Data Sources

The primary data for this study were derived from a sample of 1,009 summary court documents across China. These court cases include major forms of criminal activities (e.g., violent, property, and white-collar offenses), and cover 29 of the 30 provinces in China (data from Tibet were not available). The cases were tried between 1986 and 2001, encompassing three levels of Chinese courts (i.e., district, intermediate, and superior courts).

We drew our sample of 1,009 cases from seven published collections of criminal court cases (a list of these collections is included in Appendix A). Each book describes the process of case selection for inclusion. The language used by the editors of these collections gives some indication of the basis for selection. For example, a regional sample of cases included in the 2000 Shanghai Case Collection is stated to be “carefully selected.” The cases cover “a broad range of areas” and are “typical” of each type of case (Qiao 2000:1). A national sample of criminal cases put together by the Institute for Applied Legal Studies of the Supreme Court was meant to reflect emerging and changing crime patterns and behaviors. This collection tends to focus on serious cases, cases involving individuals of high status, and cases involving complex facts and laws (Cao 2000). By contrast, the collection compiled by the College of National Judge and the Chinese People's University was designed to “provide useful information for legal scholars and practitioners” (Zhu 1999:1). Therefore, maintaining “authenticity and objectivity” were two of the major goals (Zhu 1999:1). Cases were selected particularly for their representation of the time when they were adjudicated, as both crime and definition of crime may be changing over time (Zhu 1999).

As a result of the different bases for inclusion in a particular collection, it is difficult to gauge the overall representativeness of these data to all criminal cases in China. As a group, these national criminal cases cover more serious offenses than is probably true of general criminal offenses in China. This is indicated by the large proportion of cases in the sample that involved serious property loss/damage or physical injury, prison sentences upon conviction, and sentences as severe as “life” and “death” (see Table 1). The description of cases in the collection documents also suggests that these criminal offenses were probably more complicated and of greater national interest than the typical criminal offense.

Table 1. Coding of Variables and Frequency Distribution Before and After Legal Reform

Notes: ns=nonsignificant difference;

^* p<0.10;

^** p<0.05;

^*** p<0.01.

The descriptive documents do not indicate whether or not confession was a basis for case selection in these published collections. This is important because it eliminates a type of sample selection bias as an explanation for possible changes in confession rates over time.Footnote ⁹ We used statistical control through multivariate analyses to evaluate the impact of other sources of sample differences over time (e.g., the inclusion of more serious offenses, offenders with more serious criminal records, or cases with high rates of legal representation in the post-reform sample) on our substantive conclusions about changes in the likelihood of criminal confessions in the post-reform period.

Although the nonrandom inclusion of cases into the various collections restricts the generalizability of our findings, the overall sample of 1,009 cases is sufficiently diverse to provide a preliminary assessment of the extent, nature, and consequences of confessions on case disposition within a changing legal environment in China. It is within this context that we describe our empirical findings as primarily exploratory.

Variable Coding

We call court case documents in this study summary judgments. A presiding judge or a panel of judges prepares these documents after a case is closed. Summary judgments are the official formal record of the court proceedings, and their structure and content are similar to a pre-sentence investigation report (PSI) in U.S. felony cases. These documents include descriptive narrative summaries of the court proceedings and detailed information about the offender (e.g., age, gender, employment status, occupation, education, type of confession given, legal representation, prior criminal record), the offense (e.g., offense severity, co-offending pattern), and case processing outcomes (e.g., pretrial detention, acquittal or convictions, and sentencing decisions). Table 1 presents the coding of the variable and frequency distributions associated with them for each time period.

The primary variable in this study involves the nature of the offender's confession. The measure of confession involves four categories that derive from the depictions of offenders' attitudes and the court's reactions, which are presented in the summary judgment documents. Given that both the offenders' actual admissions of guilt and their attitude (e.g., contrition, sincerity) are legally relevant factors in case disposition, we included this information about confessions explicitly in all 1,009 cases of the summary judgments.

