Documente Academic
Documente Profesional
Documente Cultură
http://er.aera.net
Published on behalf of
and
http://www.sagepublications.com
Additional services and information for Educational Researcher can be found at:
Email Alerts: http://er.aera.net/alerts
Subscriptions: http://er.aera.net/subscriptions
Reprints: http://www.aera.net/reprints
Permissions: http://www.aera.net/permissions
ach day parents watch their children go off to school, trusting the system of education to keep them safe. Because the
experiences of their children and periodic media coverage
may increase their concern, parents and others naturally ask:
How safe are our schools? The response of researchers has been
that schools are generally safe (Dinkes, Cataldi, & Lin-Kelly,
2007), but this alone has not fully quelled these concerns because
of a lack of consensus about what constitutes safety. Do the relatively rare yet tragic high-profile school shootings represent a
greater concern than the long-term psychological effects of dayto-day bullying, intimidation, and incivility experienced by students at school? What is the connection between school safety
statistical reports based primarily on survey indicators and the
real-life experiences of students in schools?
General Awareness of School Violence and Disorder
Aspects of school violence have been a concern throughout modern American history (see Cornell & Mayer, this issue of
Educational Researcher, pp. 715) but came to the forefront with
the issuance of the 1978 report Violent SchoolsSafe Schools: The
Safe School Study Report to Congress (U.S. National Institute of
Education, 1978). A report in 1984, Disorder in Our Public
Schools (U.S. Department of Education, 1984), pointed to ongoing issues with school disorder nationally, concomitant with
increases in juvenile crime (Snyder & Sickmund, 1999). Although
not widely recognized, school violence incidents peaked around
16
educational Researcher
School Disorder
Outsider Adult
Shootings at School
School-Associated Shootings
(e.g., Columbine)
Marginally Acceptable
Behaviors
Theft
FIGURE 1. Conceptual representation of school disorder compared with overall student behaviors. Adapted from School violence and
disruption by M. J. Mayer, 2008, SAGE Encyclopedia of Educational Psychology, pp. 880-888, Thousand Oaks, CA: Sage.
given as the number (or rate) of incidents of a condition of interest (event or disease) present in a population at a given point in
time. Incidence is generally given as the rate of new events emerging within a specified time frame, as in the instances of new disease outbreak annually, and represents the risk of experiencing
the condition. As discussed by Borum, Cornell, Modzeleski,
and Jimerson in this special issue of Educational Researcher
(pp. 2737), the public could perceive schools as unsafe based on
media reports of the incidence of school-based homicides (e.g.,
21 during a particular year), although the actual prevalence of
school-based homicides is extremely low (the typical school
would experience such an occurrence once every 6,000 years).
The school safety waters are further muddied when the concept of harma relative constructis considered. Do horrific
school shootings, albeit rare, yet devastating for all involved, constitute a public health risk? Statistically, based on an average of 21
student deaths per year, nationally, from 1996 to 2006, the
answer would likely be no and lead to the conclusion that schools
are among the safest places for Americas youth (Modzeleski et al.,
2008). However, this provides little consolation to the bereaved
families, friends, teachers, and others affected, not to mention
parents who fear that similar harm may befall their children while
in school. Conversely, should widespread and long-term day-today bullying, intimidation, and incivility in schools be considered a serious public health risk factor? In this instance, the
answer would probably be yes, based on recent data wherein 32%
of secondary students reported being bullied at or around school
(Dinkes et al., 2009) and several lines of research that point to
pronounced long-term psychosocial harm to large numbers of
affected youth (Arseneault et al., 2006; Bierman, 2004; Boxer,
Edwards-Leeper, Goldstein, Musher-Eizenman, & Dubow,
2003; Ladd, 2003; Nansel et al., 2001). Thus far, the school
safety field lacks consensus on how to approach these matters,
and this situation is impeding movement toward forming a
national safe school agenda.
Despite limited consensus on essential school safety indicators, researchers have developed a picture of trends in school violence and disorder at the national level taken from multiple data
17
180
160
140
120
100
80
60
40
20
0
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
At School
92
72
73
64
73
55
56
63
78
74
61
55
60
48
46
49
Year
FIGURE 2. Rate of total crime against students ages 12 to 18, per 1,000 students, at and away from school, 19922006. Total crime
includes theft, violent crime, and serious violent crime.
