Journal of Social Work Values & Ethics

JSWVE is best viewed in Firefox.
Spread Firefox Button

Managing Editor: Association of Social Work Boards

The contents of this website and any publications, advertisements, and other materials contained herein do not necessarily reflect the views of the ASWB and are not to be considered an endorsement or indicator of support for any such publications, advertisements and other materials.  ASWB is a not-for-profit, 501(c)(3) organization whose membership and mission statement can be found at     

ISSN: 1553-6947

All pages on this site © White Hat Communications. All rights reserved.



Choose Language

Who's Online

We have 2 guests online
It It Ethical? 101 Scenarios in Everyday Social Work Practice: A Discussion Workbook

Notice: As of January 1, 2013, the Journal of Social Work Values and Ethics is published by the Association of Social Work Boards at This site will remain active as an archive of the journal's editions from 2004-2012. Thank you!

Home arrow SPRING 2005: VOL. 2, #1 arrow Felony Convictions and Program Admissions: Theoretical Perspectives to Guide Decision-Making
Felony Convictions and Program Admissions: Theoretical Perspectives to Guide Decision-Making Print E-mail

Felony Convictions and Program Admissions: Theoretical Perspectives to Guide Decision-Making


Gail M. Leedy, MSW, Ph.D., and James E. Smith, MSW, Ph.D.

University of Wyoming  


Social work education programs face an ethical dilemma when determining whether to admit an applicant who has been convicted of a felony.  Decisions must be made which protect future clients while also providing educational opportunities to qualified students. A hybrid decision-making model, integrating  statistical modeling and intuitive processes, is presented.

Key Words:  Social Work Education; Gate-keeping; Felony Convictions; Intuition; Emotional Awareness; Ethical Decision-Making


Social work programs in the United States accredited through the Council on Social Work Education (CSWE) are required to have policies regarding admission procedures for both undergraduate and graduate levels.  However, the Educational Policy and Accreditation Standards ratified by CSWE in 2002 (2003)  provide little guidance regarding admissions policies or decisions.  Earlier accreditation standards (CSWE, 1994) stated, "Criteria and processes of admission should be designed and implemented to accept from the group of applicants those who, in accordance with the program's educational goals, are best qualified to become professional social workers at a [beginning level BSW program or advanced level MSW program] of practice " (Evaluative Standard 5.0).   There has been a strong tradition for social work programs to consider the admissions process as part of the gate-keeping role for the profession, and many authors have attempted to better define those qualities which mark an applicant as suitable, thereby promoting the gate-keeping role (Born and Carroll, 1988; Dunlap, Henley, & Fraser, 1998; GlenMaye & Oakes, 2002; Miller & Koerin, 1998; Pelech, Stalker, Regehr,  & Jacobs, 1995).

This gate-keeping role presents educational programs with a potential ethical dilemma.  Programs are to admit those applicants who are "best qualified to become professional social workers” (CSWE, 1994).  Admissions committees are compelled to make predictions of the applicant’s future competence, integrity, and commitment to the profession’s core values.  

One of the particularly difficult decisions in gate-keeping admissions policies is deciding whether to admit an other-wise qualified applicant with a previous criminal conviction.  Such decisions require that we deal with the legal issues as well as with the ethical dilemmas (Cole & Lewis, 1991; Gelman & Wardell, 1988; Gibbs, 1994). While the legal questions are usually handled by the general counsel for the university, the ethical issues must be resolved at the level of the program itself.  Programs first turn to the National Association of Social Workers’ Code of Ethics (NASW) for an overview of the value base of the profession (NASW, 1996).  

The most salient value conflict centers around the protection of the client and the rights of the individual applicant.  As stated in Section 1.01 of the Code (NASW, 1996), the social worker's primary responsibility is to the client.  In Section 2.09, there is the explicit admonishment to "discourage, prevent, expose, and correct the unethical conduct of colleagues." Further, in Sections 4.04 and 4.05, standards related to dishonesty and impairment of professionals, are presented. These values may be viewed as conflicting with the rights of applicants and the need to respect their worth, if indeed the applicant is rehabilitated.  As Younes (1998) points out, this dilemma is compounded even further when the applicants themselves are from an at-risk  population.  

Magen and Emerman (2000) and Scott and Zeiger (2000) have presented us with an excellent ‘Point and Counterpoint’ debate of possible solutions for this value-based dilemma: a blanket policy to reject all applicants with a felony record versus a case by case review of the individual applicant. Magen and Emerman present two central points for not accepting applicants with felony records.  Their first argument is that felony convictions are a form of social sanction.  With a social sanction comes a loss of rights, including the loss of opportunities to engage in certain professions or educational opportunities including social work. 