The first category of confession includes offenders who refused to admit any criminal wrongdoings (juburenzui—refusal of admission of guilt). This type of confession is illustrated in the summary judgments by particular wording that states explicitly that the defendant refused to admit guilt and thus deserved more severe punishment. When we treated confession as a simple dummy variable, we used this category to define nonconfessors.

The second category of confession refers to those who admitted to their crime after being arrested, but their admission was interpreted by legal officials as incomplete, nonspontaneous, insincere, or instrumental in nature (e.g., confessing to receive a reduced sentence). This type of confession is called renzui (i.e., “admission of guilt”). A typical narrative statement that reflects this type of confession is “the offender admitted the charges—his attitude was ‘so-so’” (case #90). Consistent with the traditional sentencing practice in China, judges generally make no specific comments about either lenient or severe punishment for these “marginally sincere” confessions.

The third category includes offenders who confessed to their crime after the police discovered their major criminal activities but showed sincere remorse and repentance (tanbai—confession). In some cases, those offenders also apologized to the victim and made restitution. According to the longstanding Chinese legal tradition for treating remorseful confessors with leniency, judges have discretion in granting lenient punishments for these offenders. Some Chinese scholars even argue that laws should explicitly reward those individuals who confess remorsefully, even after their capture, in order to educate the public of the importance of remorse and conformity with the social norms (Reference Li and XieLi & Xie 2000). In coding this type of confession, we looked for the following types of comments in the summary judgments: (1) “the offender's attitude is good” but with no explicit comments of lenient case disposition (case #869), (2) “the offender's attitude is good and deserves lenient punishment” (case #295), and (3) “after her capture, the offender admitted the wrongdoing and showed sincere remorse—she also compensated the victim's loss [and] thus deserves lenient treatment” (case #125).

The fourth category of confession involves offenders who voluntarily turned themselves in before the officials discovered their wrongdoings (zishou—literally, self-submission). These offenders, according to the 1997 CL (Article 67), represent the highest degree of confession and, by law, qualify for a lighter or mitigated sentencing reduction. In these types of cases, the court's language is quite uniform that “the offender voluntarily turned himself in, qualifying for zishou by law, thus should receive reduced sentences” (e.g., case #123).

Other legal variables that are mandated to influence court decisions under Chinese law include offense severity, prior record, and co-offending pattern. We measured offense severity by two dummy variables: (1) whether there was serious physical harm to the victim (coded “yes” or “no”) and (2) whether there was serious damage or loss of property (coded “yes” or “no”).Footnote ¹⁰ The excluded category in this coding scheme represents cases in which there was no serious physical injury or property damage/loss. We also dummy-coded prior record to represent whether the defendant had a prior criminal conviction. We coded co-offending patterns to contrast offenses with and without multiple offenders. Crimes involving multiple offenders are considered more serious in China because of the greater planning and the treatment of these offenses as criminal conspiracies (Reference Zhang and MessnerZhang & Messner 1994).

Extralegal factors involve offender and case attributes that are not legally mandated to influence case disposition. These factors include characteristics of the defendant (e.g., age,Footnote ¹¹ gender, occupation, and residency status) and offense or case attributes (e.g., victim compensation by the offender or the defense attorney used in the case). We measured occupation as an ordinal variable, ranging from the unemployed, manual laborers (e.g., clerks, factory workers, self-employed laborers), and high-status offenders (e.g., managers and governmental officials). Residency status referred to whether an individual was a permanent resident or recent migrant to the area where he/she committed the crime. We also dummy-coded the year of the criminal case to compare types of confessions and case dispositions before and after the 1996 legal reform.

While age, gender, occupation, and legal representation have been regularly used in assessing legal decisions in both Western and Chinese studies (Reference Liu, Zhou, Liska, Messner, Krohn, Zhang and LuLiu et al. 1998; Reference Liu, Zhang and MessnerLiu, Zhang, & Messner 2001; Reference Lu and MietheLu & Miethe 2002; Reference MietheMiethe 1987; Reference NagelNagel 1983; Reference Zhang, Zhou, Messner, Liska, Krohn, Liu and LuZhang et al. 1996), residency status and victim compensation by the offender are unique in China. Because transients are increasingly blamed for urban crimes in China, they are expected to be treated more harshly by the criminal justice system. Victim compensation by the offender signifies reconciliation and restoration in the communitarian cultural tradition. Previous studies have found that compensation made by the offender significantly reduces the severity of punishment (Reference Lu and DrassLu & Drass 2002).

