Learning to Sentence: An Empirical Study Of Judicial Attitudes Towards Judicial Training In Romania

This paper aims to present an innovative type of empirical research in judicial studies. It focuses on assessing the attitudes and perceptions judges have towards the way they learn how to sentence throughout the course of their judicial careers. The first main assumption is that judges learn all throughout their lives, both through the formal training offered and other informal sources of learning (such as the practice of sentencing, or peer advice). The second main assumption is that the judges’ learning needs and perceptions towards training change as they gain more experience and they are exposed to various learning contexts. So far the study was conducted in one European jurisdiction (Romania), totalling 510 judicial respondents ranging from 0 to 40 years judicial experience. This paper presents the findings with regards to their views on the judicial training they receive.


Introduction
The study of judicial decision-making has its origins in the early 20 th century American scholarship in political science and law.The judicial studies movement was a reaction against the traditional legal model of judicial decision-making, which saw law and legal rules as the only relevant factors in judicial decision-making.As judicial studies developed, new models of decision-making emerged.Initially the attitudinal model suggested that judicial decisions simply embodied judges' personal attitudes.Later the strategic and institutionalist models acknowledged that judges act in more complex ways, and that other judges and institutional rules can affect and constrain judges' decisions.But all of these models maintain that judicial decision-making is only about achieving judges' policy preferences.In recent decades scholars have begun to acknowledge that the leading models in judicial studies are too narrow, 2 and a more complex and holistic approach to understanding judicial decision-making is needed. 3is study seeks to add an extra dimension to the empirical study of judicial decision-making by looking at how judges learn to make decisions in court.By looking at the components of the learning process, this study adds a developmental dimension to the study of judicial decision-making -because it hypothesizes that decision-making is not a static, unchangeable exercise, but evolves with time and experience.This aspect was inadvertently captured in some past empirical studies which also measured age or years in post, 4 but that did not seem to prevent researchers from drawing 1 Diana Richards is PhD Candidate at University College London.Her research is currently focusing on assessment of judicial attitudes towards formal judicial training.She can be contacted by: diana.richards@ucl.ac.uk.Address: UCL Laws, Bentham House, 4-8 Endsleigh Gardens, London WC1H 0EG, UK, telephone: +40 7551570859.-The author is incredibly grateful to the Romanian National Institute of Magistracy and to the Romanian judiciary for agreeing to contribute to this study.Additional thanks go to Professor Cheryl Thomas, UCL Judicial Institute, Dr Nigel Balmer, UCL Centre for Legal Empirical Research, and the UCL Laws faculty for the significant support and mentoring in designing and conducting this research.Finally, the author is very grateful to Dr Markus Zimmer and Prof Philip Langbroek, as well as to the two anonymous reviewers for providing insightful feedback on the piece. 2 Lawrence Baum, Judges and Their Audiences: A Perspective on Judicial Behavior (Princeton University Press 2009). 3Jennifer K Robbennolt, Robert MacCoun and John M Darley, 'Multiple Constraint Satisfaction in Judging' (Social Science Research Network 2008) SSRN Scholarly Paper ID 1133184 http://papers.ssrn.com/abstract=1133184accessed 23 September 2014. 4Sheldon Goldman, 'Voting Behavior on the United States Courts of Appeals, 1961-1964' [1966] The American Political Science Review 374; S Sidney Ulmer, 'Social Background as an Indicator to the Votes of Supreme Court Justices in Criminal Cases: 1947-1956  Terms' (1973) 17 American Journal of Political Science 622; Orley Ashenfelter, Theodore Eisenberg and Stewart J Schwab, 'Politics and the Judiciary: The Influence of Judicial Background on Case Outcomes ' (1995) 24 Journal of Legal Studies 257; Kenneth L Manning, Bruce A Carroll and Robert A Carp, 'Does Age Matter?Judicial Decision Making in Age Discrimination Cases* ' (2004) 85 generalised conclusions about factors that influence judicial decision making, without acknowledging that those factors might change with experience and with exposure to learning.
