Child Abuse Prevention Methods

This paper will outline the data collection and analysis methods for child abuse victims in alcoholic families. The hypothesis is that in nuclear families affected by alcoholism, secondary level child abuse prevention (intervening in families affected by alcoholism, but where families have not been referred for child abuse) will lead to fewer subsequent family child abuse referrals to the criminal justice and social service systems than will tertiary level child abuse prevention (intervening in alcoholic families where the family has received a referral for child abuse).

Sample

The population to be studied is families with children (persons under the age of 18) where alcoholism is present in one or both parents. In order to maximize representativeness, every parent in Winnebago County AODA treatment groups will compose the sampling frame. First, the AODA treatment groups will be approached, and possible participants will be informed of the study and asked for their consent to be involved. Those that choose to participate will then subjected to the sampling procedure. The sampling method that will be used is cluster sampling. The sampling frame will be divided into two groups: one will be composed of alcoholics with children who have not yet suffered abuse (the secondary level), and the other group will be alcoholics with children who have been abused (the tertiary level). The sampling unit will be the individual in treatment. Persons will be appropriately placed in each group according to whether or not they have committed acts of child abuse. Names from each group will be drawn randomly from a hat or placed into a computer program that mimics that process.

The best way to access a segment of this population seems to be through women's and men's AODA treatment groups. In most cases, women have custody of the children, but men also occasionally get custody, and it may be useful to include men to enhance the representativeness of the study. A second benefit of using persons in treatment for AODA issues is that this population would have a high chance of being referred for child abuse at some point in the past or possibly in the future, which makes this population a good fit for secondary and tertiary levels of intervention.

In regard to confidentiality and voluntary and informed consent, the treatment agencies will be informed of the planned procedures and that the study plans only to report statistics in terms of subsequent child abuse cases, without linking any particular information to any particular person. Potential participants will be informed that participation in the study will be completely voluntary Second, they would be informed that, should they choose to participate, the information they reveal would be for research purposes only, and the study would not link any information to a particular individual. This fact will be underscored to participants multiple times so that they do not feel any pressure to respond with what they see as socially desirable responses. They would also be told the information would be kept private from any legal authorities potentially seeking action against them. The main point emphasized would be that the study was designed to research the effectiveness of intervention on improving parent-child relations.

This study itself is designed in such a way that everyone is already participating in an AODA treatment program, so there are no issues of whether or not participants have access to intervention. The major source of potential harm for the study lies within how well researchers ask questions of the participants. Some lacking experience in this area may use a harsh or judgmental tone, and this could lead to a participant feeling more stressed and ashamed, which reduces their growth in the process, or perhaps even could cause them to drop out of the study entirely. This is an area that needs to be closely monitored.

Instrument

Data will be collected via in-home (or anywhere else at their convenience) personal interviews with the participants. The unit of analysis is a referral to social services or the criminal justice system for child abuse. Here is an example of what an interviewer's question sheet may look like:

1. Begin with polite conversation about the weather/current events.

2. Thank participant for their participation in the study, and remind them of the potential benefits they and others may receive from the research in the study.

3. Remind participant about the confidentiality of any revealed information.

4. Remind participant of the financial benefits of participation in the study.

5. How are you doing?

6. How's your job going?

7. How's your wife/significant other doing?

8. How are your children doing?

9. What are they learning in school?

10. What progress have you made during the course of AODA treatment?

11. What struggles have you had during the course of AODA treatment? You should go easy on yourself and comment yourself on the progress you have made so far.

12. Have you had any referrals to area social services or county CPS for suspected child abuse or neglect (remind participant again about confidentiality of information and financial rewards for study participation)?

13. Thank the participant for his/her participation. Congratulate them on any positive progress made. Remind them again of confidentiality of information and financial reward for participation.