We include several measures of legal process and case disposition in this study as dependent variables. These include separate dummy variables that represent whether the defendant confessed to the crime (confession), received pretrial detention or bail (detention), was acquitted or convicted of the charge (acquittal), or received a prison sentence upon conviction (prison). We measured the length of the sentence on an ordinal scale, including the following categories: (1) less than five years of incarceration, (2) more than five and less than ten years of incarceration, (3) between ten and fifteen years of imprisonment,Footnote ¹² (4) life imprisonment, and (5) the death penalty.

Univariate Frequency Distributions

Our primary interest involves the nature of criminal confessions, and our sample data suggest that the extent and nature of confessions have changed over time. As shown in Table 1, the proportion of Chinese defendants who refused to admit their guilt significantly increased over time (from 21 to 33% of the cases in the post-reform period), and the proportion who confessed but were considered to be insincere in their remorse dropped from 56 to 46%. The other two types of confessions (i.e., confession with remorse and defendants who turned themselves in voluntarily for their crimes) remained relatively stable over time and were relatively rare within this sample of criminal cases.

Concerning criminal case disposition, approximately 90% of the defendants in the sample were detained prior to trial, and this proportion did not significantly change over time. Only about 9% of the defendants in each period were acquitted of all charges. The proportion of convicted defendants given a prison sentence increased from 83% in the pre-reform era to 88% in the post-reform period. The length and severity of the prison sentence for these defendants were significantly reduced after the legal reform in 1996. Half of the convicted defendants received less than five years of imprisonment in the post-reform period, compared to only about one-third of the pre-reform defendants. The proportion of criminal cases that involved a death sentence was more than twice as large in the pre-reform period as in the post-reform period.

Of the offender and case variables, there were significant changes over time in the age and occupation of the defendant, legal representation, seriousness of physical harm, and victim compensation. These results indicate that offenders in the post-reform period were older and more likely to be unemployed and represented by counsel than their counterparts, and the offenses were less likely to involve physical injury or death, than in the pre-reform era. The proportion of criminal cases involving male offenders (90%), local residents (66%), defendants with criminal records (7%), co-defendants (71%), and serious property loss/damage (46%) did not significantly change over time.

Predicting the Likelihood of Confessions

Given that confessions are significantly less common in the post-reform era, an important question involves whether or not these differences persisted after we introduced controls for differences in offender and case attributes across the samples. Table 2 summarizes the results of this multivariate logistic regression analysis.

Table 2. Logistic and Multiple Regression Analysis of Confessions, Pretrial Detention, Acquittal, Prison Sentence, and the Severity of Punishment

^a Notes: This small value represents that all defendants with a prior record in the post-reform period were convicted. d.f. represents the degrees of freedom.

^* p<0.10; ^**p<0.05; ^***p<0.01.

The likelihood of a defendant's confession has decreased in the post-reform period, and this pattern held in both the bivariate and multivariate analyses. When we introduced statistical controls for offender and case attributes, the conditional odds of a confession were 1.6 times less likely after the legal reforms. This finding of a significant net reduction in the odds of confession in the post-reform era provides more direct support for the claim that the procedural reforms in the criminal law were associated with a decreased willingness of criminal defendants to admit their guilt. Among the offender and case attributes, the likelihood of a confession was significantly higher, ceteris paribus, for defendants who were residents of the particular jurisdiction, unemployed, not represented by counsel, and who provided the victim compensation. The finding that legal representation was associated with a net decrease in the probability of confession is consistent with the assertion that access to legal counsel provides criminal defendants some legal protection and safeguard against unbridled pressures toward admissions of guilt and self-incrimination. Inclusion of an interaction term to represent whether legal counsel had context-specific effects (e.g., if counsel was more influential on confession decisions in the post-reform era), however, was nonsignificant, indicating that the legal counsel had a similar effect on reducing the likelihood of confession in both the pre- and post-reform periods.