Secondly, Magen and Emerman (2000), in agreement with Born and Carroll (1988), argue that students are not the clients of social work programs.  Rather, the clients are the recipients of service provided by the program's graduates.  Thus in accordance with the NASW Code of Ethics (1996) these authors contend that the potential risk to the clients should not outweigh the costs of denying an educational opportunity to an applicant, even one who is truly rehabilitated.

In their counter argument, Scott and Zeiger (2000) contend that the ultimate charge is to determine if an otherwise qualified applicant poses a risk to clients, to colleagues, to the profession, or to society.  This risk to clients and the profession must be weighed against the cost of failing to accept such an applicant, including failing to honor the worth and dignity of the individual applicant.  Admitted applicants should be those who are active contributors to the enhancement of clients’ well-being.  Therefore, if an applicant has the potential to become an excellent social worker, denying them their opportunity based on their past behavior is a disservice not only to the individual but also to clients and the profession.  Rather than taking a blanket approach to rejecting all applicants with a criminal background, Scott and Zeiger suggest reviewing each applicant individually for admission, including an interview assessment. They delineate several characteristics of the applicant which may have predictive value, including the type of offence committed, time elapsed since the criminal act, as well as other indicators.     

Decision-Making Perspectives

The present paper argues that programs that choose to review applicants with a criminal background must develop clearly explicated policies. However, even with clear policies, it will ultimately be necessary to make a decision in spite of uncertainty.  Therefore it is crucial that the decision-making process itself be based on a clearly defined model derived from sound decision-making principles.  We assert that this process should be based on a theoretical framework of risk management in decision-making.  Two different decision-making models are presented here, along with our developed model for a decision-making support system.  These models and the support system are applicable to admissions decisions and are consistent with the social work value base.

Ethical Decision-making in Social Work

The NASW Code of Ethics (1999) does not provide specific rules for practice, but rather provides a frame of reference for making decisions, with the understanding that the context in which practice occurs must be factored into the decision-making process.  Based on the Code of Ethics, various models for resolving ethical dilemmas in social work practice have been proposed.  Loewenberg, Dolgoff and Harrington (2000) present a decision-making model which combines an ethics assessment screen and ethical rules screen, based on a rank-ordering of ethical principles.  Mattison’s (2000) model takes this a step further by emphasizing the importance of the preferences and value system of the social worker.  She argues that increasing the self awareness of the decision-maker is necessary for understanding the patterning involved in making decisions when social work principles are in conflict.

Determining Thresholds: Signal Detection Theory

Concepts developed from Signal Detection Theory can facilitate our understanding of probability-based decision making processes.  This theory was originally designed to assist operators in detecting and reporting a "signal" (such as a radar signal) which was presented at or near the threshold for detection.  Signals in this range result in uncertainty and rules for reporting detection were needed (Swets, 1986).   Four possible outcomes of making a yes/no decision exist under conditions of uncertainty.  A ‘Hit’ is when 1) a signal (dangerous condition) exists and 2) the signal is reported.  A ‘Miss’ is when 1) a signal (dangerous condition) exists but 2) it is not detected/reported.  A ‘False Positive’ is  when 1) no signal exists but 2) the operator falsely believes it was detected and therefore reports it. "Peace and Quiet" is when 1) no signal exists and 2) none is reported.     

Mossman (1994) and Swets (1992; 2000) have applied the signal detection model to diagnostic decision making.  In cases of diagnostic uncertainty, two key concepts emerge: decision threshold and decision accuracy.  According to these authors, the concept of balancing between false positives and misses is a trade off between sensitivity (detecting real signals) and specificity (not sounding false alarms).  The decision of how strict the threshold should be depends upon this balance.  When it is important that all signals are reported due to the severe costs of missing a signal, a lenient threshold will be used.  With a lenient threshold, a signal will be reported even when there is a high degree of uncertainty of its existence.  In this way the number of hits is maximized and the sensitivity is high.  However, because signals are reported even though there is considerable doubt regarding their existence, there is a corresponding high rate of false positives.  Thus, sensitivity is high for a lenient threshold and specificity is low.

In contrast, when the outcome of a false positive is severe in relation to the outcome of a miss, a strict threshold might be recommended.   In this way, a signal will be reported only if there is a high degree of certainty that the signal exists.  With a strict threshold, the false positive rate is reduced but there will be a corresponding increase in misses. In other words, the specificity is high but the sensitivity is low.

Another important consideration in determining the threshold for detection is the base rate of occurrence (Swets, 1992; 2000. When a signal, or dangerous condition, is highly unlikely to occur, it is reasonable to set a strict threshold for reporting. All of these factors, the cost of false positives and misses, the benefits of hits, and the base rates of occurrence, interact in determining where the threshold should be set.