When thinking of the learning process judges have to go through, the first and foremost thing that comes to mind is the specific formal training judges have to undergo by virtue of being judges, a type of vocational training dedicated to adults.In most jurisdictions, newly-appointed judges have to undergo initial or induction training, while judicial training institutions also offer continuous training throughout the judges' careers.Initial training is the training typically offered in continental civil law systems where the judicial trainees have a law degree (are professional judges, as opposed to lay judges) but are not required to have significant previous practical legal experience at the time of recruitment.For this reason, the initial training lasts longer (one or two years) and is meant to prepare the trainee for the judicial office by compensating for the lack of experience with a more intensive and longer theoretical and practical training.In contrast, induction training is available in jurisdictions (such as the common law jurisdictions) where judges are recruited among legal professionals who have already gained significant practical legal experience.Induction training is typically shorter than initial training.Because initial and induction training refer to different types of candidate profiles, a judge typically undergoes only one of these forms of training, never both. 5In addition, it is typically the case that judges are encouraged (or required) to update their skills and knowledge by undergoing continuous training offered by the judicial training institutes, which normally consists of subject-specific short sessions, organised either centrally or regionally.Continuous training anticipates that judges are expected to sign up for sessions regularly throughout their career, based on the assumption that one is never too old to learn something new but also on the assumption that, by keeping their know-how up to date, judges maintain their independence and relevance in society while making tough decisions. 6 look at how judges are trained when they first begin their job, and to see if the training is indeed helping them learn how to do their job in court, is a key component in understanding how judges learn to judge -and the current study includes an attitudinal study of initial judicial training needs and expectations.But judges, like any other professionals, learn all throughout their professional lives.This phenomenon was captured by theorists of education when speaking of experiential education and lifelong learning, 7 as well as by educational and governmental institutions when defining different types of learning. 8This, in turn, has three consequences for the current study, out of which the first two are key for the current piece.First, it made sense to not just look at initial judicial training but at various forms of continuous training as well, be it face-to-face or online, and to see if judicial experience correlates with a different perspective on the usefulness of various types of teaching methods or legal content.In that sense, it was expected that the continuous training vs. the initial training serve different purposes and have to employ different methods to address needs of professionals at different levels of their careers.This is already assumed by judicial training institutes when they design their sessions and pick their methods, but so far their intuitions were never externally validated by measuring what the participants actually expect.For this reason, the research design included respondents from all types of training (initial, induction, continuous) and also observed the entire variety of methods and content offered in all types of training.
Second, in having experiential learning as its fundamental assumption, the study had to take the developmental thesis a step further.It is not just that the initial training has to differ from continuous training in its deliverables, because it serves learners with no or with some prior experience; in fact it has to measure the degrees of experience and see, for instance, if judges who undertake continuous training at different levels of their career might still have significantly different needs and expectations from one another, even if they are included in the same category of learners. 9Because of this aspect, Social Science Quarterly 1; Lee Epstein, Andrew D Martin and Andrew W Martin, 'Does Age (Really) Matter?A Response to Manning,  Carroll, and Carp' (2004) 85 Social Science Quarterly 19. 5 That does not mean that a system cannot have both types of recruitment and therefore both types of training, which will become apparent soon. 6"To do their work well and meet society's expectations, courts need to have judges with the highest ethical standards and extensive legal and societal knowledge.As a result, judges should continuously enrich their knowledge, maintain their skills and acquire new ones.It would appear that the need for judicial training is nowadays greater than ever -increasingly complex and sensitive issues arise in litigation processes, which judges need to master fully.In this sense, the variable of prior experience needed to be not just binary but continuous.
two key variables were taken particularly seriously -the type and amount of experience gained, on one hand, and the type and amount of prior exposure to training, on the other -to determine whether they explain differences in perceptions towards all the forms of formal learning just mentioned.
The third consequence of a developmental approach to judicial learning that had impact on the manner in which this study was designed, but unfortunately cannot be treated in this article, was that it broadened the scope of "learning" to mean more than just "training".While formal training is an important component in helping judges learn how to make decisions, one could hypothesise that there are other instances in which judges learn how to do their job, either by actually having to make decisions in concrete cases (learning by doing), by using a variety of tools in their practice (legislation, precedential outcomes in similar cases, guidelines, reports, recommendations) or by asking their peers for advice in difficult situations (peer learning).The findings pertaining to the informal aspects of learning will be treated in a future article for the sake of clarity.
