Women in Prison: A CrossSectional Analysis of Female Incarceration

1
Women in Prison A Cross-Sectional Analysis of
Female Incarceration

• Zohre Salehezadeh
• Kenneth Kickham
• Robert Bentley
• Oklahoma Department of Human Services
• National Association for Welfare Research and
  Statistics Conference
• Tuesday, August 24th 2004, Oklahoma City, OK

2
Women in PrisonNational Data

• In the United States, the rate of female
  incarceration has increased an average of 5.2
  percent per year since 1995.1
• Between 1988 and 1997, the number of crime cases
  involving females under age 16 increased 89
  percent, while the number of cases involving
  females age 16 or older grew 74 percent.2
• In 2002, sixty out of every 100,000 female U.S.
  residents were behind bars.3

3
Women in PrisonNational Data

• Between 1990 and 1998
• The number of female defendants convicted of
  felonies in state courts has grown at twice the
  rate of the increase in male defendants 4
• The number of women per capita involved in
  corrections overall has grown 48 compared to a
  27 increase in the number of men per capita.5
• Of the female offenders released from prison
  during 1994, about 58 percent were rearrested.6

4
Comparing the States
Source U.S. Dept. of Justice, Bureau of Justice
Statistics
5
Female Incarceration and Male Incarceration

• State female incarceration rate is highly
  correlated with male incarceration rate.
• Oklahoma ranks fourth among states in total
  incarceration rate.
• Oklahoma ranks fifth in male incarceration rate.
• Oklahoma leads all states in terms of female
  incarceration rate.7

6
Crime and Incarceration in Oklahoma

• From 1974 to 2001, Oklahomas crime rate
  increased 14 percent, from 4,050 per 100,000
  population in 1974 to 4,619 in 2001.
• The incarceration rate increased by more than 440
  percent, from 121 per 100,000 in 1974 to 658 in
  2001.
• The incarceration rate continued to increase over
  the 1996-2000 period, in spite of declining crime
  rate.8

7
(No Transcript)
8
Research Questions

• Why is Oklahomas female incarceration rate so
  high?
• What are the possible factors correlated with
  state variations in female incarceration rates?

9
Research Approach

• Review the literature to find the theoretical
  causes of female incarceration.
• Search for data that fits the theories.
• Develop a regression model with variables that
  tend to explain the variation in female
  incarceration rates across the states.
• Discuss the implications.

10
What Does the Literature Suggest?

• Many possible causes, including
• History of abuse or neglect
• Unstable family structure
• Poverty
• Mental health and substance abuse
• Public policies
• Severity of criminal justice system
• Effectiveness of indigent defense system
• Availability of social services

11
Oklahoma Data Supporting Possible Theoretical
Causes

• In a survey9 of Oklahoma female inmates, more
  than 71 percent said they had been involved in an
  abusive relationship.
• Forty-two percent lived with a single parent or
  other relatives as a child. 10
• Sixty-one percent of these women had used drugs
  and 33 percent had used alcohol. 11

12
Data Set and Model

• Our data set contains the fifty states, with the
  state values for over 70 variables, being
  measured at various times between 1996 and 2002.
  
• We use regression modeling to see which variables
  can explain variation among the states in female
  incarceration rates.
• The dependent variable is the number of
  incarcerated females (with sentences of more than
  one year) per 100,000 female residents in 2002.
• The state values range from 11 (MA and RI) to 131
  (OK) female prisoners (natl. avg. 54).12

13
Many Variables Didnt Work

• There were numerous variables that were expected
  to affect the rate of female incarceration, yet
  did not meet the criteria for inclusion in the
  final model. For example
• Age demographics
• Crime rate
• Economic indicators, such as poverty rate, the
  female unemployment rate, and state government
  general revenue per capita and,
• Education variables.
• None worked well in the presence of other
  predictors.

14
Other Important Variables Not Included

• Also, there were some variables we couldnt test
  due to a lack of data.
• Number of domestic violence occurrences in each
  state. This data is not collected at national
  level and not available.
• The rate of illicit drug use among females. This
  data is not available by state.
• Amount of private investment in prison system.
  We expect higher incarceration rates will lead to
  higher returns on investment, giving private
  prisons a reason to lobby for tougher sentences.