The directions and entire purpose of this instrument are deceptively simple. The only goal of this interview is to establish whether or not the person has had subsequent referrals to the county CPS or any other social service system for child abuse or neglect. The entire rest of the interview is only a setup to get the participant relaxed and comfortable to the point where they would feel at ease sharing very personal information with the interviewer. Another good strategy may be to have the same interviewer work with the same participants, as this can maximize participant comfort with the interviewer, and therefore the participant's likelihood of revealing desired information is also increased. The bottom line is to minimize participant resistance and maximize continued participation so as to get the most accurate data possible. The level of measurement to be used is nominal. Either participants have had referrals and subsequent referrals for child abuse or they have not. In some cases, it may be helpful to talk to the participant's social worker, as this can help determine how accurate the information is that the participants is relaying. The interview itself does not need to be any longer than forty-five minutes, probably closer to thirty on average. The structure is pretty important in that it constantly emphasizes confidentiality and financial rewards, which are two important factors that determine the accuracy of information revealed and the continued participation of the participant.

In regard to validity, this study is set up well to measure what it is in fact designed to measure. It is very linear in design, and really is only interested in one piece of data: the existence, or not, of documented child abuse. One factor that may affect validity is participation. One study involving Project Safecare had just fifty-three of 266 families complete the study, which could last up to six months. After six months, the families were awarded a gift certificate to a grocery store, the amount of which was not noted. Obviously, the reward was not provided on a regular enough basis or was not of a large enough amount for participants to continue with the study. This study is intended to be longitudinal, and if the techniques in this study still lead to high levels of attrition, or if statistics in this study are very inconsistent with child abuse rates found in other studies (which could be due to feelings caused by poor interviewing skills or participants giving misleading information), then this study may not be valid. A good check will be to see what percentage of participants finish the full course of research, and what percentage drop out. The approximately 25% completion rate in Project Safecare seems to be standing on pretty shaky ground from which to draw empirical generalizations. Another threat to validity is that this study is measuring formal referrals for child abuse, which tend to be the more extreme cases of child abuse. This may mean that the final data are not representative of all child abuse cases, but the more extreme ones. This method was chosen because it is easier to measure and more cost-effective than investigating participants using a standard measurement tool and trying to establish whether or not child abuse consistently occurs in a particular household. As it stands, this study is set on a firm foundation for being valid in relation to the more extreme cases of child abuse. In regard to reliability, similar issues appear. It seems that the primary issue is again participant attrition, and this study attempts to minimize that by offering consistent financial rewards which are hopefully significant enough to encourage participation. Reliability will be checked by comparing the data to existing data on child abuse intervention success (Gershater-Molko, Lutzker, & Wesch, 2003).

Design

The design to be used is the Multiple Time Series with a Nonequivalent Control Group Design. There will not be a control group receiving no intervention; each group will be participating in AODA treatment. The participants in the secondary and tertiary groups will be compared on a longitudinal basis to see if there are any differences between the secondary and tertiary levels of intervention. Observations for the secondary group will be made on an annual basis until the child turns eighteen, and likewise, observations for the tertiary group will also be made on an annual basis until the child turns eighteen.