Signal Detection Theory in conjunction with statistical information on recidivism (described below) can be applied to the present ethical dilemma of whether a social work program should accept an otherwise qualified applicant who has a criminal record.  The four contingencies of the Signal Detection Theory applied to this decision are presented in Figure 1.   

Figure 1.  Signal Detection contingency table regarding applicants with a criminal record (adapted from Green & Swets, 1966).

In applying the signal detection model to making decisions regarding applicants with a criminal record, a "Hit" refers to correctly recognizing that the applicant will re-offend and thus endanger the client and/or the profession.  Having recognized this, the committee rejects the candidate from the program.  A "Miss" refers to failing to recognize that the applicant will indeed commit further offenses.  Because this danger is not detected, the committee incorrectly accepts the applicant into the program.  A "False Alarm" is making the (incorrect) determination that the applicant will re-offend, when in reality they do not pose a danger.  In this case, the committee rejects a qualified applicant.  Finally, "Peace and Quiet" is accepting a qualified applicant who poses no danger.

For this example, our dilemma is further complicated because we must base our decision on our prediction of the applicant's future behavior.  Unlike radar signals, there is no absolute knowledge of whether a signal exists.  Thus, our dilemma is similar to the diagnostic uncertainty model discussed above (Mossman 1994; Swets; 1992, 2000)For assistance in determining our course of action, we can return to the concepts of sensitivity and specificity to determine our threshold for reporting a signal,  that is, an applicant who presents a danger to clients, colleagues or the profession. This threshold for rejecting a candidate will depend on the admissions committee’s assessment of the impact of incorrect decisions.  One of the most important considerations in determining the threshold is the type of offense which was committed.  For instance, if an applicant has a record of repeated abuse against vulnerable people, the threshold may be very lenient.  In other words, if there is any degree of evidence that the applicant will re-offend, this danger will be reported.   In contrast if an applicant has a record of a single offense of a non-violent crime such as making a false statement to secure a loan, the threshold may be more stringent.  In this case, the impact of committing another similar offense, while certainly in conflict with the core value of integrity, would generally not be seen as especially dangerous to clients. Therefore, the applicant would be rejected only if there was a high degree of certainty that a signal (evidence of re-offending) exists. 

The impact of false positives must also be considered.  In our situation, if there is extreme competition for a limited number of student slots and there is an over-abundance of social workers, the cost of rejecting a qualified applicant would not be seen as a particular problem.  However, if there is a severe shortage of social workers and applicants to the program, then rejecting a qualified applicant would be more costly. Furthermore, the characteristics of the applicant themselves must be determined.  If the applicant is especially qualified and possesses characteristics which would be an asset to the profession, the cost of a false positive increases.  

Decision Accuracy

Decision accuracy is a measure of how closely the decision which is made matches the actual outcome.  Swets (2000) explains that accuracy is highly dependent on the quality of the information used for decision making.  Utilizing factors with good predictive validity and having sufficient information to make a decision are two prerequisites for accuracy. Several methods for increasing accuracy, relevant to the dilemma presented here, will be discussed below.

Statistical Decision-Making Models

The question at hand requires that we make a decision to accept or reject applicants with a criminal history based on predicting whether they will commit further crimes.  To aid in making this decision, statistical tables of recidivism rates should be consulted.  According to the U. S. Department of Justice (Department of Justice, 2002), 67.5% of all persons released from prisons in 1994 in the 15 states studied were re-arrested within three years.   The re-arrest rates ranged from 40.7% for kidnapping to 78.8% for motor vehicle theft.  These rates also varied by personal demographics.  For instance, women had a 57.6% re-arrest rate, while the rate for men was 68.4%.   The younger the person at the time of release, the higher the re-arrest rate.  Minority re-arrest rates were higher than those for White people.  

However, the raw data lend only marginal predictive power when looking at individuals.  Therefore, numerous efforts have been made to develop a statistical model which would be useful in determining the risk of recidivism based on individual characteristics.  Silver, Smith and Banks (2000) review and compare the most common models.  Out of almost 100 possible risk factors associated with recidivism, 14 have been found to account for the majority of the variance.  These include age at sentencing, employment status, number of previous arrests, and substance dependence.  Using either linear or multivariate logistic discriminant analysis equations, very good predictive accuracy may be attained based on these variables. Indeed, according to Silver, et al., (2000), these models are more accurate than expert clinical judgment.  Likewise, Mossman (1994), wrote that “a nonclinician furnished with knowledge of past behavior may outperform a mental health professional relying solely on information garnered from a clinical interview” (p. 790).

Similar statistical models have been applied directly to social work fields of practice, including child protective service agencies when determining which reports of child abuse and neglect to investigate (Johnson, Brown and Wells, 2002).  Buttell and Carney (2002) report on a statistical model used to predict whether men who are court-ordered into treatment for battering will complete the program.  In both instances, statistical models provided significant information regarding the interactions of variables.  

