categories × 3 behavior types × 2 valence) that had (1) the highest item correlation with the original LAS total score, and (2) a nonsignificant association with gender (in 6 of the 24 cells, the item with the highest LAS Total score correlation was not selected due to a significant association with gender). The 24 items comprising the LAS Short Form, with their respective categorization and frequency of endorsement, are shown in Table 1. The mean item correlation (absolute value) of items with LAS total score was r = .46 (p < .001) and ranged from .30 to .59. The mean correlation (absolute value) of items with gender was r = .10 (NS; range = .01 to .17).
| Item | Categorya | % answering true | |
|---|---|---|---|
| 1. | I take care of my possessions so that they will last as long as possible. | ADR + | 85.1 |
| 2. | I choose to listen to music that has a death related theme (e.g., some heavy metal music like Ozzy Osbourne's “Suicide Solution”). | ADR − | 18.1 |
| 3. | I try to eat foods that are good for me. | AHR + | 71.3 |
| 4. | I have gone on occasional drinking sprees. | AHR − | 38.8 |
| 5. | I avoid unnecessary risks. | AIR + | 62.8 |
| 6. | At least once a month I have driven or have been driven more than 20 miles per hour over the speed limit. | AIR − | 54.3 |
| 7. | I rarely do things that violate my standards. | ASR + | 72.3 |
| 8. | I spend a lot of time doing things that are unproductive or unfulfilling. | ASR − | 30.3 |
| 9. | I look forward to a long life. | FDR + | 92.6 |
| 10. | I enjoy thinking about death. | FDR − | 9.6 |
| 11. | I enjoy eating “right.” | FHR + | 67.0 |
| 12. | I don't really care much about what I eat (e.g., fried foods, sugar, etc.). | FHR − | 46.3 |
| 13. | I enjoy spending time with people who are cautious and avoid unnecessary risks. | FIR + | 48.4 |
| 14. | Sometimes I feel so frustrated that I would like to hit my fist against the wall (or do something that could hurt me). | FIR − | 58.5 |
| 15. | I feel good because any activities are meaningful and have purpose. | FSR + | 84.0 |
| 16. | I wish that I was someone else. | FSR − | 20.7 |
| 17. | I expect to have a long and interesting life. | TDR + | 89.9 |
| 18. | Killing myself would solve many of my problems. | TDR − | 10.6 |
| 19. | It is important to brush one's teeth after every meal. | THR + | 66.5 |
| 20. | The danger of smoking cigarettes has been exaggerated. | THR − | 10.1 |
| 21. | The chance of my being injured in an accident in the next year is very low (less than 10%). | TIR + | 61.7 |
| 22. | Sometimes I think about injuring myself (e.g., smashing my fist into a window). | TIR − | 21.8 |
| 23. | I believe that I am a good person. | TSR + | 92.6 |
| 24. | I think that I am worthless. | TSR − | 6.9 |
- a The first letter of the category code indicates the behavior type (A = action, F = feeling, T = thought). The second letter indicates content category (DR = death-related, HR = health-related, IR = injury-related, SR = self-related behavior). The sign indicates valence (negative, positive).
Psychometric Properties of Short Form Scales
Psychometric properties (i.e., internal consistency and 30-day test–retest reliability) of the LAS Short Form scales, along with mean level of endorsement, are shown in Table 2. For comparison purposes, internal consistency and test–retest reliability values for the same scales in the original 96-item LAS are shown in the last two columns. The mean alpha of the eight scales in the Short Form was .65, compared to .83 in the original LAS. Mean test–retest values for the scales dropped slightly from .82 in the original LAS form to .73 in the Short Form. Both test–retest values indicate good temporal stability for the LAS.
| LAS Short Form | Original LAS | ||||||
|---|---|---|---|---|---|---|---|
| No. of items | Mean (SD) | Alpha | T1–T2r | No. of items | Alpha | T1–T2r | |
| LAS Total | 24 | 6.32 (4.39) | .84 | .84 | 96 | .92 | .85 |
| Actions | 8 | 2.43 (1.73) | .60 | .71 | 32 | .82 | .80 |
| Thoughts | 8 | 2.50 (1.83) | .59 | .70 | 32 | .82 | .90 |
| Feelings | 8 | 1.39 (1.54) | .64 | .70 | 32 | .81 | .71 |
| Death-related | 6 | 0.71 (1.18) | .67 | .70 | 24 | .77 | .82 |
| Health-related | 6 | 1.90 (1.60) | .62 | .75 | 24 | .77 | .80 |
| Injury-related | 6 | 2.62 (1.65) | .58 | .67 | 24 | .85 | .88 |
| Self-related | 6 | 1.09 (1.38) | .67 | .80 | 24 | .87 | .79 |
Also shown in Table 2 are the mean scores for the various LAS Short Form scales. On average, adolescents endorsed 6 of the items in the negative direction (i.e., either endorsed the negative behavior or denied the positive behavior). Scores were higher for thoughts and actions than for feelings. In the four content categories, endorsement was highest for the injury-related (IR) behaviors and lowest for the death-related (DR) behaviors.