15
Results of the Regression Model
Predictor Variable 13 ß t
? (Constant) 57.0
2.30 .026 People executed per million
population (2000) 11.0
2.25 .030 Child abuse/neglect deaths per
million pop. (1998) 2.15
1.85 .072 Federal spending for social
services per capita (1996)
-.55 -2.37 .022 Percent of people in
families headed by females
living below 125 of poverty
(2000) .53
1.89 .066 Mental health and subst. abuse
serv. availability (1997) 8.18
-2.23 .031 Percent of children living with
grandparents (2000) 3.98
3.40 .001 Indigent defense system (state
level) -9.64 -1.83
.075
16
Results of the Regression

• The predictor variables explain 64.3 of the
  variation in female incarceration rates.
• The estimated coefficients are all statistically
  significant ( gt 90 confidence) with the expected
  signs.
• A decrease of one in Oklahomas female
  incarceration rate equates to approximately 18
  women.

17
Interpreting the Results

• Severity of criminal justice system
• Number of people executed per million population,
  2000 (Stat. Abstract)
• Oklahoma ranks 1 3.2 per million pop. (natl.
  avg. 0.24)
• If Oklahomas severity measure decreased by
  half (to 1.6), we would expect the female
  incarceration rate to decrease by 18 (from 131 to
  113).

18
Interpreting the Results

• Abuse and neglect
• Child abuse/neglect deaths per million
  population, 1998 (Urban Institute data)
• Oklahoma ranks 1 13.04 per million pop. (natl.
  avg. 3.6)
• If Oklahomas rate dropped to the national
  average, we would expect the female incarceration
  rate to decrease by 19.4 (from 131 to 112), all
  else being equal.

19
Interpreting the Results

• Availability of social services
• Federal spending for social services per capita,
  1996 (UI data)
• Total federal spending on child welfare services,
  including funds from IV-B, IV-E, Medicaid, SSBG,
  EA and other sources.
• Oklahoma ranks 38 16.04 per capita (natl.
  avg. 23.52)
• If Oklahomas expenditures increased to the
  national average, we would expect the female
  incarceration rate to decrease by 4 (from 131 to
  127).

20
Interpreting the Results

• Economic Stress
• Of people in families headed by females, percent
  living below 125 of poverty, 2000 (Current
  Population Survey data)
• Oklahoma ranks 5 52.5 (natl. avg. 43.6)
• If Oklahomas rate dropped to the national
  average, we would expect the female incarceration
  rate to decrease by 4.7 (from 131 to 126.3).

21
Interpreting the Results

• Mental Health and Substance Abuse (We used
  expenditures for services because incidence data
  was not available. Increased expenditures should
  reduce these problems, thus decreasing female
  incarceration.)
• Federal expenditures for services per capita in
  1997 (Census data)
• Oklahoma ranks 27 4.89 (natl. avg. 4.99)
• If Oklahomas expenditures increased to 6.39 per
  capita (as in CA), we would expect the female
  incarceration rate to decrease by 12.5 (from 131
  to 118.5).

22
Interpreting the Results

• Unstable family structure
• Percent of children living with grandparents,
  2000 (Census data)
• Oklahoma ranks 18 6.5 (natl. avg. 5.8)
• If Oklahomas rate dropped to the national
  average, we would expect the female incarceration
  rate to decrease by 3 (from 131 to 128).

23
Interpreting the Results

• Effectiveness of indigent defense system
• Is the system funded primarily or totally at the
  state level, as opposed to county-level funding?
  (Spangenberg Group, 1998)
• Oklahoma, KS, OH, SC are mixed states.
• If Oklahomas indigent defense system were funded
  primarily or totally at the state level, we would
  expect the female incarceration rate to decrease
  by 4.8 (from 131 to 127.2).

24
Conclusion

• The regression results suggest that
• The severity of Oklahomas criminal justice
  system, and its mixed approach to indigent
  defense funding, contribute to the female
  incarceration problem.
• The severity of Oklahomas criminal justice
  system is also supported by the fact that
  Oklahoma crime rate is 15th in the nation whereas
  it ranks 1st in female incarceration rate, 4th in
  total incarceration rate, and 5th in male
  incarceration rate in the nation. 14
• Social service expenditures, particularly in the
  areas of mental health and substance abuse, have
  an inverse relationship with state variation in
  female incarceration rates.