Statistical models have many advantages, including reasonably accurate prediction rates, reliance on proven objective factors, the ability to combine an extraordinary amount of information, and the lack of individual bias entering into the equation.  However, there may also be significant drawbacks to their use.  Silver and Miller (2002) caution us that statistical models are not constrained by ethics.  For instance, since race may be an important predictor variable for recidivism, these models automatically factor it in to the equation without regard to the possible role of discrimination or oppression.  Continuing to rely on race within the statistical models can result in further marginalization of various groups.  Using this statistical model then, rather than working to alleviate the conditions which place the person at higher risk, actually perpetuate discrimination.  

Emotional Awareness and Intuition Models

In contrast to the statistical processes described above, when using an intuitive model the decision maker does not attempt to be detached and dispassionate.  Intuitive decision-making models recognize that we must rely on our judgment, our previous experiences and our “gut-feelings.” 

According to Barbalet (2001), decision-making is at the core of intellect and rationality.  The process of coming to decisions necessitates using all available information: direct, indirect, experiential, verbal and nonverbal, cognitive and emotional.  This is especially true when making judgments about others and prospects for the future.  The informational role for emotional content and context in decision-making can assist people in interpreting what they see and hear (Clore, 1994; Barbalet, 2001), and may act as the beginning point for making decisions (Bandura, 1986; Evans, 2001; Forgas & Vargas, 2000).  

Consistent with this, Shweder (1994) suggests that emotions activate a personal “schema” that has the meaning and shape of an emotional story based on the decision maker's individual experiences, perceptions, and meanings, both direct and indirect. These cognitive attributes include the socially constructed qualities of what is right, wrong, good, bad, normal, and abnormal; what is liked, disliked, fair, unfair, and just or unjust. Goleman (1995), in his work on emotional intelligence, suggests, “The emotional mind is far quicker than the rational mind, springing into action without pausing, even a moment, to consider what it is doing. Its quickness precludes the deliberate, analytic reflection that is the hallmark of the thinking mind... Actions that spring from the emotional mind carry a particularly strong sense of certainty, a by-product of streamlined, simplified way of looking at things that can be absolutely bewildering to the rational mind” (p. 28).  Further, Bandura (1986) cautions that faulty decisions and actions can arise from our failure to consider pertinent affective information, misperceptions of relevant affective information, or from deficient cognitive processing of that information.  

Our emotions may play a very powerful role when deciding whether to accept an applicant with a criminal record into a social work program.  We may expect that our emotional response to the type of crime committed will outweigh all other variables, including race, gender, age, qualifications, and the discriminatory practices of the criminal justice system itself.  Understanding these emotional responses can assist us in proceeding with the decision-making process to insure fairness and a sense of social justice to the individual involved.

Intuition, a concept related to emotional intelligence, has been examined as a factor in the decision-making processes (Bayard, 2001; Haidt, 2001; Khatri & Ng, 2000; Luoma, 1998).  As defined in these studies, however, intuition is more than just an emotional response or a "gut feeling."  Lieberman (2000) stressed the role of implicit learning as the underlying mechanism for intuition, and Khatri and Ng (2000) contend that intuition is subconscious, complex, quick and central to all decisions.  Rather than being emotionally based and biased, they argue that intuition is based on an in-depth but unconscious understanding of complex situations.  

Based on these assumptions, Khatri and Ng (2000) have developed an Intuitive Synthesis approach which accentuates the importance of examining the totality of a situation and synthesizing information into an integrated picture or story.  Their model integrates judgment, experience and unconscious knowing with performance and environmental data to model decision-making in business settings. They conclude that intuitive synthesis is widely used, and is especially relevant when making strategic or non-routine decisions, and when there is a degree of uncertainty involved.  

Examining intuition from a social cognitive neuroscience approach, as Lieberman (2000) has done, can provide significant insight into the process.  Although there are many definitions of intuition available through the literature, Lieberman defines it as “the subjective experience of a mostly nonconscious process that is fast, a-logical, and inaccessible to consciousness that, dependent on exposure to the domain or problem space, is capable of accurately extracting probabilistic contingencies" (p. 111).   Based on neurobiological studies, Lieberman argues that the utility of intuitive responding is related to the situation.  In those instances for which a logical decision-making process can be used, intuition actually will result in errors and biases due to an over-dependence on personally salient information, rather than an overall assessment of the cognitive criteria.  In contrast, Lieberman states that situations which involve implicit rather than explicit learning are especially amenable to interpretation by intuition.  For instance, decoding another’s emotional state from their non-verbal behavior relies heavily on intuitive processes.  Also, encoding one’s own emotional states is intuitive. In these cases, intuition is superior to conscious, cognitive based processing.