The Impact of Confession on Criminal Case Disposition

We examined the impact of confession on criminal processing decisions by comparing nonconfessors and various types of confessions. We conducted both bivariate and multivariate analyses to explore these relationships.

About 90% of the defendants in this sample were detained pending trial, and there were no significant differences between confessors (91%) and nonconfessors (90%). Among the types of confessors, however, the proportion of defendants detained before trial was lower for those who showed remorse (83%) or voluntarily confessed to undetected criminal activity (84%) than for those whose admissions were considered less than sincere (94%).

Nonconfessors had a higher rate of acquittal on all criminal charges than confessors with insufficient sincerity (14% vs. 5%), but acquittal rates for other types of confessors were similar to those for nonconfessors.Footnote ¹³

An examination of the bivariate relations revealed a significant difference between nonconfessors and confessors in the likelihood of receiving a prison sentence upon conviction. While about 95% of nonconfessors received a prison sentence upon conviction, only about 81% of confessors received a prison sentence. Among the different types of confessors, the likelihood of a prison sentence was lowest among defendants who turned themselves in (69%) or confessed with remorse (73%), whereas imprisonment risks were much higher (85%) among confessors whose sincerity was challenged by court officials. Nonconfessors were also given more severe prison sentences and punishments than confessors. For example, nearly one-third of convicted offenders who did not confess were given a death sentence, compared to less than one-fifth of the confessors. Defendants who admitted guilt without sufficient remorse were given the most severe punishments among the different types of confessors.

Some of the bivariate relationships between confession and case disposition persisted, while others changed after we introduced controls for other offender and case attributes through multivariate analyses. As shown in Table 2, logistic regression analyses revealed that confession had no significant net impact on the likelihood of pretrial detention and acquittal/conviction decisions. Similar to the bivariate results, confessors were less likely to receive a prison sentence and were given significantly shorter and less severe prison sentences than nonconfessors after controlling for other variables.

Aside from the impact of confessions, criminal case dispositions in this sample were also influenced by several offender and case attributes. For example, the conditional odds of pretrial detention, ceteris paribus, were significantly higher among migrants, the unemployed and less-skilled workers, and offenses involving serious property loss/damage. The net likelihood of receiving an acquittal of all charges was significantly higher among older defendants, persons not represented by counsel, single offenders, those charged with less serious property offenses, defendants who compensated victims, and individuals who were not detained before trial. The conditional odds of a prison sentence were higher among male offenders, migrants, unemployed and less-skilled workers, offenders with prior records, serious property offenses, defendants who did not compensate victims, and those defendants who were detained prior to trial. Finally, the multiple regression analysis indicated that more severe punishments were given to female defendants, migrants, those of higher occupational status and prior records, crime situations involving multiple offenders, serious physical and property crimes, and noncompensators of crime victims.

Context-Specific Effects of Confession on Case Disposition

Although the legal stipulation of sentencing reductions for voluntary confessors has remained the same over the last two decades, judicial discretion in criminal processing decisions may still have been affected by other changes in legal structure and legal culture after the 1996 legal reforms.

To explore more fully whether the impact of confession on case disposition is context-specific (i.e., dependent upon time period), we entered interaction terms into the respective logistic regression and multiple regression models to determine whether these outcomes were significantly different after the legal reforms. As shown in Table 2, these interaction effects were not significant for all of the dependent variables, suggesting that the net impact of confession on case disposition did not change appreciably after the reform effort. Overall, the multivariate analysis indicated that differences between confessors and nonconfessors in case outcomes are similar in both the pre- and post-legal reform periods in China.

