Contact Barbara Rubel for Your Upcoming Event

Critique by Barbara Rubel

May, 1999

CRITIQUE OF

ARE TEACHERS OF CHILDREN AND YOUNG ADOLESCENTS RESPONSIVE TO SUICIDE PREVENTION TRAINING MODULES? YES

Michael W. Davidson and Lillian M. Range

University of Southern Mississippi, Hattiesburg, Mississippi, USA

The C.R.I.S.I.S. Model and Program addresses Caregivers Responding In Situations Involving Suicide.  Caregivers include educators and so I researched the literature and found a study of interest in Death Studies, 23: 61-71, 1999. The question presented in Davidson and Range’s research study was "Are teachers of children and young adolescents responsive to suicide prevention training modules?" The research question and hypotheses seek to deal with an issue, which is to determine whether teachers would be responsive to suicide prevention training. In exploring the merit of the research question, I want to find out if there is justification in the question. Do we need to know the answer and if so, is the conclusion warranted?

I believe that there is great merit to this research question because suicide is the third leading cause of death in 15-19 year olds. Research has found that suicidal youth may be unresponsive to the intervention programs directed at them in the school system. Therefore, teachers need to be the target population in efforts to help these children. Teachers are on the front line in handling student problems, and although they are not mental health professionals, they can be gate keepers in intervention. The researcher’s question exhibits clarity in limiting the suicide ideation to children and young adolescents; moreover, it addresses the use of a suicide prevention training module for teachers and their responsiveness to this approach.

The design of the study included a sample of consenting anonymous practice teachers. Most were Caucasian, American and women. Distribution of the subjects was based on their being practice teachers who were teaching full-time under supervision and were completing their final semester before graduation with a degree in elementary education. They were going to teach children between the ages of 6 and 13. In theory, random assignment produces groups that have equal skill levels. It is possible that these practice teachers had the same skill level. They offered the one hour in-service training module in the middle of the practice-teaching semester, typically the last semester before graduation. This quantitative research intervention included information about suicide and warning signs, community resources, a question and answer period for subjects, and addressed practice-teacher’s anxieties about dealing with what could be a difficult subject.

To measure the subject’s responsiveness to the questionnaire, the researchers included related topics. They consulted other social scientists and reviewed research already conducted. Regarding criterion validity, the researchers used the same questionnaire as Davidson & Range, 1997, so there was a ‘gold standard’ that already existed. This saved time in writing new questions and preserved some compatibility with earlier research.

The variable to be measured was the subject’s responsiveness to a suicide prevention training module.  They gave subjects a pre-test, an author generated, a 4-sentence vignette about a suicidal student ("Pat"). Following the vignette were eight 7-point Likert-type questions about teachers’ responses (from 1 = completely unlikely to 7 = completely likely). The practice teachers then participated in the training module. At the conclusion of the training, a posttest was administered. The teachers were asked to turn the page over and complete the back, which included the same vignette and eight questions as in the pretest. By keeping the tests the same, they were off to a very good start. Data was recorded both before and after the training module and the researchers gave the subjects the same instrument in the posttest.

In addition to the questionnaire, the data monitored included escorting a student to counselor, no-suicide agreements, parents called, seriousness of a suicide ideation, and comfortably in similar situation. Compared with pretraining, at post-training the teachers were more likely to say that they would send or escort Pat to the counselor’s office, use written or verbal no-suicide agreements, call Pat’s parents, believe Pat to be serious rather than simply seeking attention, and comfort in similar situation. The researchers found that increased proactive attitudes after one hour of training implies that teachers would benefit from periodic suicide awareness and prevention training modules.

The researchers fulfilled the four criteria for a good research study justifying the research effort. It was an interest to teachers and educators, it was feasible, it was practical and the topic was narrow enough and sufficiently focused so that the research could be performed in a reasonable period of time. It was relevant because the topic of suicide could be used in the real world and will solve a problem of suicide. It appears that the study was ethical because no one was maltreated in the study. However, the researchers stated that there were limitations in the design. The posttest data were collected immediately after the training module was presented. They also stated that demand characteristics in experimental design might influence the subject’s attitudes. Thus, the subjects simply might have complied because they wanted to cooperate. Though the test was reliable there might be a question how valid the answers are. I base this validity issue on the fact that though they aim the questions at getting at the truth, the authors did not take into consideration those practice teachers would have expectations to grade higher on the post-test that was immediately after.   They might have wanted to be deemed more desirable as a respondent and as a student themselves.