Association of LAS Short Form with Associated Measures
Although efforts had been taken to ensure that the LAS not be synonymous with measures of depression and hopelessness, the LAS Short Form total score was significantly (p < .001) correlated with CES-D (r = .43), Hopelessness Scale (r = .61), and social desirability (r = .39). Individual scales within the LAS Short Form (actions, thoughts, feelings, DR, HR, IR, SR) were most highly correlated with hopelessness; the average correlation of scales with CES-D, Hopelessness Scale, and social desirability were r = .34, .50, and .30, respectively (all p < .01). The LAS Short Form scales were highly associated with LAS Short Form Total score, ranging from r = .58 to .88 (all p < .001). The mean correlation of scales with LAS Short Form Total score was .80 and dropped to .70 when associations with depression, hopelessness, and social desirability were statistically controlled.
Although each of the individual LAS items had been selected on the basis of a nonsignificant association with gender, the Short Form Total score was significantly related to gender, with males having a significantly higher average Total score than females [7.34 vs. 5.46; t(185) = 3.00, p < .01]. Males also had significantly higher scores on three of the Short Form scales: (1) Actions [2.80 vs. 2.12; t(185) = 2.72, p < .01]; (2) Thoughts [2.89 vs. 2.17; t(185) = 2.72, p < .01]; and (3) Injury-related [3.13 vs. 2.17; t(185) = 4.18, pk < .001]. Gender differences on the other LAS Short Form scales were nonsignificant.
Association of LAS Short Form with Original LAS
Comparability of the LAS Short Form with the original LAS was assessed by computing the correlational matrix of scales from the two versions. Correlations of identical scales across the two LAS forms (e.g., DR scale on LAS Short Form with DR scale on original LAS) appear in the diagonal of the correlational matrix shown in Table 3.
| Original LAS scalesa | ||||||||
|---|---|---|---|---|---|---|---|---|
| Total | Actions | Thoughts | Feelings | DR | HR | IR | SR | |
| LAS Short Form scalesa | ||||||||
| Total | .93 | .86 | .86 | .84 | .76 | .76 | .70 | .75 |
| Actions | .83 | .71 | .75 | .83 | .72 | .67 | .57 | .70 |
| Thoughts | .78 | .86 | .66 | .62 | .54 | .69 | .67 | .57 |
| Feelings | .79 | .64 | .83 | .72 | .73 | .60 | .54 | .68 |
| DR | .69 | .57 | .66 | .67 | .81 | .42 | .42 | .59 |
| HR | .70 | .70 | .64 | .58 | .47 | .85 | .43 | .49 |
| IR | .72 | .72 | .65 | .61 | .49 | .51 | .83 | .41 |
| SR | .71 | .58 | .66 | .70 | .61 | .47 | .36 | .84 |
- a Total = LAS total score; DR = death-related; HR = health-related; IR = injury-related; SR = self-related behavior. Significance of all correlations greater than p < .001.
Association of LAS Short Form with Suicidal Behavior
Using the 32 subjects of Sample 2, data were available regarding the ability of the LAS Short Form to identify subjects with lifetime suicidal ideation (n = 10). Only 4 subjects (all of whom had suicidal ideation) reported a past suicide attempt; therefore, screening characteristics of the LAS Short Form for past attempt were not computed. Classification characteristics in predicting lifetime suicidal ideation were maximized at an LAS Short Form Total score of 10. Thirteen of the 32 subjects scored at or above 10, of whom 8 had suicidal ideation. Thus, use of this cut-point identified 8 of the 10 ideators at the cost of 5 “false-positive” subjects. Translating these results into the terminology of screeners, sensitivity (i.e., proportion of true ideators correctly identified by LAS Short Form) was .80; specificity (i.e., proportion of true nonideators correctly identified by LAS Short Form) was .77; and positive predictive value (i.e., PPV, or the proportion of true ideators among those identified by LAS Short Form as ideators) was .62. Over three fourths of the adolescents (78.1%) were correctly classified by the LAS Short Form. These screening characteristics, especially the relatively high PPV, indicate that the LAS Short Form is a respectable screener for lifetime suicidal ideation. Use of the entire 96-item LAS was unable to improve upon these classification abilities; the optimal cut-point for the original LAS was 39 and correctly classified the same portion of subjects (i.e., screening characteristics of the entire LAS were identical to the LAS Short Form).
Looking at just the 24 girls in Sample 2 (7 of whom had lifetime suicidal ideation), the optimal screening characteristics were obtained the slightly lower cut-point of 7 (sensitivity = .86, specificity = .82, PPV = .67, percent correctly classified = 83.3). However, the cut-point of 10 resulted in equally respectable classification properties with the girls (sensitivity = .71, specificity = .88, PPV = .81, percent correctly classified = 83.3). The small number of males (n = 8) prevented examination of the screening characteristics of the LAS Short Form separately for the boys.