Explaining the Decline in Criminal Confessions

The major economic reforms in the 1980s had a profound impact on Chinese society. They introduced a Western value system that seriously challenged traditional communitarian values and cultural beliefs. The common Chinese citizen experienced a personal transformation from a contextual actor (i.e., a person whose identity is linked to a group or relational context) to a self-serving individual actor. These social forces, in turn, precipitated declines in informal social control that have been linked to unprecedented increases in crime rates and strong public outcry for more punitive measures toward offenders. Subsequent changes in legal structure (e.g., increased legal counsel, more adversarial proceedings) and legal culture (e.g., greater use of law for dispute resolution, more formal challenges to state authority) occurred following the initial economic reforms.

Although it is relatively simple to document these social and legal changes in China, it is far more difficult to identify the major contributor to the decline in criminal confessions over time. Possible explanations for decreases in confessions in the post-reform period include (1) declining communitarian values, (2) changing legal structure, (3) changing legal culture, and (4) basic changes in cost/benefit evaluations of confession by criminal defendants. We summarize the strengths and limitations of each explanation below.

Any viable explanation for declining confession rates in China's post-reform period must begin with the decline in the communitarian value system. Communitarian societies are linked to particular functional imperatives (e.g., mutual obligations, interdependency, and group interests superseding individual interests) that promote confession as morally appropriate behavior necessary for the restoration and reintegration of the “whole person” back into his/her social group. While communitarianism remains the dominant value system in China, individualism has continued to grow in China over the post-reform period.

Rising self interest under declining communitarianism, however, cannot itself explain the decline in confessions unless concurrent changes have also occurred in legal structures. This is the case because elements of the prevailing legal structures (e.g., access to legal counsel and other resources, individual rights) place serious constraints on defendants' opportunities to confess. Without a facilitating legal structure, a defendant's desire to fight the allegations through nonconfession is neither a practical nor a feasible option.

Various changes in the legal structure have occurred in the post-legal reform period, but the most visible change involves the greater access to legal counsel. In fact, the broadened rights to legal representation are the major source of change in China's legal structure because other changes in procedural law (e.g., the movement away from an inquisitorial system, more adjudicative functioning by judges, increased defendants' rights to cross-examining witnesses, and pretrial release) are directly tied to access to defense counsel. Aside from providing basic legal protection for their clients, legal counsel may enhance defendants' awareness of the consequences of their admissions of guilt, thereby reducing their likelihood of confession. The presence of legal counsel may also increase the chances that sentencing concessions will be granted if defendants do confess. However, the rise in legal representation is not a sufficient explanation for the declining rates of confession because we found lower risks of confession in the post-reform era even after we included statistical controls for differences in legal representation across pre- and post-reform samples. This latter finding suggests that it is more than just the availability of defense counsel that precipitated the decline in criminal confessions.

Changes in legal culture that may explain the post-reform decline in criminal confessions involve the consequences of increased acceptance and use of litigation. As mentioned earlier, there has been a dramatic increase in civil litigation in China over the last two decades. One possible consequence of this increased use of law to resolve civil disputes is greater public awareness of individual rights and the subsequent emergence of a cultural climate in which nonconfession is becoming a more socially acceptable challenge to the state's authority. However, a legal culture conducive to nonconfession behaviors also requires a facilitating legal structure that provides defendants the physical opportunity to challenge the longstanding tradition of confession in China's legal history.

Another plausible explanation for defendants' greater reluctance to confess in the post-legal reform era perhaps lies in pragmatic concerns about the lack of certainty of dispositional decisions and the harsh punishments defendants are likely to receive after they have confessed. These concerns about the “costs” and “benefits” of confession fit into a more general rational choice framework that views legal actors as rational decision makers (see Reference RamseyerRamseyer 1985, Reference Ramseyer1988).

From this rational choice perspective, confession decisions reflect a rational calculation of the relative risks and benefits of confession. Potential benefits of confession for the defendant include the possibility of a reduced sentence for confession and positive appraisals of his/her moral worth by following the traditional cultural imperative that associates confession with “good citizenry” and as being necessary for the moral reform of the offender. The primary costs of confession involve the greater certainty of a criminal conviction and a virtual guarantee of severe punishment even if the defendant confesses when the initial charge is serious.