25
Limitations of the Data

• All correlations were done with incarceration
  rate data. This may be incorrectly computed due
  to some flaws in the comparability of the data
  among states. Some additional issues to
  consider
• Do States with weak local funding sources have
  higher incarceration rates since local
  incarcerations may not be included in the data?
• County jail inmates are not included in a states
  incarceration rate, although sentenced for a
  comparable amount of time.
• Some crimes punished at the local level are not
  included in the states incarceration rate data.
• Sentencing lengths, percentage of sentence
  served, and time served for comparable crimes
  among states are not taken into account.

26
References And Notes

• 1. Harrison, Paige M. and Allen J. Beck.
  2003. Prisoners in 2002, Bureau of Justice
  Statistics Bulletin, U.S. Dept. of Justice, p. 5.
  http//www.ojp.usdoj.gov/bjs/pub/pdf/p02.pdf
• 2. Scahill, Meghan C. 2000. Female Delinquency
  Cases, 1997, U.S. Dept. of Justice, Office of
  Justice Programs, Office of Juvenile Justice and
  Delinquency Prevention. http//www.ncjrs.org/pdffi
  les1/ojjdp/fs200016.pdf
• 3. Harrison and Beck, p. 5.
• 4. Greenfeld, Lawrence A. and Tracy L. Snell.
  1999. Women Offenders, Bureau of Justice
  Statistics Special Report, U.S. Dept. of Justice,
  p. 1. http//www.ojp.usdoj.gov/bjs/pub/pdf/wo.pdf
• 5. Ibid, p. 6.
• 6. Langan, Patrick A. and David J. Levin. 2002.
  Recidivism of Prisoners in 1994, Bureau of
  Justice Statistics Special Report, U.S. Dept. of
  Justice, Office of Justice Programs, p. 2.
  http//www.ojp.usdoj.gov/bjs/pub/pdf/rpr94.pdf
• 7. Harrison and Beck, p. 4-5.
• 8. http//www.doc.state.ok.us/

27
References And Notes

• 9. Oklahoma Department of Corrections, 2002
  survey (an unpublished document) and Sharp, Susan
  F. and Susan T. Marcus-Mendoza. 2003. Female Drug
  Offenders in Oklahoma, A survey prepared for the
  Task Force on the Incarceration of Women in
  Oklahoma, also see http//www.mapinc.org/drugnews/
  v03/n1057/a06.html?212
• 10. In 1980, according to Census data, 21.2
  percent of children in Oklahoma lived with a
  single parent or other relatives. This number
  increased to 31.6 percent in the year 2000.
• 11. According to the Statistical Abstract of the
  United States (2002, Table 183), 5.1 percent of
  Oklahoma population used any illicit drug in
  1999. Female drug and alcohol abuse in Oklahoma
  is not available.
• 12. Harrison and Beck, p. 5.
• 13. The final slide lists some other variables we
  tested but did not meet the criteria to be
  included in the model.
• 14. Statistical Abstract of the United States,
  2002. Also see http//www.doc.state.ok.us/
• If you have any questions or comments, or would
  like copies of the surveys used in this
  presentation please contact me at
  zohre.salehezadeh_at_okdhs.org

28
Some Additional Variables We Tested

• States' ranking based on AFDC benefit levels in
  1996, 1lowest amount, 5highest amount (Harknet,
  et.al. study, 2003)
• Unemployment rate for females averaged
  93,4,5,7,8,9
• Unemployment rate for males averaged 93,4,5,7,8,9
• Burglary rate -per 100,000 pop.- in 2000 (Stat.
  Abs. 2002, Table 285)
• Motor vehicle theft crime rate -per 100,000 pop.-
  in 2000 (Stat. Abs. 2002, Table 285)
• Divorces per 1,000 population in 1998 (National
  Center for Health Statistics)
• Percent of teens age 16 - 19 who are dropouts
  2000
• Percent of persons age (25 and up) with bachelor
  degree or more
• Elementary and secondary education expenditures
  per capita 98-99
• Female population age 18 - 24 as percent of total
  population
• Female population age 18 - 24 as percent of total
  FEMALE population
• Female pop. age 25 - 44 as percent of total
  population
• Female pop. age 25 - 44 as percent of total
  FEMALE population
• Total federal spending on child welfare services,
  including funds from IV-E, IV-B, Medicaid, SSBG,
  EA, and other sources 1996"