In regard to the internal validity threat of history, many things could happen that could confound the results. The parent could have a relapse that might land them in jail, making them unavailable or highly resistant to continued study. There could be other events such as natural disasters, the finding of a new job, or the family moving to a new area that could possibly lead to participant attrition and confounded results. In regard to the internal validity threat of maturation and passage of time, a person simply having a child abuse referral may be enough to get them to realize they need to make some changes in their life, which could blur the lines as to whether or not intervention made the real difference in a person's life. People might also simply mature and move forward from the attitudes and lifestyles they had which enabled them to commit child abuse, and this may confound the results of this longitudinal study. However, this study is designed to be longitudinal and this will help to minimize this threat, as it can show the fluctuation and maturation of program participants. In regard to the internal validity threat of testing, participants may figure out that what the study is really interested in is subsequent child abuse referrals. However, the interview is set up in such a way that it is more of a friendly conversation to see how things are going, rather than something designed to investigate child abuse; the idea is that this minimizes the participant feeling threatened or interrogated by the study, and therefore they will not feel the need to respond to questions in certain ways in order to improve their test performance. It may be necessary at some point in time to switch the questions up somewhat so that participants are unable to figure out the "best" way to respond to the questions. In regard to the internal validity threat of instrumentation, the study uses the same measurement of the dependent variable, and that is subsequent referrals, or not, for child abuse. The same thing is going to be measured each time, so this threat will be minimized. In regard to the internal validity threat of statistical regression, this study has some vulnerability, especially since it is longitudinal. One thing that guards against this issue is the usage of women's and men's AODA treatment groups, which, overall, would tend to have more representative cases of child abuse than would CPS cases. In regard to selection bias, this study is designed well. Both women and men in AODA treatment groups are being used, then are placed in secondary and tertiary comparison groups, and finally are selected randomly for study. The groups are comparable, as they both involve the same measure of child abuse. In regard to the internal validity issue of experimental mortality, this study is highly vulnerable, as noted by the previous statistic of people dropping out of Project Safecare at a rate around 75%. The Project Safecare reward, given after six months of participation in an intervention, was probably not large enough to make people feel participation to be necessary. In this study, rewards will be of lower sums, but placed on a more frequent basis. Paying participants $20-25 in cash or with a gift certificate for an interview that will not last longer than an hour seems to be a pretty good reward that most people would find appealing, especially since many persons struggling with AODA issues also struggle with poverty. The other method that could be implemented would be allowing a selection of different gift certificates to different stores; the more people are allowed to participate in decisions affecting them, the more interested they become in participating in an activity. People do not enjoy simply being told what to do, given what they are supposed to, and then hurried out the door. Hopefully, this strategy will effectively address this issue. In regard to the internal validity issue of ambiguity about the direction of causal inference, this study faces a challenge as well. For example, a person in the tertiary group who has committed child abuse may not commit child abuse in the future not because of participation in the AODA treatment program, but because of abstinence from chemical dependency. One way to control for this may be to develop questions for program participants that ask them whether they feel the improvements in their lives are due more to the AODA intervention or personal abstinence from chemical dependency, or perhaps using some standardized scale that can measure this effect. Finally, in regard to the internal validity issue of diffusion or imitation of treatment, probably almost every study faces this challenge. This study faces that challenge in that participants in secondary or tertiary groups may attend the same AODA treatment groups, and talk about the study will occur to some extent. However, the interview this study uses is designed to focus on the participants and off of what the interviewers want to find, so hopefully that will help to reduce talk among participants. Interviewers may also talk amongst one another and influence study results; however, the interviews can be done while working alone, and the interviewers will also be consistently reminded how important it is that they remain objective so as not to confound the study results.

This study seems to be set up well to have a fair amount of external validity. The major challenge, as has been noted several times, is attrition, and hopefully this study has been set up to minimize that effect. The study is very linear in that it really is only interested in one piece of information, child abuse referrals, and that linear design greatly reduces complexity, which also enhances the level of external validity. The results could definitely be generalized to persons who have received professional treatment for AODA issues. However, the results would not be generalizable to child abuse in general, which can happen in homes without AODA issues and can happen for a variety of different reasons. The results may also only be generalizable to Winnebago County and other regions with a similar standard of living. Milwaukee, for example, may experience higher child abuse rates because many areas of it have much lower standards of living and include more violence and poverty than is seen anywhere in Winnebago County.

One future area of study would be to look at much more high-risk areas such as Milwaukee. Also, it may be wise to study child abuse from other aspects where AODA issues are not present-such as in areas of rampant poverty, in homes where anger runs out of control, or in homes with other significant mental illness present. Studying these areas would give us a much clearer and more representative understanding of child abuse, and the many factors that lead to it. This study, however, will make a valuable contribution to child abuse as it relates to AODA issues.

Bibliography

Gershater-Molko, R.M., Lutzker, J.R., & Wesch, D. Project SafeCare: Improving Health,

Safety, and Parenting Skills in Families Reported for, and At-Risk for Child

Maltreatment [Electronic version]. Journal of Family Violence, 19(6), 377-386.