While it has a dynamic and innovative conception of what counts as learning, and therefore a rather ambitious aim in terms of the variables it aimed to measure, this study had to narrow down its scope in two ways so as to become feasible.First, it decided to look at one type of judicial decision-making rather than try to muddle up very different types of judicial tasks.It chose sentencing as one of the key, incredibly difficult, tasks judges in criminal courts have to perform.On one hand, because there is purportedly a higher pressure to get things right in a criminal case than in other types of cases.On the other hand, because in sentencing the judge has to learn how to balance a myriad of aggravating and mitigating factors, as opposed to other kinds of decisions where the decision might include fewer elements.Some very insightful studies on sentencing were conducted so far, but they were conducted on small samples and/or they did not take the developmental, experiential element into account and therefore just offered a "snapshot". 10cond, although an ideal study would be conducted in all world jurisdictions in order to be able to offer a more generalizable picture of judicial decision making, and perhaps even part of a longitudinal study initiative that would also account for time changes, the current study needed to start with one jurisdiction, with the hope that the following years will give opportunities for expanding the study to other jurisdictions.The jurisdiction studied is Romania.Why Romania?Romania represents an interesting case study for several, perhaps unexpected, reasons.
First, it has implemented a hybrid judicial appointment and training model during the last decade.It did not just focus on appointment of young, inexperienced judges, as civil law systems typically do.It also ran a parallel appointment and training route for experienced lawyers, akin to what common law jurisdictions run around the world.From the standpoint of comparative empirical legal research, this was a unique opportunity to study two categories of judges who have the same cultural and political background, but have very different life experience and working knowledge of the law.In other words, by studying the Romanian hybrid judicial system, one manages to compare the differences between common and civil law jurisdictions without having to worry, as in other comparative approaches, that there are confounding factors such as different political systems, or variable socio-economic factors, or different legal educational backgrounds between the respondents. 11Moreover, the two different subsets of judges -the young, inexperienced judges from the "NIM route" 12 (comparable to judges from other continental European jurisdictions) and judges with prior advocacy experience from the "direct route" (comparable to new appointees from common law jurisdictions) -received training from the same trainers, on the same legislation and legal procedure, under the same educational vision of the leadership of the NIM.With so many societal and environmental variables controlled, one might perhaps draw the safe conclusion that the differences in attitudes between the two types of Romanian judges reflect differences in their advocacy and life experience in general.
Second, although rather recently turned into a democracy with independent judicial appointment and training bodies, Romania had benefited during the last decade from visionary and proactive young managers both in the Superior Council  (Prison Reform Trust 2007). 11 In fact, the age difference between "NIM route" and "direct route" judges is small enough to warrant that both types of judges were raised post-1989, therefore they experienced the same post-communist transitional democracy. 12The NIM represents the Romanian National Institute of Magistracy, the institution in charge with the appointment and training of judges and prosecutors in Romania.In Romania, the "NIM route" represents the judicial appointment and training route whereby young law graduates with no significant practical experience are recruited, then trained for up to 2 years at the Institute.
of Magistracy and the National Institute of Magistracy.This translated into the implementation of the best appointment and training practices for the Romanian judiciary.This effort has been recently praised by the European Judicial Training Network (EJTN) in a report approved by the European Commission on the best training practices for judges and prosecutors throughout Europe.The EJTN has praised the Romanian NIM for good practices in four key areas: 1) the training needs assessment, 2) in providing a comprehensive package to deliver large-scale training on new legal instruments, 3) the recruitment and evaluation of judicial trainers, as well as 4) making training content available online. 13ile Romania is, both theoretically and practically, an interesting and timely case study for empirical research assessing judicial attitudes, it nevertheless is just a first, and hopefully not last, case study for the current methodology.The methodological framework of this study can be applied in any jurisdiction of the world -not just in a European, continental, or post-communist jurisdiction.The main reason is that the current research methodology is not jurisdiction-bound; it was designed to apply to any jurisdiction as long as judicial training is offered within that jurisdiction.The real challenge would not be in adapting the research instrument -but in making the correct interpretations as to what causes differences across jurisdictions, as potential confounding factors must be controlled for or at least identified.This issue needs to be addressed in future research.