Percentage of Respondents
95% C.I.
12
10
8
6
4
2
0
1993
1995
1997
1999
2001
2003
2005
2007
7.3
8.4
7.4
7.7
8.9
9.2
7.9
7.8
4.4
4.5
5.2
6.6
5.4
5.5
Year
educational Researcher
Downloaded from http://er.aera.net at Midlands State University on August 25, 2014
We believe that the low number of schools identified as persistently dangerous is also a byproduct of ambiguity and a lack of
consensus over which safety indicators and databases are most
salient and what is actually being measured. Crime incident data
can lead to biased estimates of the extent of violent acts and number of perpetrators in a community because they reflect arrests, not
adjudications and convictions, and do not include behaviors that
escape law enforcement due to lack of detection, victims failure to
report crimes, or system inability to pursue cases. Victimization
self-reports via the NCVS are also vulnerable to problems, including sampling frame and instrumentation, and respondent errors,
such as poor recall, comprehension difficulties, and telescoping
effects (inaccurately recalling sooner or later than events occurred;
Biemer, Groves, Lyberg, Mathiowetz, & Sudman, 1991). Major
national surveys have demonstrated risk of bias driven by extreme
response sets (responses that tend to be anchored at one part of a
scale; Furlong, Sharkey, Bates, & Smith, 2004) and inferential
limitations due to uses for which the surveys were never validated
(Furlong & Sharkey, 2006). For example, citing work by Bachman
and colleagues, Furlong and Sharkey reported on the Monitoring
the Future Study, where validity of questions about substance use
was not empirically established. Large-scale surveys tapping into
similar domains can vary significantly with respect to analysis
goals linked to instrumentation design and can produce data that
are incongruent across differently defined and targeted populations, types of violence or disorder, and time frames (Leone et al.,
2000; Reiss & Roth, 1993; Sharkey, Furlong, & Yetter, 2006).
More localized surveys have demonstrated problems in terms of
consistently following standardized administration protocols,
directions to respondents differing across administrations, and
local survey providers often being untrained and unprepared for
survey administration (Cross & Newman-Gonchar, 2004).
Beyond survey administration issues, there are problems with
how results and other school safety data are organized, interpreted, and disseminated among schools and allied agencies.
There is no uniform school safety data collection and recording
framework, and practices vary, for example in the recording of
duplicated versus unduplicated (e.g., counting multiple suspensions of a particular student as one suspension) counts and protocols for reporting based on individuals versus incidents.
Systems-change initiatives (e.g., positive behavioral supports)
that often use outcome measures such as office disciplinary referrals have been criticized for not necessarily capturing the most
relevant data reflecting violence and disruption in a school
(Morrison, Peterson, OFarrell, & Redding, 2004). Likewise, differing approaches to recording suspension data across school districts can result in reliability and validity problems, hindering
meaningful interpretation (Morrison, Redding, Fisher, &
Peterson, 2006). A noteworthy investigation in New York State
uncovered serious underreporting of school violence in many
schools, where about one third of a sample of audited schools
failed to report approximately 80% of violent incidents (Office
of the New York State Comptroller, 2006). The State of
California, in fact, started formal school crime reporting twice,
and both efforts were discontinued because it was evident that
not all crimes were being reported and that reporting was influenced by administrator idiosyncratic interpretation of the reporting requirements (California Legislative Analyst Office, 2002).
Beyond considering definitional and measurement issues, questions emerge regarding linkage to theory.
Linkage to Theoretical Models
School safety and order encompass a wide range of student and
adult behaviors and systemic processes, going beyond a narrower
focus on school violence. Cross-cutting threads in school safety
involve well-run schools with positive student engagement and
outcomes and overall psychological and physical safety for all
stakeholders. A reasonable expectation for researchers and other
stakeholders in the social and behavioral sciences would be that
major national surveys of important social issues (e.g., violence,
drug use, health habits) would be linked to established theoretical
foundations and well-conceived conceptual frameworks. But this
is not necessarily the case. Johnston, OMalley, Schulenberg, and
Bachman (2001) provided insightful discussion of how theory
applied to drug abuse meshed more or less well with goals of the
Monitoring the Future Survey. Several key points they raised with
respect to drug use might be applied to violence and other socialpsychological concerns.