The authors stated that another limitation is that practice-teachers were relatively inexperienced. They said that future studies should include teachers of various levels of experience. When an Inservice addresses brand new teachers, we give them the tools to save lives of their students from the very beginning of their career, and so this study was admirable. However, all teachers should receive this type of training. Practice teachers might have been suffering from depression and suicide ideation or have lost someone they know to suicide. This needs to be addressed as it can influence their state of mind and therefore, their score. They might feel guilty about their role in past intervention, or may be fearful of their own suicide ideation and this may influence the score and is unpredictable.

Another design that I believe would be a better tool in seeing whether training modules work is a two group, pretest-posttest design. One group would be experienced teachers working with children and young adolescents and the other group would be practice teachers who will be teaching the same age group. However, this would produce non-equivalent groups.  The researchers can compare their findings with more experienced teachers and see if the training module had the same effect. Subject variables were controlled by including only one type of subject in the experiment. Therefore, selection was a threat to internal validity as there was only one group in this design and no means of comparison to a different treatment group. The question addressed all teachers and the researchers only studied practice teachers.

Collecting data involves obtaining information objectively reflecting truth as told to the researcher. It was obvious what sorts of attitudinal changes were anticipated, and present teachers may have cooperatively complied with these expectations. It was expected that responsiveness would be associated with higher scores. Validity is the extent to which conclusions can be drawn about the causal effects of one variable on the other. The variable of training modules affects responsiveness scores to intervention. As practice teachers, the subjects do not have the confidence and experience of trained teachers. Were the subjects telling the truth? The subjects might have believed that the researchers were looking for higher responses after the training. The subjects were practice teachers who had not yet graduated. Therefore, without even realizing it, these feelings of wanting to obtain a better score might have influenced their responses. I think the researchers did capture the true relationship between their variables but the design of the study should have included teachers that were experienced and those of different cultures as well. I wonder what the response rate would be in a catholic school where suicide is considered a sin. I also wonder what African-American teachers would report, as there is little suicide ideation in that population.

The survey should have also included a three to a six-month follow-up questionnaire, same as pre and post test to see if they integrate the knowledge from the intervention. The administration of the posttest immediately after the training intervention might have influenced the positive responsiveness. There was little chance of the threat of maturation as the intervention was only one hour and their personal attitudes would not change because they were growing older. There was also no threat of 'History'.  There was no time for 'History' to intervene as the researchers did not suggest that there was a cluster of suicides, or any particular school was dealing with a recent suicide. There was no mention of anything similar to the above going on in the outside world at the same time as the study. There was no instrumental threat as there were no changes in the measuring procedures. Both the pretest and post test were the same. The practice teachers might have talked to one another during the training and their interactions during the intervention might have influenced their posttest scores.. During the module there was a question and answer period. During that time the practice teachers were allowed to ask questions and share experiences. The experiences shared might have influenced the responses.

Approximately 90% of those attending the in-service training completed both questionnaires. Those who declined to participate simply turned in a blank sheet of paper. A threat to internal validity is mortality that is any drop out of subjects at any phase in the study. Regardless, a 70% response rate is deemed tolerable. Though it is deemed tolerable, I question why there was attrition in the first place with the understanding that teachers want to help their students do well, to learn and succeed. The researchers did not attempt to reduce non response rate. The researchers could have implemented a qualitative method to figure out why there was attrition by asking the subjects questions directly. They could then addressed those issues in the design of any future studies. However, with 75 subjects, 90% retention (10% attrition) is minor.

The one group pretest/posttest design did not show thoroughness because it neglected to distribute the subjects equally according to race and gender. The sample is an effort to represent the larger population. The group of people the study was interested in knowing something about are the new teachers and they represent the sample subset of population. The sample must represent population in ages, ethnicity and gender, similar to the target population. The groups should be heterogeneous and diverse. The fact that the sample was white American women is not a clear representation of the population. For external validity the researchers had to make explicit the population and setting to which generalization is sought. I think that by only having practice teachers it was not general enough.