There is one last important aspect of the research design that can initially puzzle a reader belonging to common law jurisdiction with adversarial practices, but it is very important in conducting this kind of research in inquisitorial systems and hopefully its value will become apparent.In inquisitorial systems, judges and prosecutors collaborate on discovering the truth in the case -and for that reason their rapport with one another is collegial rather than that between an umpire and a player with competing interests.In many inquisitorial systems around the world, it is considered not only normal, but also welcome, that judges and prosecutors are recruited and trained together.The birth of judicial studies took place in the United States where the adversarial nature of the justice system draws a stark distinction between the roles of judges and of prosecutors.For this reason, judicial studies are seen, by definition, as only including judges.Yet transplanting the judicial studies methodology into Europe has to take into account the specificities of many European judiciaries in order to be an accurate description of reality.While on the issue of who makes the final decision in a case, it is still the judge who has the final word, on the issue of judicial training, measuring the prosecutors' attitudes towards the learning process at the same time as the judges' gives an important vantage point from which to see if the particularities of the judicial role correlates with a significant change in attitude.For this reason, prosecutors were included as a secondary sample in this study.As it will be explained in the research design, prosecutors undergo the same selection process and get the same training as their fellow judges.While they do not constitute the main focus of this study, their insights are valuable when they differ significantly from their fellow judges, and that is an added, if not unexpected, strength of this study.
The following sections describe the research questions, the stratified sampling and the methodology employed in the study on the attitudes and expectations Romanian judges have towards the judicial training offered by the National Institute of Magistracy, while flagging up interesting differences in attitudes correlated with their various levels of experience.

The Research Design
With regard to formal training, the study was designed to measure whether judges14 with 1) different levels of experience and/or with 2) different levels of exposure to judicial training regard the various aspects of their formal judicial training differently.They were asked about their: 1. perceptions with regards to their (past) initial training, its adequacy as preparation for judicial service and its best training methods; 2. perceptions on continuous training, its overall usefulness and best methods; 3. reflections on the overall usefulness of (past) university law training in the light of current experience; 4. the perceived usefulness and feasibility of judicial skills training (judge craft); and 5. preference, if any, for different e-learning methods.
The idea behind the questions asked was, on one hand, to get a sense of the importance that various types of judges assign to the judicial training for their professional development, on the other hand, to identify more specifically those methods and content choices that judges, from their own experience, consider most fruitful in their learning process.Of course, as mentioned before, the main hypothesis of the study was that the respondents' perceptions of the 5 themes above would be explainable through different levels of judicial experience as well as different exposures to training.In this sense, experience and exposure to training are the two main input variables of the study.
Sampling.Different jurisdictions offer different combinations of samples on these two input variables.As explained in the introduction, in common law systems, newly-appointed judges typically have significant legal expertise (typically as advocates, but not only) and have had time to observe sentencing in court, although they lack judicial experience.Moreover, in common law jurisdictions, training is not necessarily mandatory and it is not part of the recruitment process.In contrast, in civil law systems the newly-appointed judges do not need to have any practical legal experience, but they are required as part of the appointment process to go through initial training, which often includes supervised practice in court proceedings.The paradigms just described are extremes on a continuum and, of course, every world jurisdiction might borrow elements from both models, as it will be visible in the case study below.Indeed it has been noted that "many if not most European judiciaries now appoint at least some experienced professionals to the judiciary later in their careers, and their initial training needs are therefore similar to new appointees in common law systems.In addition, common law judiciaries are increasingly becoming 'career' judiciaries in which more appointments are being made from among younger, less experienced lawyers and where progress to higher judicial posts is not just possible but encouraged."15These differences between judicial systems are important to note here because they reveal that the variable of "experience" can be understood in at least three different ways, so it needed to be further refined in the research design: 1) prior judicial experience (experience on the bench), 2) prior legal experience (experience gained by practicing law) and 3) prior experience in criminal cases (since respondents are asked about their sentencing practice and they might have not spent their entire judicial career judging criminal law cases).Respondents were thus asked, where applicable, about their judicial, legal and sentencing experience.