Several aspects of the Chinese legal system, however, make these cost/benefit calculations more difficult than in other contexts. For example, plea agreements in the United States are a type of confession that clearly specifies a particular concession for the guilty plea (e.g., a charge reduction, dismissal of other counts, probation rather than jail time, concurrent rather than consecutive sentences). In China, however, confession in most cases does not guarantee a concession. Instead, the criminal code and the penal policy in China only specify that persons who confess with sincere remorse or before the police detect the crime will be given a more lenient sentence (but the degree or amount of leniency is not defined). In addition, the code and policy allow judges to easily circumvent provisions for lenient treatment through their official interpretations of the sincerity of the defendant's confession. Chinese defendants are also disadvantaged in these risk/benefit calculations because of (1) an enormous variability in sentencing for most offenses,Footnote ¹⁴ (2) less public awareness of the “going rate” for particular crimes and greater uncertainty that these “standard” sentences would be applied even if known, (3) high prospects for severe punishments for many defendants even if they confess,Footnote ¹⁵ (4) less access to defense counsel for protecting defendants' rights, and (5) less advocacy by those who have legal representation than is true of most Western developed countries. Sound cost/benefit calculations under these conditions of uncertainty are extremely difficult.

From the perspective of the rational legal actor, the high rate of confession in both the pre- and post-reform periods must be attributable to an excess of perceived benefits, whereas the post-reform decline in confessions must be due to a change in the nature of these cost/benefit ratios. Because these cost/benefit calculations should derive from rational defendants' assessments of the prevailing social and legal conditions, this explanation for the post-reform decline in criminal confessions is also ultimately tied to the nature of legal structure, legal culture, and the wider social context in which these decisions are made.

Collectively, these interpretations suggest that the post-legal reform decline in confessions in China is not the direct consequence of any singular factor. The additive combination of all of these social forces is also not an adequate explanation for this finding because even though confessions have declined over time, they still remain the dominant trend in about two-thirds of the criminal cases in the post-reform sample. Instead, it is the particular conjunction of a changing value system and legal cultures, legal structures, and modifications in the subjective calculations of a cost-benefit ratio that best account for the observed decline in criminal confessions in the post-legal reform period.

For those concerned with the general study of law and society, these explanations for declining confessions in China's post-legal reform period suggest that this basic relationship between social and legal change is both complex and dynamic. It is not simply an additive combination of particular events in isolation. In fact, even when the temporal ordering between social and legal change appears to be relatively straightforward, the lingering effect of cultural traditions similar to China's communitarian value system may continue to influence legal decisions, even within a rapidly evolving sociolegal context. An immediate question for further research that derives from the current study is whether declines in criminal confessions will become more apparent in China as its communitarian foundation and legal structure continue to change over time.

The Consequence of Confession on Case Outcome

It is widely known that sanctions rendered by the criminal justice system can be very severe in China (Reference CaiCai 1997). Consistent with the literature, the current study reveals that most offenders were detained during investigation and trial (90%), were convicted (91%) and given a prison sentence (84%), and received lengthy imprisonment terms (approximately 64% of offenders who received a prison sentence served longer than five years). Once controls were introduced for offender and offense characteristics, whether or not a defendant confessed had no significant impact on the likelihood of pretrial detention or conviction decisions. However, confession had a significant net impact on reducing the punishment for convicted offenders. Confessors were significantly less likely to get a prison sentence and were given less harsh punishment if they received a prison term.

While confession was associated with less severe punishment in both the pre- and post-reform periods, the fact remains that Chinese defendants still face harsh punishments even when they do confess. For example, 85% of the confessors in the post-legal reform period were still given a prison sentence upon conviction, and nearly one-third of them received punishments more severe than ten years in prison. In considering that confessions had similar effects on sentencing decisions over time (see Table 2), these findings suggest that sociolegal changes in the reform period did not “spill over” and differentially affected the consequences of confessions. Regardless of the time period, it is clear from these findings that Chinese defendants in this sample did not reap the same type of sentence concessions for their admissions of guilt as defendants in criminal court proceedings of other countries.