Hall (2002) reviews the substantial literature on the use of intuition in making medical decisions.  She emphasizes that uncertainty is always a part of decision making and that this uncertainty can result in anxiety, and often negatively impacts the decision-making process.  According to Hall, several types of biases or errors can result.   Representativeness refers to using non-predictive information when comparing the current case to known cases.  While the use of such data can improve the decision-maker’s confidence in their decision, it may discourage the decision-maker from searching for information which is relevant and predictive.  Another bias presented by Hall is based on the availability of cases to use as comparisons to the index case.  When only a few comparison cases are available, these will receive greater weight than they should.  Also, the most recent case will typically be weighed more substantially than those cases further removed in time, prejudicing the decision process.  Intuitively-based errors in decision making also occur due to the ease with which a negative outcome can be imagined.  If the decision-maker has seen negative outcomes before, and especially if these outcomes were dramatic, then the decision maker may be unduly biased against this decision alternative.  A final bias worth mentioning here, described by Hall, is that people generally are optimistic and over-emphasize their own capabilities.  Thus they may be prone to take a riskier alternative believing that they can prevent the negative outcomes.  

In spite of these potential drawbacks, intuition and emotions play a powerful role in decision making.  For the present dilemma posed by applicants to social work programs with criminal backgrounds, the intuitive synthesis model would appear to be an essential component of the decision making process.  Knowing that intuition is especially useful in decoding emotional and personality information from non-verbal cues, it would seem that the use of face to face interviews with such applicants would be desirable.  Furthermore, the benefits of intuition can be maximized as the admissions committee members develop an understanding of the potential biases as well as the most appropriate venues for its use.       

Hybrid Models

Various hybrid models which incorporate both statistical procedures and intuitive reasoning have been proposed.  Benbenishty and Treistman (1998) recommend the use of Decision Support Systems to aid military mental health officers who must decide if a soldier should be discharged due to psychiatric illness.  Their decision support system combines statistical analyses with models derived from a study of expert clinical decision making processes.  These authors advocate for consistent use of such support systems in order to reduce idiosyncratic responses in decision making.  

Another hybrid model, characterized as explanation-based, has been applied to jury decisions (Hastie & Pennington, 2000).  This model focuses on reasoning about the evidence or facts and utilizes a narrative approach. According to these authors, people naturally develop a story or narrative based on the evidence presented in the case.  The goal of this narrative development is to construct a story which provides the best fitting explanation to account for both the evidence presented in the case and the author’s personal knowledge and world knowledge base. Using their proposed model, this narrative is then judged as to its validity, completeness, exclusion of other potential stories, and integration of conflicting information.  Once this story has been critically assessed, it can then be matched to the possible decisions to be made and the potential outcomes of such decisions.  The degree of certainty or confidence that the decision-maker has in their constructed narrative and the goodness of fit with the potential outcomes determines their final decision.

To apply this model to the current decision, that is whether to accept or reject an otherwise qualified applicant because of a criminal record, we would presumably hear or read all of the information regarding the applicant and their background.  To complete our narrative, we would integrate this information with our knowledge of human behavior and our own “practice” knowledge and personal experiences.   We would compare this narrative to our possible decisions of acceptance or rejection of the applicant. Finally, we would make a judgment based on the totality of this information.

Proposed Decision- making Support System

We are proposing a decision-making support system for ruling on applicants who have criminal backgrounds.  This decision support system acknowledges the importance of the applicant’s rights as well as the priority of protecting the social work program, the social work profession and the eventual clients that the applicant might serve.  The primary purpose of this model is not to provide a decision tree to use in making these admissions decisions, but rather to increase faculty members’ knowledge of the decision making process itself and to encourage them to integrate several types of information when making the decision.  The proposed model, as designed, allows each program to develop their own standards and processes.  It also offers mechanisms for developing standardized decision processes for use across applicants and the types of offenses committed.  This model is a hybrid of statistical models and intuitive models and encourages the integration of each type of information as appropriate.  Furthermore, it is based on the strengths perspective. It begins with the belief that the applicant will not commit further offenses. Only when information is found which indicates that the applicant continues to pose a risk to the profession and/or clients, will they be denied admission.

Prior to implementation of this decision-making support system, each faculty member should become familiar with the concepts incorporated into the model.  They should be aware of the relationship between sensitivity and specificity and understand how these relate to the process of setting thresholds. They should understand the methods for increasing accuracy in decision-making and understand their own intuitive and emotional contributions to the process.  The proposed model is illustrated in Figure 2.

Figure 2.  The decision-making support system model for assessing applicants with criminal convictions.