Romania is a European, continental law jurisdiction, a rather typical example for Central and East European continental countries.Among a total population of 20 million, it currently has 6,949 judges and prosecutors, collectively called 'magistrates' and considered part of the judiciary.Judges and prosecutors undergo the same state exam in entering the profession and are trained together throughout their career.As described earlier in the introduction, this explains why, although the current study was focused on the professional training of judges, prosecutors were also included as a secondary sample, for both theoretical and practical reasons.The main practical reason is that prosecutors were present at all sessions where the survey was applied, so it would have been a mark of disrespect not to provide them with a copy of the survey while their colleague judges were filling in the questionnaire.The main theoretical reason is, as explained in the introduction, that prosecutors as a separate sample provide an interesting vantage point on the judicial training -one can now try to distinguish if specific judicial attitudes towards training could be due to the specific nature of activity as a judge or not.
In Romania, the study was meant to cover judges (and prosecutors) with various levels of judicial experience and having undergone all three possible types of judicial training -initial, induction and continuous training.The main phase consisted in a pen-and-paper large-scale survey containing multiple choice, rating and ranking questions, designed to be filled in by judges during their training sessions (10 minutes maximum).Three versions of the questionnaire were drafted to reflect the three main samples above.In addition, the prosecutors undergoing continuous training were given slightly adapted version of the questionnaire, although most variables and questions remained the same so they could be compared as a cross-sample.In total, each questionnaire had 20-22 questions and provided on average data about 113 variables.The main phase took place between February and April 2014.Various techniques for maximising the response rate were used, ensuring an overall response rate of 70% (510 surveys received).Out of all the questionnaires, 92% (470 surveys) were considered valid and ended up being used in the data analysis.
Data gathered.The following table summarises the data gathered in the first jurisdiction:  had not yet been assigned judicial and prosecutorial functions during their first year.They were nevertheless asked about their career option and that option was taken into account as a proxy for their likely profession.This split was used in all subsequent questions to verify if there are any statistically significant differences between judges and prosecutors in their perceptions.
While the main sample remains the judges' sample, it was sometimes useful to look at how prosecutors responded, and whenever there are interesting differences they are mentioned specifically.188) have no experience in post (the initial and the induction trainees), while the rest (60%) have between one year and 35 years in the post, with 11.5 years of judicial experience on average (the mean was unbiased by the inexperienced ones).On average, judges have 12 years of judicial experience.
Since initial and induction trainees lacked judicial experience when surveyed, they were further asked about their legal experience prior to joining the judiciary in order to see if that variable makes a significant difference in their behaviour.All induction judges had prior legal experience (ranging from one to 17 years, with seven years on average), while only 23% initial trainees had prior legal experience (ranging from six months to eight years, with two-thirds having three years or less).The difference in legal experience between the two groups is, perhaps unsurprisingly, statistically significant.Induction trainees must have at least five years of experience when applying for judicial positions, while initial trainees must not.Those initial trainees who had prior legal experience (excluding internships) earned it either as police officers (47%), trainee lawyers or legal advisors (53%).In contrast, more than half of judges undergoing induction training have a police background (54%) while 39% have advocacy experience.As you can notice, advocacy and policing are the two main prior professions for judicial trainees in Romania.
Out of 446 respondents, 30% (134) had no prior criminal legal experience. 16The prior criminal legal experience of respondents ranges from six months to 40 years, with 10 years of experience on average.Averages vary across categories: while experienced judges are very similar to experienced prosecutors (mean 11 years of criminal legal experience), induction judges are less experienced (mean 6 years) and initial trainees even less so (mean two years).In what follows a selection of findings pertaining to judicial training will be presented.