From the Western perspective on plea bargaining, it is an interesting and relatively unique dynamic that confession in China's post-legal reform period remains the dominant “choice,” and criminal defendants receive comparatively little benefit in sentencing concessions for their admissions. However, this particular pattern appears to be explained by the continued restrictions on individual rights,Footnote ¹⁶ the steadfast faith by lawmakers of the probative value of confession, and/or China's communitarian tradition that places more importance on the moral value of confession than its subjective utility to the offender as a means of receiving a more lenient punishment.

Implications for Comparative Research

A comparison of confessions and their impact on case disposition between the Chinese and the Japanese system reveals that an overwhelming majority of criminal defendants confess to legal authority about their criminal involvement in both countries (Reference Ramseyer and RasmusenRamseyer & Rasmusen 2001). As discussed previously, the strong inclination toward submission to authorities by individuals in these two societies may lie in the Confucian cultural influence that stressed a group-oriented, hierarchically, and morally ordered society. These essentially communitarian societies have produced a different system of confession from that practiced in an individualistic society. While plea bargaining in the United States is functionally equivalent to the system of confession in Japan and China in that they both help reduce the severity of legal sanctions and save resources for the criminal justice system, plea bargaining is fundamentally different from confession. Plea bargaining focuses on the probative value of confessions. In contrast, the corrective value of confessions is more emphasized in both of these communitarian societies.

While the Japanese society has remained true to its core belief and values about communitarianism, the changing economic, social, and legal arrangements in China, accompanied with greater changes in the culture system, signify a major departure from its communitarian roots to a more individualistic society. As we have suggested earlier, these structural and cultural changes may ultimately underlie the reduced confession rate in China in recent years. Major changes in Japan's confession rates may not have occurred because similar changes in the legal structure and its culture have not taken place (Reference Sanders, Jacob, Blankenburg, Kritzer, Provine and SandersSanders 1996).

Although it is both structurally and culturally expected that confessions receive leniency in China and Japan, the degree of leniency does not appear to be comparable in the two systems. Japanese criminal defendants receive much more lenient sanctions than their Chinese counterparts. Compared to our Chinese sample, Japanese defendants are more likely to be convicted, but the overwhelming majority of those convicted receive a suspended sentence. Among those who are convicted and receive a nonsuspended sentence, the majority of sentences are less than five years (Reference BayleyBayley 1991; Reference ReichelReichel 2002).

In addition to the major differences in the changing nature of confession and the legal responses to confessions between the two systems, there are some notable differences with regard to how confessions are conducted in these two societies. For example, compensation is “ordinarily done” in Japan (Reference HaleyHaley 1995:130), but only a small proportion of our Chinese offenders (31%) compensated the victim. Japanese offenders are also expected to apologize to the victim and seek their forgiveness, but no such expectation exists in China. Of all the 1,009 cases, in only four cases did the court comment that defendants made compensation and apologized to the victims and their family. Furthermore, victims and/or their family members play an important role in the criminal justice proceeding with letters of forgiveness in Japan (Reference HaleyHaley 1995). This contrasts with little victim involvement in the Chinese legal system. In fact, there was no mention of the victim's views about the defendant and/or the offense in any of the official Chinese court records examined in this study.

The differences in the practice of confession between the two societies may lie in the Chinese socialist legal system, which departs from the communitarian legal culture. While the communitarian society emphasizes the mediation of the relationship between individuals, a socialist system tends to focus on the maintenance of state and individual relations (Reference Sanders and HamiltonSanders & Hamilton 1992). For this reason, an apology to the victim symbolizes the restoration of individual relationships in a communitarian context, whereas admission of guilt to legal authorities represents the restored order between the state and the individual in socialist China. In either context, however, confession plays an important role in case disposition. How rates of confession will respond to further sociolegal changes in both societies remains an interesting question for future research.