DISCUSSION
In the current study, a 24-item short form of the Life Attitudes Schedule was created. Each item included in the LAS Short Form represents one cell from the original LAS theoretical matrix (4 content categories × 2 behavior types × 2 valence). Extensive psychometric data were available on the 96 items (4 items per cell) contained in the original LAS. These data were used to determine which item should be retained in each cell for the LAS Short Form. Items were retained on the basis of a high correlation with the LAS total score and with a nonsignificant correlation with gender.
As anticipated, the chosen LAS Short Form items replicated the positive psychometric properties of the original LAS. Specifically, the obtained test–retest correlations indicated good stability for the shortened instrument. In addition, the Short Form content scales had moderately good internal consistency in this sample (e.g., ranging from alpha = .58 to .67), suggesting that they can still be used separately in addition to the total score. However, the LAS Short Form Total score had a very acceptable measure of internal consistency (alpha = .84), suggesting that the overall measure may be the most useful and robust score to consider. The high correlation between the LAS Short Form and the original LAS Total score (r = .93) indicates that the abbreviated form results in a similar rank ordering of subjects. In fact, squaring the correlations between the short and long LAS form Total scores reveals that the LAS Short Form explained 86% of the original LAS variance.
Surprisingly, gender differences were obtained on the LAS Short Form. This was unexpected given that items were specifically selected for inclusion on the basis of a nonsignificant correlation with gender. We had strongly considered developing separate LAS Short Forms for males and females but decided that a single form would be preferable for both practical and scientific reasons. The finding of gender differences on several of the LAS Short Form scales is consistent with results obtained on the original LAS (Langhinrichsen-Rohling et al., 1995). Specifically, males engage in significantly higher rates of injury-related behavior than do females. This is in keeping with past research showing higher rates of “foolhardiness” and recklessness in adolescent males than females (e.g., Clark et al., 1990). Although future research is needed to help clarify the importance of these gender differences, it is noteworthy that the majority of self-report devices for the assessment of suicidal behavior (ideation and attempts) find higher scores for girls than for boys (e.g., Lewinsohn, Rohde, & Seeley, 1996) in spite of the fact that males complete suicide at substantially higher rates than females (National Center for Education Statistics, 1993). Broadening the scope of the construct to include injury-related behaviors has resulted in identifying more males as engaging in life-threatening behaviors. This finding may have important prevention and intervention implications.
Finally, data in the current study were used to determine a cutoff score on the LAS Short Form. Results suggest that adolescents scoring 10 or more (either by endorsing a negative item or by failing to endorse a positive item) were correctly classified as having a past suicide attempt approximately 78% of the time. Not enough males were available in Sample 2 to evaluate the screening properties separately by gender, although comparable results were noted for the girls. Overall, reasonable sensitivity and specificity rates were obtained, and the screening capabilities of the LAS Short Form appear to be good. In fact, although the measure had been reduced to one fourth of its original length, screening characteristics for the LAS Short Form did not differ from the original LAS form.
Several limitations to the current study temper the positive findings. First, this research was conducted in the Northwest and the majority of subjects were middle-class Caucasian students. Future research is needed to determine the generalizability of these findings to other populations. Ideally, norms for this instrument would be developed on a nationally representative sample. Second, data regarding age were not obtained. Although unfortunate, there is probably not a great deal of variance in LAS scores over this short age span. Our research with suicidal ideation shows little variance across the high school years (Lewinsohn, Rohde, & Seeley, 1996), and exploratory analyses examining LAS Short Form item correlations with Short Form Total scores as a function of school grade did not reveal striking differences. More importantly, we do not know the limits of effectiveness of the LAS Short Form with older or younger subjects. Thus, the appropriate age limitations for use of the LAS Short Form are currently unavailable. It should also be noted that psychometric properties for any abbreviated scale need to be reexamined, as they may be lower in replication samples.
Until it has been more fully validated, the primary applications of the LAS Short Form appear to be in research settings, rather than clinical work. The first priority is to undertake a prospective study to determine the extent to which elevated scores on the LAS Short Form predict future engagement in life-threatening and life-enhancing behavior. Obviously, the best way to assess for current or past suicidal behavior is to ask specifically about those behaviors. The LAS Short Form could also be used to study the concurrent associations of a variety of positive and negative health behaviors. Another use of the LAS Short Form would be as the dependent variable in studies evaluating impact of interventions aimed at impacting suicidal and other life-threatening behaviors. This instrument could also be used to identify groups of adolescents at particularly high risk for engagement in a variety of life-threatening behaviors. Perhaps, one day, the LAS Short Form will function as a screener for identifying high-risk adolescents in schools, communities, and other large-scale settings. One purpose in developing the LAS was our belief that a more broadly defined measure was preferable to a number of much more narrowly defined instruments. We are unaware of any instrument that assesses as wide a domain of behaviors as covered by the LAS.
Previously, we established the reliability and validity of the 96-item Life Attitudes Schedule. We determined that this new instrument may be useful in identifying individuals whose rate of engagement in life-threatening and life-enhancing behaviors is problematic. In the current study, we have extended this line of research by developing a shortened version of the LAS that appears to maintain good psychometric properties. We believe that, over time, both the original LAS and the LAS Short Form will serve a variety of uses for researchers and clinicians working in the area of suicidal and risk-taking behaviors.