First, the history of social and behavioral inquiry, including
research on substance abuse, includes dozens of theoretical
models and their derivatives, addressing (a) types of involvement;
(b) stages of development; (c) sets of social, emotional, and behavioral factors purportedly linked to outcomes; (d) foundational
theories of human behavior; and (e) units of measurement and
analysis. Second, no theories are embraced by most of the field,
and theory development remains in process. Third, citing comments of Merton (1957), Johnston et al. (2001) pointed to much
of the research embracing middle-range theories that map to
minor testable hypotheses; this research did not entail a coherent explanatory approach driving variable selection but, rather, a
more general orientation to relevant variables. Fourth, citing
Cattells (1966) description of an inductive-hypothetico-deductive spiral, Johnston et al. argued that in the absence of coherent
theory, a somewhat more eclectic approach is necessary, where
through iteration, particular theoretical positions drive specific
empirical tests, which in turn influence refinement of theory.
Taken as a group, these considerations suggest that the school
violence field is developing in a manner similar to other areas of
social-behavioral research, such as substance abuse, in that at
present it lacks a unifying framework and is more often studied
in a piecemeal fashion within discipline domainseducation,
psychology, sociology, criminology, and public healththat
communicate incompletely with one another. This theoretical
fragmentation may contribute to the so-called research-to-practice gap.
The desire to better understand the status of school safety is
linked to future research, policy, funding, and programmatic
implementation to address need. Developing theoretical coherence is critical for purposes of developing a system of meaning
that informs research, practice, and planning for the future. For
school violence and safety data collection and analysis, the interest often is focused on learning more about (a) victimization
experiences; (b) characteristics of the individuals and schools;
(c) systemic factors, such as how the schools system of rules is
understood and implemented; (d) risk and protective factors
across ecological levels; and (e) related contextual variables, such
january/February 2010
19
issues, as well as respondent ages, response coding rules, and patterns of concordance of different surveys over time.
Stakeholders examining year-to-year changes in data need to
be mindful of point versus interval estimates (e.g., means and
confidence intervals). Statistics from national school safety surveys using probability samples include interval estimates that are
driven by standard error. Although it is tempting to read the data
as point estimates, this can be misleading. For instance, the 25%
increase in serious violent crime from 201,800 in 1997 to
252,700 in 1998 seems noteworthy but is statistically nonsignificant at the .05 level as a result of high standard errors. Other
factors must also be considered when evaluating trend data for
change. Major national surveys reviewed here use complex sampling designs (clustering, stratification, and unequal selection
probability weighting) that can entail multiple methods for calculating the standard error of statistics, which are typically larger
than those of a simple random sample due to the design effect (the
ratio of the variance of a statistic under a particular sample design
to the variance of that statistic under simple random sampling for
a sample of equal size; Kish, 1964; Kish & Frankel, 1974). This,
in turn, affects interpretation of change.
Do seeming contradictions in trends based on different indicator variables constitute a problem, or are they a benefit, helping
investigators better understand disparate aspects of school safety?
Rand and Rennison (2002) discussed apparent contradictions
between a NCVS-reported 15% drop in crime from 1999 to
2000 and a Uniform Crime Reporting Program report showing
year 2000 figures at a stable level compared with previous year
data. They noted that year-to-year changes in violent crimes
across the two surveys moved in the same direction about 60% of
the time. Rand and Rennison identified multiple aspects of the
surveys that accounted for differing statistics on crime, including
who was being measured, counting rules for multiple victims versus incidents, victimization measurement of persons versus
households, and protocols for recording a series of victimizations.
For example, if a student was victimized several times in one day
by the same bully perpetrator, should this be recorded as a single
victimization or several? Protocols can vary. In addition, socialenvironmental factors may contribute to differential reporting
behaviors regarding crime and victimization, such as crime victims perception of system responses to their reports.
Meaning and utilitynot just beautyrests in the eye of the
beholder, particularly when public policy issues are involved.
Above and beyond the topical content focus of a data collection
system are the end users who dictate relevance and utility. Justice
system officials developing policy, legislators addressing crime,
law enforcement officials developing interventions, and criminologists studying these issues all would be interested in officially
reported crime data to support their efforts. Epidemiologists
studying patterns of violence, policy makers developing a national
strategic agenda, lawmakers approving funding, and local human
services agencies providing support programs are each uniquely
dependent on victimization data collected through surveys.
