Capital Punishment in the U.S.: Statistical Analysis 1973–1993

Introduction

The intent of this analysis is to evaluate the data provided in the U.S. Department of Justice study titled Capital Punishment in the United States, 1973–1993, with principal investigators from the United States Department of Justice, Bureau of Justice Statistics (U.S. Department of Justice, 2008). The data collected for this study concentrated on prisoners who had been under sentence of death and whose sentences were commuted or pardoned during the period of 1973 through 1993. The study provides a dataset that includes socio-demographic and criminal history data for each respondent, enabling an assessment of whether significant statistical relationships exist that can explain why a sentence was commuted or not. Also included in the dataset are prior felony convictions for criminal homicide and the legal status of respondents at the time of their capital offense. Inmates executed after the initial pardon or commutation of sentence are also included. The dataset is comprised of 5,555 cases.

A longitudinal study of inmates spanning the years 1973 through 1993 forms the basis of the methodology, with the U.S. Department of Justice accumulating socio-demographic data, felony convictions, homicide records, and justice records for each respondent over time. The methodology was designed to ensure the confidentiality of inmates, defining the population as both male and female inmates incarcerated in U.S. prisons during the years of the analysis. The sampling approach was designed to capture the entire sampling frame of inmates who were at one time given death sentences and then either had those sentences commuted or were reassigned to execution. This is a unique dataset in that it captures inmates who received commuted sentences, with a subset of them having a second conviction that ultimately led to execution.

Assessment of Methodology

The dataset is intended to be universally inclusive of the entire population of male and female inmates in the U.S. penal system, yet it may vary from comparable sources in other government agencies for several reasons. First, National Prison Statistics (NPS) captures capital punishment from the last day of the calendar year rather than from the specific date of execution. Second, if an inmate enters prison under a death sentence and then has that sentence commuted, the date recorded reflects the court judgment date rather than the actual date of execution. Third, there is variation in the date when the inmate is formally entered into a state or federal correctional facility.

These three variations in how dates are recorded introduce a high level of imprecision in defining how variables specific to the commutation of sentences varied by date of re-evaluation, date of re-entry into a correctional facility, or the impact of socio-demographic variables on an inmate's ability to gain commutation. This imprecision also makes it difficult to predict statistically why some inmates evade execution while others do not. This is one of the shortcomings of the dataset and reduces its value for in-depth longitudinal analysis. Due to the three different approaches to how the date variable is presented, the reason for an inmate's removal from a death sentence (V31) will serve as the primary variable for this analysis.

The data forming the basis of this analysis is V31 — the reason for an inmate's removal from under sentence of death. Presuming that every inmate seeks to attain this status and that relatives and attorneys would devise strategies and defenses to gain the inmate's freedom from death and eventually from prison, this variable is therefore influenced by all available variables in the dataset.

Variable for Analysis and Rationale

Broader socio-economic and cultural factors have been shown to influence mortality rates of inmates by region of the country (McCann, 2008). The influence of highly conservative areas of the United States has an auto-correlative effect on executions of repeat felons and those with longer histories of criminal convictions (McCann, 2008). There has also been significant analysis of executions in Texas specifically, and this hypothesis has been supported at a statistically significant level (Cunningham & Sorensen, 2007). The region in which an inmate is incarcerated appears to have a greater effect on the likelihood of gaining a commuted sentence than demographics alone — another reason for selecting V31 as the basis of this assessment.

There is also the dimension of how racial variation among inmates underlies discrimination, made more pronounced by the specific region in which inmates are held (Peffley & Hurwitz, 2007). Based on this background research, gaining a commuted sentence has as much to do with an inmate's geographic location as with socio-demographic attributes or the quality of legal defense. As a result, V31 — the reason for an inmate's removal from under sentence of death — makes the most logical sense as the variable to include in this assessment.

Analysis of Data

Because V31 is pivotal to understanding how socio-demographic factors influence whether an inmate receives a commuted sentence, initial measures of central tendency and dispersion for this variable were first examined. The mean value of 4.41, with a standard deviation of 1.512 and a variance of 2.287, indicates significant differences in the factors that lead to an inmate's removal from under sentence of death. Of the total 5,555 cases in the dataset, 51% — or 2,839 cases — met the condition of having their death sentence commuted. More specifically, 5.1% of all inmates in this subset had their sentences commuted outright; 21.6% had their conviction and sentence overturned; 43.1% had their conviction affirmed and their sentence carried out; and 16.9% had their convictions overturned on grounds that their case was declared unconstitutional.

The observation that 43.1% of all inmates had their conviction affirmed and were sent to execution contributes to the wide variation of outcomes. The initial analysis also indicates that constitutional review of cases can add meaningful judgment to individual cases. With 51% of the total inmate population meeting the condition of review, the constitutional evaluation of cases appears to be critically important to sentencing outcomes.

Using cross-tabulations of V31 against socio-demographic variables in the dataset provides further insight into factors contributing to an inmate's removal from a death sentence. For purposes of this analysis, V31 was cross-tabulated with all socio-demographic variables available, including state, gender, ethnicity, Hispanic origin, year of birth, educational level attained, current status, and details of any new sentence. Key findings from these cross-tabulations are described below. First, Texas leads the United States in this dataset both in sentences commuted (30.1% of relevant cases) and in executions (31.4%). This is consistent with prior research on how Texas interprets capital punishment (McCann, 2008; Cunningham & Sorensen, 2007). Among the demographic variables that most contribute to explaining variation in commuted sentences are career criminal history, marital status, education, legal status, and Hispanic origin. When V31 is cross-tabulated with legal status, inmates who were under sentence and on parole were more likely to receive a commuted sentence (7.3%). Additionally, those who had charges dropped were immediately granted a commuted sentence in 64.2% of cases — representing 79 inmates in the sample — illustrating the need for secondary and tertiary screening of death sentences.