The Findings: Formal Training Perceptions and Needs
This part of the study aims to gauge the importance that different types of judicial training have/had on the subsequent career of judges and prosecutors.It contains both general questions reviewing the overall perceived usefulness of training and more specific questions concerning which training methods are fit for purpose.Below, both types of findings are presented; but apart from the descriptive results, the interesting aspects lie in the correlations found between these training perceptions and needs and other variables such as experience, role or background.The most important finding here can be summarised as "appreciation comes with experience".Namely, experience seems to generally have a positive effect on the respondents' appreciation of initial training; more experienced respondents tend to appreciate initial training more than inexperienced respondents do. 16It is important to note here that this variable might seem ambiguous when being corroborated across samples: it means either 'amount of sentencing experience' (for judges with experience) or 'amount of exposure to sentencing' (for prosecutors, who obviously do not sentence, but by being involved in criminal cases in court are exposed to sentencing, and for initial and induction trainees, who have no sentencing experience yet but might have had exposure to sentencing during their previous legal experience as advocates or police officers).This rule seems to differ slightly between judges and prosecutors.Judges who have at least some sentencing experience and at least some prior legal experience before becoming judges tend to appreciate initial training more than those who have no sentencing or prior legal experience.The amount of experience itself does not seem to make a difference in the judges' case.In contrast, for prosecutors the amount of experience as prosecutors seems to make a difference in their opinions (it increases their appreciation for initial training).

Initial Training
From just looking at these correlations it is not clear why these significant correlations and these significant differences between judges and prosecutors exist.After looking at bivariate correlations, an ordinal regression analysis was run.The existence (or lack) of experience remained significant in explaining the difference in perceptions, while the variable profession (judge vs. prosecutor) was controlled for (Wald χ 1 =36.647,p.0, 95% CI, -1.935 to -.989).This means that respondents who have at least some judicial experience are significantly more likely to think initial training was very useful to them than those who have no judicial experience.What needs to be noted though is that similar results were found for law school17 training as well: the more experienced respondents are, the more likely they are to appreciate their university law studies (see Figure 3 below).This is true for both judges and prosecutors.The fact that the same correspondences between experience and appreciation for initial training and university training were found suggests that, the more experienced respondents accrue, they more they appreciate their theoretical studies; trainees who are currently undergoing initial training or have just graduated are very keen to start earning practical experience and feel oversaturated with what they believe is "too much theory".As it will be seen in the next section, this eagerness to get "their hands dirty" is confirmed by additional findings.When presented with the results, the management team of the Romanian NIM agreed that, from their experience, this is the attitude that they face with initial trainees.

Figure 3: Factors correlating with appreciation of university education
There is no evidence that prior legal experience (before joining the judiciary) influences perception of university training, so this could lead to the conclusion that judicial experience per se affects respondents' opinions unlike general experience in the legal profession.It seems to suggest that the legal knowledge acquired during university becomes more relevant for judicial actors as opposed to other legal actors.But this is not the only possible interpretation; it could also be that university law courses were better fit for purpose in the past than they are nowadays.
An ordinal regression model was also run, with the aim of controlling differences in profession and in verifying if judicial experience impacts on the perceived usefulness of university training.Indeed, it was not just that respondents with no experience tended to see university training as less useful than those with experience, irrespective of profession (Wald =15.594, p.0, 95% CI, -1.134 to -.382).It was also the case that the more judicial experience they have, the more appreciative they become (Wald =42.562, p.0, 95% CI, .081 to .151).
But what were the specific initial training activities that respondents appreciated most?All respondents were asked to rank nine initial training activities given their perceived impact on the respondents' judicial practice.Figure 4 summarises preferences for the nine activities, ranked by the total number of respondents considering them among their top three choices.The chart below presents the aggregated preferences for initial training methods from all samples, although the differences between samples in their preferences were significant and will be discussed below.
Overall, the NIM seminars, the court internship and the practical advice given by trainers during training are the top preferred activities, with more than 30% of all respondents placing them among their top three options.In contrast, the discussions with peers, the role-play exercises and the NIM capacity exam displayed low importance, with less than 15% of all respondents considering them among their top choices.