Review of Application:  Initially every application is assessed using the standard review process.  As part of this process, a history of criminal convictions should be undertaken.  This process can be initiated within the application itself by having applicants respond to a question such as, “Have you ever been convicted of a felony or pled nolo contendere/no contest to any felony?”  Additionally, applicants' criminal backgrounds may be researched through the state's Central Registry or Department of Criminal Investigation.  A fee is often charged for such reports.  If the applicant is rated as acceptable but is found to have a history of a conviction then the decision-making support system is employed.  

Step 1:  Based on the type of crime committed, the admissions committee members should carefully assess the risks associated with a "Miss," or failing to detect and report a danger.  It is important to note that this risk is not based not on the applicant, but rather on the criminal offense which was committed. Committee members should also review how factors such as race, ethnicity, gender and socio-economic status can bias the criminal justice system.  The cost of a "False Positive," or rejecting an applicant who will not re-offend, must also be determined, within a social justice framework. Acknowledging the specific attributes of the applicant and what they can potentially contribute to the field should be taken into account.  This may be especially relevant if the applicant is a member of an under-represented group in the profession.

Step 2:  In this step of the model, the threshold for rejecting the applicant is determined.   In other words, the committee members must determine, based on the above review, how willing they are to make each type of error.  A lax threshold might be used when the outcome of a ‘miss’ is severe.  This ensures a high level of sensitivity in detecting potentially dangerous students.  For instance, if an applicant has a history of sexually or physically abusing children or other vulnerable populations, a lax threshold should be applied.  With such a threshold, even a small likelihood of re-offending will be reported as a danger and the applicant will be rejected from the program.  While this will result in a higher probability of false positives, (rejecting those applicants who are truly rehabilitated). it places the safety of potential clients at the forefront.  In contrast, a stricter threshold might be chosen when recidivism would not be especially dangerous to potential clients or the profession. The use of this strict threshold would forfeit sensitivity for specificity, reducing the rate of false positives.  

We recommend that the admissions committee as a whole work together to determine this threshold.  Disagreement amongst committee members in this step would be expected to result in lack of agreement in the final decision.  The role of emotional responses for each committee member must be carefully assessed. Likewise, recidivism rates should be closely examined.  

Step3:  In this stage, the committee members develop a narrative regarding the applicant, the criminal activity and post-conviction rehabilitation.  The goal of developing this narrative is to determine if there is a "signal," that is, an indication that the applicant will re-offend and thus represents a danger to clients and/or the profession.  The narrative will help the committee members understand the unique circumstances surrounding the applicant's life situation during the time the offense was committed and the applicant’s post-conviction behaviors, treatment and/or rehabilitation efforts.  The narrative should include all facets of information, incorporate alternative explanations, and integrate professional knowledge, including base rates of recidivism.

Likewise, understanding characteristics of  the justice system may also be important to take into account.   An important consideration is whether the person has negotiated a plea bargain.  Plea bargaining can result in a reclassification of the crime, leading the admissions committee to underestimate the severity of the criminal act.  Conversely, plea bargaining has been reported to result in a higher conviction rate than standing trial (Gorr, 2000).   Understanding the role of racial profiling, which has been documented by the U.S. Bureau of Justice Statistics (Bureau of Justice, 2001; Meehan & Ponder, 2002) must also be factored into the decision.

Additionally in Step 3, the role of personal intuition might be included in order to increase the accuracy of the narrative.  Since intuition is most valuable when responding to non-verbal communication, the use of an interview is recommended.  The committee members are encouraged to assess their personal responses to the applicant and to integrate these responses with their previous experiences and knowledge. Furthermore, the members must incorporate an emotional awareness into this stage, in order to reduce biases.

Step 4:  Based on the narrative developed in Step 3, the committee members will assess the completeness of story, potential biases and errors, and contradictory information, and will now make a decision as to whether they believe the applicant will re-offend.  It is expected that some level of uncertainty will remain with this decision.  Therefore the committee members must also determine their level of confidence in their decision. 

Step 5:  In this step, the committee members compare their confidence in their decision of whether they believe the applicant will re-offend, from Step 4, to the predetermined threshold for rejecting the applicant determined in Step 2.  As discussed above when a lenient threshold is determined, rejection of the applicant may be the required decision even though there is very little evidence that the applicant will re-offend.  With a moderate threshold, if there is little evidence to support the judgment that the applicant continues to pose a risk, then the applicant can be accepted.   With a strict threshold, an applicant might be accepted even though there is moderate evidence that they will commit further offenses.  

Step 6:  The final step in the model is for the admissions committee as a whole to make a decision.  A consensus building process, rather than a simple majority vote, might be advantageous.  In this way, the personal experiences and intuition of each of the committee members can be expressed and evaluated by other members.