Specific survey data efforts such as the CDCs YRBSS may better
support the work of epidemiologists than the needs of federal,
state, and local policy makers trying to respond to violence and
disruption in the schools. The YRBSS can help track the emergence and development of targeted risk behaviors of interest,
educational Researcher
Downloaded from http://er.aera.net at Midlands State University on August 25, 2014
21
These highlights taken from national surveys point to continuing safety problems in schools. Some of these indicators are
at less than half the levels seen in the early 1990s, when school
violence was considered by many as having reached epidemic levels (Elliott et al., 1998), which suggests that progress has been
made toward the reduction of school violence incidents. However,
there is no definitive proof of what drove the reductions, and
other statistics have remained fairly consistent. As a group, these
data suggest that significant violence, bullying, and related
threatening and intimidating behaviors continue in schools
(Mayer, in press), a circumstance associated with students experiencing fear and anxiety and altered academic and social-emotional trajectories (see Cornell & Mayer, this issue).
While considering these data on school violence and disorder,
it may be instructive to consider an analogy to the harm caused to
children by child abuse and neglect. The literature in psychiatry,
psychology, education, social work, and juvenile justice paints a
general picture of great difficulties children experience as a result
of multiple forms of abuse and neglect (Johnson, Cohen, Brown,
Smailes, & Bernstein, 1999; Keiley, Howe, Dodge, Bates, &
Pettit, 2001). Research on resilience (Luthar & Cicchetti, 2000;
Masten et al., 1999; McGloin & Widom, 2001) suggests that
some factors act as buffers or sources of protection that reduce the
odds of having deleterious developmental outcomes. There is general agreement regarding the likelihood of harm from severe
abuse. However, the effects for children who experience chronic,
long-term neglect, although overlapping, are somewhat different.
Children who experience neglect over the long haul are generally at elevated risk for poor academic performance, developing
limited social communication skills, experiencing poor peer relations, and having some degree of emotional-behavioral difficulties (Hildyard & Wolfe, 2002; Kendall-Tackett& Eckenrode,
1996; Widom & Maxfield, 1996). The harm is real. Yet social
systems aimed at protecting children from abuse and neglect may
not respond as well in such instances because (a) much of the
damage occurs very slowly over time, (b) there is no highly tangible evidence of harm (e.g., bruises on the body) often associated
with physical abuse, and (c) the systemic resources required to
ameliorate such long-term neglect are more extensive and require
long-term strategic approaches.
Much like instances of obvious severe physical and sexual
abuse, the educational and community systems respond quickly
and decisively to instances of extreme violence in the schools
(e.g., school shootings). The harm is clearly evident, but the systems try to stabilize the situation through crisis intervention to
make sure no further harm occurs as a result of the violence.
Despite the attention that incidents of deadly violence have
received, situations of serious but not life-threatening disruption
or violence are more common, and more difficult for educational
systems to address. In fact, in Davis v. Monroe County Board of
Education (1999), the U.S. Supreme Court ruled in favor of a
student claiming harm resulting from chronic harassment at
school that school authorities did not adequately address. In light
of this ruling, it is possible to consider these ongoing day-to-day
problems as constituting a form of systemic educational neglect
when not meaningfully addressed. Like neglect on an individual
level, the persistence in schools of lower level bullying, harassment, threatening behaviors, and incivility is more difficult for
22
educational Researcher
Downloaded from http://er.aera.net at Midlands State University on August 25, 2014
(Braga & Kennedy, 2001); (b) prevalence of local firearm ownership, accounting for 47% of the variance in firearm-related youth
deaths (Murnan, Dake, & Price, 2004); (c) increased likelihood
of having loaded and unlocked guns at home among families in
communities with high crime and gang activity (Vacha &
McLaughlin, 2004); (d) widespread practice, in homes with children (2 million or more), of storing loaded and unlocked guns
(Okoro et al., 2005; Schuster, Franke, Bastian, Sor, & Halfon,
2000); (e) higher weapon-carrying rates among inner-city youth,
compared with suburban and rural youth, in school and community (Sheley, McGee, & Wright, 1995); and (f ) relatively easy
availability of firearms, per student self-reports (Brown, 2004;
Sheley & Wright, 1998). Clearly, managing gun access by youth
is a national-, state-, and community-level challenge far beyond
the purview of schools.