The main finding regarding preferences for initial training methods depending on experience and/or training is that, while for experienced judges and prosecutors NIM seminars and courses were the top two preferred methods, for initial and induction trainees the court internship was the most important training element.Is this difference caused by a difference in experience or a difference in training, or both?A potential explanation is given by looking more closely at the correlations between different characteristics of the participants and their preference for the second-year court internship (Figure 5).What is key here to note is that participants with no judicial or sentencing experience are much more likely to consider the court internship important than those with at least some judicial experience.In other words, inexperienced trainees are very keen on getting practical experience as soon as possible, and this often supersedes their appreciation for other training methods.This seems to be all the more accurate for judges than for prosecutors.The correlations seem to be reversed in respondents' preferences for NIM seminars (see Figure 6): the more judicial and/or sentencing experience respondents have, the more likely they are to appreciate the value of the past NIM seminars on criminal law/procedure on their subsequent activity.In contrast to practice in court, seminars offer the environment where trainees can discuss legal stipulations and the principles/legislative reasons behind them, the procedural steps to take in specific situations and can ask the trainer about any ambiguities in interpreting the law.Unlike courses (i.e.lectures), the seminars have an intrinsically interactive character which, as considerable research has shown, is much more effective in adult education and training.
Why the mock trial exercises are so unpopular among Romanian judicial trainees is unclear.Mock trials are organised as part of the initial and induction training.The mock trial observations that took place as part of the exploratory phase showed that the standard of organisation was high, and the mock trials were of a high quality, with full involvement of the trainers and the management of the institute.Yet observations also hinted that, if given the chance, initial trainees would prefer having a day visit to court than going through a mock trial -although the trainees did not explain why when they voiced their preference to the trainer.The quantitative results seem to confirm that, indeed, mock trials are indeed among the least preferred initial training methods, and there is no significant difference between initial and induction trainees on this matter.This perception could be interpreted in several different ways.
First, it could be a matter of design: in a mock trial, only one of the trainees gets to play the role of the judge and another gets to play the role of the prosecutor.The large majority of trainees are not actively involved in the role-play, but rather passively observe and comment on the mistakes their peers made.The mock trials and the individual interviews with trainees seem to suggest that, on one hand, the trainee in the active role feels an enormous pressure under the scrutiny of peers and trainers, but at the same time deems that experience personally useful.On the other hand, the trainees who do not get the chance to play any judicial role seem frustrated by the mistakes their colleagues make.The 89% of the respondents who did not pick mock trials as their top three options might well be the 'passive watchers' who did not get the chance to benefit from the exercise. 18Second, an alternative interpretation is given by the trainees' preference to go to court rather than go through a mock trial: in their 'hunger' for practical experience, they might see court visits as more 'authentic' in terms of providing them with the insights of 'how it's like to be a judge' or 'how a day in court looks like'.It may be that after seeing experienced judges and prosecutors in action, the trainees might be more receptive to participation in mock trials because they have some sense of how the roles are to be played.These findings seem to suggest that the perception of judges and prosecutors changes over time, and the judicial institute does well to offer a mix of in-class and in-court methods of training.The inexperienced trainees have an obvious need for practical, in-court methods (such as the court internship or day trips to courts); and the training could be extended to include other similar methods such as job-shadowing.This preference is further reflected in the section on induction 18 When presented with the results in April 2015, the NIM management acknowledged that this could be a good hypothesis of why the mock trials are only preferred by some respondents (those who had an active role in the exercise) but not others.training below, where induction trainees show their support for job-shadowing as part of the training.At the same time, more experienced respondents come to appreciate the specific in-class methods such as the seminars; interpretations of why this is the case are numerous, but correlation with other findings (regarding continuous training and e-learning) might suggest they tend to appreciate the law-oriented, dialogic activities where they get the chance to discuss the interpretation of the law and establish common understandings of what the law requires.

Induction Training
Respondents undergoing induction training (N 27) were further asked what suggestions they would have for the improvement of induction training.As mentioned in the very beginning, this particular dataset is comprised of newlyappointed judges, with no years in post, no prior criminal/sentencing experience, who have never gone through initial training or continuous training before.
According to the findings, most judges would like the induction training be longer (63%) and, in addition, they would like to be able to shadow a judge in their own court before beginning their own judicial activity (70%) (see Figure 7).The latter seems to support the interpretation that trainees without judicial experience are keen to maximise the amount of in-court experience and exposure before beginning their own activity.