I would like to address face validity, the extent to which their questionnaire was measuring what it attempted to measure, it did include appropriate questions. Construct validity is the degree to which both the independent and the dependent variables accurately measure the constructs of interest. The independent variable is the training module. The dependent variable is the responsiveness on the questionnaire before and after the training. The hypothesis shows a relationship between construct A (training modules) and construct B (questionnaire). The overall hypotheses is when teachers participate in training modules, their responses are confined to knowledge gained as a product of that intervention and it is an example of construct validity.

The hypothesis is the prediction of what might happen between two variables, the training module and the questionnaire which indicated the response to a suicidal student. It made a statement that told me something about the relationship between these two variables. Davidson and Range took the essence of a question and made a prediction. That prediction is teachers of children and young adolescents are responsive to suicide prevention training modules and they supported the hypothesis. The hypothesis committed the researchers to a particular prediction. They asserted the prediction between the two variables and confirmed it. They had to document that when a teacher took part in training module, the more confident the teacher would be in helping the student, and the chance of their student completing suicide is less likely because of teacher intervention. The selection of the topic is feasible, of interest and relevant.

The internal validity is the function of the design. The relationship between high scores on the test does not mean that the intervention caused the increase in scores. Testing might have been contaminated if the teachers talked with one another before taking the posttest as this is a concern when you have two groups. The conclusion the researchers made is that teachers became even more definite about their proactive attitudes after an hour of training about suicide intervention. However, there was a possibility of the Hawthorne effect as there was only one group. There could be an external factor as they were being given attention, being watched and their responses were tested.

In summarizing the research reviewed here is to focus on the scores. Compared with pretraining, after training these teachers reported that they were more likely to send Pat to the counselor’s office (M = 6.00 vs. 6.53) p <.001; escort Pat to counselor’s office (M = 5.62 vs. 6.27 ); use write no-suicide agreement (M = 4.39 vs. 6.16),p <.001 ; use a verbal no-suicide agreement (M= 4.72 vs. 5.57), p < .001; believe that Pat would attempt suicide (M=4.91 vs. 5.51),p < .01; call Pat’s parents (M= 5.76 vs. 6.20), p <.05; be comfortable in handling a similar situation (M=3.67 vs. 4.28, p < .01; and view Pat as not simply seeking attention (M = 4.23 vs. 4.84), p <.01. The mean is the average of all scores and the results support the author’s conclusions. The p levels indicate that the chance of the scores being an accident are very low.

The teachers would take specific action to the threat of suicide and take the threat seriously. With respect to taking concrete action, these teachers were positive before the training and were even more positive afterwards. With respect to attitudes, as in taking Pat seriously, teachers were neutral to slightly positive before to moderately to highly positive afterwards. Results indicated that the answers were significantly higher at posttest than at pretest, indicating an increased willingness to view the suicide threat seriously and take specific actions to prevent it.

Posttest improvement was true for each of the eight items. There was a drastic change in these teachers’ attitudes toward no-suicide agreements, changed from barely positive to certain that they would use them. Perhaps the practice teachers were not even aware that these no-suicide agreements exist. The scores revealed that they would actually take them to counselor, and call parents. The intervention made them more aware of the risk involved when a student is talking about suicide. Before the intervention they were almost neutral in responding to a suicidal student. They did not think the threat was a serious one. By becoming aware of the suicide statistics and being educated about prevention, they learned and believed they would implement the training.

External validity is the function of sampling and ideally, we aim for high generalizability. To what extent can we generalize from the research sample of practice teachers to the populations of experienced teachers already in the school system? The researchers compromised the external validity of their work when they defined their population narrowly. Nevertheless, they came close to achieving good representation. How they take their findings and apply them to external validity of a study is a function of what went on in the sampling procedures used in that study. The results of the study were statistically significant. I believe that the researchers can administer the same tests to teachers throughout the United States and be confident that the results of the research would be the same for both practice and experienced teachers. Though there is a theoretical debate as to whether this type of intervention works or not, the study can translate statistical significance into theoretical significance, proving that teachers do not want their students to complete suicide and if possible will take an active role in prevention. The results show that teachers need to be educated about suicide ideation and how they can effectively prevent suicide. I believe that suicide prevention training modules positively impact on teachers of children and young adolescents and this study was worthwhile.

I look forward to your comments regarding this critique. Send me an email at  griefwork2@aol.com

        Copyright (c) 1999 Griefworkcenter, Inc.         All rights reserved  

        Please see disclaimer          You are visitor number