What is known about the infusion of community-level violence into the schools? Most school violence studies have focused
on school or student characteristics but have not concurrently
analyzed data across multiple levels of the ecology (Watkins,
2008). Three recent studies addressed this issue. Welsh, Greene,
and Jenkins (1999) studied communities and schools in
Philadelphia, finding that schools could offer programming that
mitigated the harmful effects of community violence and that
schools were not necessarily at the mercy of nearby community
violence. That study, which received methodological criticism
(Hoffman & Johnson, 2000), stood somewhat alone until a
study of schools in Chicago addressed similar issues (Kirk, 2006).
The Chicago study likewise found that schools could be successful while embedded in a community experiencing high violence
rates, with school program effects operating somewhat independent of community factors. The research also identified the ameliorative role of families active participation in school activities
and administrationan outcome negatively associated with poverty (Evans, 2004)and found an association of reduced delinquency with increased academic engagement of students. Watkins
(2008) examined community, school, and student influences on
weapon carrying in school, analyzing survey responses from a
subset of students in Grades 8 to 12 from the 19941995 administration of the National Longitudinal Study of Adolescent
Health. The analysis, while having limitations, showed no systematic relation between student weapon carrying and surrounding community poverty, residential turnover, and community
violent crime.
The previous discussions suggest the need for continued
research to better explain factors and processes across ecological
levels. There may be more questions than answers regarding
school violence, what is going on in Americas schools, and what
can be done about it.
Unanswered Questions and Next Steps in Research,
Policy, and Practice
Although there are dozens of unanswered questions about school
violence and disorder and the current safety of schools, it is
important to identify a core set as high-priority targets. These
questions pertain to this articles focus on the overall safety of
schools but go further to address other concerns associated with
school safety. Here we identify five key questions that we argue
23
Arseneault, L., Walsh, E., Trzesniewski, K., Newcombe, R., Caspi, A., &
Moffitt, T. (2006). Bullying victimization uniquely contributes to
adjustment problems in young children: A nationally representative
cohort study. Pediatrics, 118, 130138.
Biemer, P. P., Groves, R. M., Lyberg, L. E., Mathiowetz, N. A., &
Sudman, S. (1991). Measurement errors in surveys. New York: John
Wiley.
Bierman, K. L. (2004). Peer rejection. New York: Guilford.
Borum, R., Cornell, D. G., Modzeleski, W., & Jimerson, S. R. (2010).
What can be done about school shootings? A review of the evidence.
Educational Researcher, 39, 2737.
Boxer, P., Edwards-Leeper, L., Goldstein, S. E., Musher-Eizenman, D.
R., & Dubow, E. F. (2003). Exposure to low level aggression in
school: Effects on aggressive behavior, future expectations, and perceived safety. Violence and Victims, 18, 691704.
Braga, A. A., & Kennedy, D. M. (2001). The illicit acquisition of firearms by youth and juveniles. Journal of Criminal Justice, 29, 379388.
Brown, B. (2004). Juveniles and weapons: Recent research, conceptual
considerations, and programmatic interventions. Youth Violence and
Juvenile Justice, 2, 161184.
California Legislative Analyst Office. (2002). Review of school crime
reporting. Sacramento: Author.
Centers for Disease Control and Prevention. (2008). School-associated
student homicidesUnited States, 19922006. MMWR, 57(02),
3336.
Cross, J. E., & Newman-Gonchar, R. (2004). Data quality in student
risk behavior surveys and administrator training. Journal of School
Violence, 3, 89108.
Cornell, D. G., & Mayer, M. J. (2010). Why does school order and
safety matter? Educational Researcher, 39, 715.
Davis v. Monroe County Board of Education, 526 U.S. 629 (1999).
Dinkes, R., Cataldi, E. F., & Lin-Kelly, W. (2007). Indicators of school
crime and safety: 2007 (NCES 2008021/NCJ 219553). Washington,
DC: National Center for Education Statistics, Institute of Education
Sciences, U.S. Department of Education, and Bureau of Justice
Statistics, Office of Justice Programs, U.S. Department of Justice.