After the decision has been made, further consequences may be involved.  If the applicant is denied admission, the policies and the decision-making process may be called into question by the applicant or counsel for the applicant.  If the applicant is admitted to the program, additional decisions regarding educational opportunities may be necessary.  For instance, the program may decide to limit the field placement, based on the nature of the previous criminal conviction, the agency, and the type of clients with whom the student wishes to work. Furthermore, it will be important to ensure that the applicant realizes that completion of the program does not guarantee they will be eligible to be licensed by the state in which they wish to practice. Each MSW Program should work directly with university legal advisors to develop statements and signature forms to reduce the program’s liability in such instances.


When potential students who have previous criminal convictions apply to our social work programs, it is imperative that we have a policy which can guide our decision-making process.  Two basic policies have been proposed.  The first is a blanket policy to reject all such applicants (Megan & Emerman, 2000).  This is a response to the discomfort with making risky decisions when uncertainty is involved. Applying this policy is conservative and results in the least risk to the program, the profession and potential clients.   An alternative policy is to review applicants on a case-by-case basis.  This process requires that thorough assessment of applicant characteristics and the criminal offense be reviewed.  While numerous relevant characteristics have been proposed (Scott, & Zeiger, 2000), no systematic process for making the decision exists.  

Perhaps the ongoing debate regarding the development of admission policies is a reflection of the fact that we have not yet developed a theoretical perspective to guide our decision making.  Such a model will require that we carefully weigh both external information such as recidivism rates and probabilities, with our more internal, intuitive responses.  Developing a model that takes both types of responding into account may provide the framework to move beyond our stalemate.

The current paper has proposed a model to fill this need.  We believe that understanding the processes of decision making under conditions of uncertainty, and using the proposed decision-making support system, can greatly increase the faculty member’s ability to make consistent and equitable decisions which are in keeping with our ethical mandates. If applied consistently, decision making would be more standardized and reliable, resulting in a lower probability of making idiosyncratic or biased decisions.  


Bandura, A. (1986). Social foundations of thought and action. New Jersey: Prentice Hall, Inc.

Barbalet, J. (2001). Emotion, social theory and social structure: A macrosociological approach. New York: University Press.

Bayard, D. R. (2001) Finding our balance: The investigation and clinical application of intuition. Psychotherapy Theory, Research, Practice, Training: Special Issue, 38, 97-106.  

Benbenishty, R., & Treistman, R. (1998).  The development and evaluation of a hybrid decision support system for clinical decision making: The case of discharge from the military.  Social Work Research, 22(4), 195-203.

Born, C. E., & Carroll, D. J. (1988).  Ethics in admissions.  Journal of Social Work Education, 24, 79-85.

Buttell, F. P., & Carney, M. M. (2002).  Psychological and demographic predictors of attrition among batterers court ordered into treatment.  Social Work Research, 26(1), 31-41.

Clore, G.  (1994). Why emotions require cognition. In P. Ekman & R. J. Davidson (Eds.), The Nature of emotion: Fundamental questions (pp. 181-191). New York: Oxford University Press.

Cole, B. S. (1991).  Legal issues related to social work program admissions.  Journal of Social Work Education, 27, 18-24.

Cole, B. S., & Lewis, R. G. (1993).  Gatekeeping through termination of unsuitable social work students: Legal issues and guidelines.  Journal of Social Work Education, 29, 150-159.

Council on Social Work Education (1994).  Handbook of accreditation standards and procedures.  Alexandria, VA: Author.  

Council on Social Work Education (2003).  Handbook of accreditation standards and procedures.  Alexandria, VA: Author.  

Dunlap, K. M., Henley, H. C., & Fraser, M. W. (1998).  The relationship between admissions criteria and academic performance in an MSW program. Journal of Social Work Education, 34(3), 455-462.

Evans, D. (2001). Emotion: The science of sentiment. Oxford: University Press.

Forgas, J. P. & Vargas, P. T. (2000). The effects of mood on social judgment and reasoning. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of emotions (2nd Ed.) (pp. 350-367). New York: Oxford University Press.

Gelman, S. R., & Wardell, P. J. (1988).  Who’s responsible: The field liability dilemma.  Journal of Social Work Education, 24, 70-78.

Gibbs, P. (1994).  Gatekeeping issues in BSW programs.  Journal of Social Work Education, 30, 63-74.

GlenMaye, L., & Oakes, M. (2002). Assessing suitability of MSW applicants through objective scoring of personal statements. Journal of Social Work Education, 38(1), 67-82.

Goleman, D. (1995). Emotional intelligence: Why it can matter more than IQ. New York: Bantam Books.

Gorr, M. (2000).  The morality of plea bargaining.  Social Theory and Practice, 26(1),  129-151.