Dinkes, R., Kemp, J., & Baum, K. (2009). Indicators of school crime and
safety: 2008 (NCES 2009022/NCJ 226343). Washington, DC:
National Center for Education Statistics, Institute of Education
Sciences, U.S. Department of Education, and Bureau of Justice
Statistics, Office of Justice Programs, U.S. Department of Justice.
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J., Hawkins, J.,
et al. (2008). Youth risk behavior surveillanceUnited States, 2007.
Surveillance Summaries, June 6, 2008. MMWR, 57(No. S S04), 1131.
Elliott, D., Hamburg, B., & Williams, K. (Eds). (1998). Violence in
American schools. New York: Cambridge University Press.
Evans, G. W. (2004). The environment of childhood poverty. American
Psychologist, 59, 7792.
Furlong, M. J., Paige, L. Z., & Osher, D. (2003). The Safe Schools/Healthy
Students (SS/HS) Initiative: Lessons learned from implementing
educational Researcher
Downloaded from http://er.aera.net at Midlands State University on August 25, 2014
25
Schuster, M. A., Franke, T. M., Bastian, A. M., Sor, S., & Halfon, N.
(2000). Firearm storage patterns in US homes with children. American
Journal of Public Health, 90, 588594.
Sharkey, J. D., Furlong, M. J., & Yetter, G. (2006). An overview of
measurement issues in school violence and school safety research. In
S. R. Jimerson & M. J. Furlong (Eds.), Handbook of school violence and
school safety: From research to practice (pp. 121134). Mahwah, NJ:
Lawrence Erlbaum.
Sheley, J. F., McGee, Z. T., & Wright, J. D. (1995). Weapon-related victimization in selected inner-city high school samples. Washington, DC:
National Institute of Justice.
Sheley, J. F., & Wright, J. D. (1998, October). High school youth, weapons, and violence: A national survey. Washington, DC: National
Institute of Justice.
Skiba, R. J., Simmons, A. B., Peterson, R., McKelvey, J., Forde, S., &
Gallini, S. (2004). Beyond guns, drugs, and gangs: The structure of
student perceptions of school safety. Journal of School Violence, 3,
149171.
Snell, L. (2005, January). School violence and No Child Left Behind: Best
practices to keep kids safe. Los Angeles: Reason Foundation.
Snyder, H. N., & Sickmund, M. (1999). Juvenile offenders and victims:
1999 national report. Washington, DC: Office of Juvenile Justice and
Delinquency Prevention.
Stroul, B. (2002). Issue briefSystem of Care: A framework for system
reform in childrens mental health. Washington, DC: Georgetown
University Child Development Center, National Technical Assistance
Center for Childrens Mental Health.
Taylor, T. J., Freng, A., Esbensen, F. A., & Peterson, D. (2008). Youth
gang membership and serious violent victimization: The importance
of lifestyles/routine activities. Journal of Interpersonal Violence, 23,
14411464.
Thomas, D. E., Bierman, K. L., & the Conduct Problems Prevention
Research Group. (2006). The impact of classroom aggression on the
development of aggressive behavior problems in children. Development
and Psychopathology, 18, 471487.
Thornberry, T., Huizinga, D., & Loeber, R. (2004). The causes and correlates studies: Findings and policy implications. Juvenile Justice
Journal, 9(1), 319. Washington, DC: U.S. Department of Justice,
Office of Justice Programs, Office of Juvenile Justice and Delinquency
Prevention.
United States. (2004). Keeping schools safe: The implementation of No
Child Left Behinds persistently dangerous schools provision (Field hearing before the Subcommittee on Education Reform of the Committee
on Education and the Workforce, U.S. House of Representatives,
26
MATTHEW J. MAYER is an assistant professor of educational psychology in the Graduate School of Education at Rutgers University, 10
Seminary Place, New Brunswick, NJ 08901; mayerma@rutgers.edu. His
research includes analyzing national data and modeling school violence
and disruption processes.
MICHAEL J. FURLONG is a professor the Department of Counseling,
Clinical, and School Psychology, Gevirtz Graduate School of Education,
University of California, Santa Barbara, CA 93106; mfurlong@education
.ucsb.edu. His research focuses on school violence prevention research.
He is the editor of the Journal of School Violence.
educational Researcher
Downloaded from http://er.aera.net at Midlands State University on August 25, 2014