Green, D. M., & Swets, J. A. (1966).  Signal detection theory and psychophysics.  New York: John Wiley & Sons.

Haidt, J. (2000).  The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108, 814-834.

Hall, K. H. (2002).  Reviewing intuitive decision-making and uncertainty: the implications for medical education. Medical Education, 36(3), 216-224.

Hastie, R., & Pennington, N. (2000).  Explanation-based decision making. In Connolly, Terry (Ed); Arkes, Hal R. (Ed); et al. Judgment and decision making: An interdisciplinary reader (2nd ed.). Cambridge series on judgment and decision making (pp. 66-81). New York: Cambridge University Press.

Khatri, N., & Ng, H A. (2000). The role of intuition in strategic decision making.  Human Relations, 53, 57- 86.

Johnson, M. A., Brown, C. H., & Wells, S. J. (2002).  Using classification and regression tress (CART) to support worker decision making.  Social Work Research, 26(1), 19-29.

Lieberman, M. D. (2000). Intuition: A social cognitive neuroscience approach. Psychological Bulletin, 126(1), 109-137.

Loewenberg, F., Dolgoff, R., & Harrington, D. (2000).  Ethical decisions for social work practice (6th ed.). New York: F. E. Peacock.

Luoma, B. B. (1997).  A case for intuition in the social work curriculum.  The Journal of Baccalaureate Social Work, 3(1), 49-60.

Mattison, M. (2000).  Ethical decision making: The person in the process.  Social Work, 45(3), 201-212.

Megan, R. H., & Emerman, J.  (2000).  Should convicted felons be denied admission to a social work education program? --Yes! Journal of Social Work Education, 36, 401-414.

Meehan, A. J., & Ponder, M. C. (2002).  Race and place: the ecology of racial profiling African American motorists.  Justice Quarterly, 19(3), 399-430.

Miller, J., & Koerin, B.  (1998).  Can we assess suitability at admission? A review of MSW application procedures.  Journal of Social Work Education, 34(3), 437-453.

Mossman, D. (1994)  Assessing predictions of violence: Being accurate about accuracy.  Journal of Consulting and Clinical Psychology, 62, 783-792.

National Association of Social Workers (1996).  NASW Code of Ethics.  Washington, DC: Author.

Pelech, W., Stalker, C. A., Regehr, C., & Jacobs, M. (1999).  Making the grade: The quest for validity in admissions decisions. Journal of Social Work Education, 35(2), 215-226.

Scott, N. & Zeiger, S. (2000).  Should convicted felons be denied admission to a social work education program? --No! Journal of Social Work Education, 36, 401-414.

Shweder, R. (1994). You’re not sick, you’re just in love: Emotions as an interpretive system. In P. Ekman & R. J. Davidson (Eds.), The nature of emotion: Fundamental questions (pp. 32-44). New York: Oxford University Press.

Silver, E., Smith, W. R., & Banks, S. (2000).  Constructing actuarial devices for predicting recidivism: A comparison of methods.  Criminal Justice and Behavior, 27(6), 733-764.

Silver, E., & Miller, L. L. (2002).  A cautionary note on the use of actuarial risk assessment tools for social control.  Crime and Delinquency, 48(1), 138-161.

Swets, J. A. (1986). Indices of discrimination or diagnostic accuracy: Their ROCs and implied models.  Psychological Bulletin, 99, 100-117.

Swets, J. A. (1992).  The science of choosing the right decision threshold in high-stakes diagnostics.  American Psychologist, 47, 522-532.

Swets, J. A. (2000)  Enhancing Diagnostic Decisions. In Connolly, Terry (Ed); Arkes, Hal R. (Ed); et al. Judgment and decision making: An interdisciplinary reader (2nd ed.). Cambridge series on judgment and decision making (pp. 66-81). New York: Cambridge University Press.

U. S. Bureau of Justice Statistics (2001).  Contacts between the police and the public: Findings from the 1999 survey.  Washington, DC: U.S. Department of Justice.

U. S. Bureau of Justice Statistics (2002). Recidivism of Prisoners Released in 1994, Special Report NCJ 193427.  Washington, DC: U.S.Department of Justice.

Younes, M. H. (1998). The gatekeeping dilemma in undergraduate social work programs: Collision of ideal and reality.  International Social Work, 41, 145-153.

Gail M. Leedy , MSW , PhD

Division of Social Work

Department 3632

University of Wyoming

1000 E. University Ave

Laramie , Wyoming 82071

(307) 766-2562


James E. Smith, MSW , PhD

Division of Social Work

Department 3632

University of Wyoming

1000 E. University Ave

Laramie , Wyoming 82071

(307) 766-5639

Last Updated ( Thursday, 17 March 2005 )


Joomla Templates by Joomlashack