The relatively low values for the rest of the suggestions lead us to infer that most induction trainees are happy with the gap between the two modules (89%), the amount of role-play exercises already taking place during the seminars (85%), the fact that they have to go straight into the judicial activity while there are being trained (78%) or the amount of document-writing exercises (67%).
Neither of these opinions was influenced by the amount of prior legal experience, or its type (with the exception of experience as lawyer for the desired length of the training).Surprisingly, the perceived usefulness of training also does not seem to correlate with these suggestions either.Finally, there is no evidence that these suggestions correlate with each other.All the respondents were asked to imagine what would comprise an ideal continuous training session on sentencing.The question was meant to approach activities specific to continuous training sessions, while also being relevant to the sentencing practice.The reader might rightfully ask -why not ask just respondents who had already undergone at least a continuous training session?The main explanation ties into one of the fundamental variables of this study -prior exposure to training, which might account for part of the variability in perceptions.For this reason, even newly-appointed judges and prosecutors who never experienced continuous training were included so as to measure whether even minimal prior exposure to continuous training makes a difference.If this prior exposure made a difference, it was made visible in the following paragraphs.Figure 9 below presents the aggregated responses identifying the most popular training methods overall: The most popular method for a training session dedicated to sentencing would consist of discussing case studies where sentencing is problematic (61% of all respondents ranking it among top three methods).Such a method would not focus on discussing sentencing guidelines or procedures in general or applicable to unproblematic cases, but would focus on those types of cases which can pose problems in practice.
Opinions of judges and prosecutors were split among the usefulness of discussing legal stipulations i.e. discussing the substantive requirements of the law for specific offences (50%), specific legal procedures (51%) and specific issues judges face in their daily sentencing (51%).Out of these, there were more respondents (30%) for whom discussing legal stipulations is of topmost importance among all activities, which could suggest its essential character for any judicial training on sentencing.The high preference for discussing legal stipulations (e.g.sentencing guidelines) relevant for sentencing can also be explained by the fact that respondents who have already undergone continuous training tend to prefer this method compared to initial and induction trainees who have not experienced this type of training yet, irrespective of their role. 19t although the methods above could seem equally preferable to respondents, a closer look at the significant correlations suggests that respondents of different types and levels of experience prefer one rather than another.
Experienced judges and prosecutors prefer discussing the law.The amount of experience (of those who have some) seems to be significantly negatively correlated with a preference for discussing the substantive law on offences, suggesting overall that the more experienced respondents are, the more they appreciate the necessity of this method in sentencing training. 20One interpretation is that the more legal experience (of all kinds) respondents have, the more they experience the ambiguity and interpretability of current laws and the more guidance they need in how to coherently interpret laws.This need was voiced by participants during the continuous trainings observed during this research.This is particularly true for a civil law country such as Romania where reliance on higher court precedent is not officially recognised or sanctioned.This issue of needing more policy and protocol guidance is accentuated by the enactment in 2014 of two new criminal law and criminal procedure codes that require judicial actors to carefully re-acquaint themselves with those codes and determine how to interpret and apply them without the informal benefit of an established jurisprudence based on precedent.
Respondents with prior advocacy/policing experience care about procedures.Interestingly, respondents who had prior legal experience before entering the judiciary are more likely to see discussions of legal procedures more useful than those with no prior legal experience.Moreover, the amount of legal experience seems to matter.Could it be because their past practical experience showed them that knowing and understanding the procedures is important in judicial decisionmaking and sentencing more specifically?
Inexperienced judges need to hear about other judges' challenges.Inexperienced judges prefer discussing specific issues judges face daily in real cases, but no such correlation was found for prosecutors.This could suggest that it is the judicial experience sui generis that makes this kind of discussions less useful during continuous training.It could either be that experienced judges already enjoy these discussions in their own courts, with their colleague judges, so they do not feel the need to discuss them during NIM training.Or it could be that they become wary of discussing specific issues they face in their cases with their NIM peers, believing that would be inappropriate.
Simulations again prove unpopular.Surprisingly, only 33% of all respondents considered sentencing simulations/exercises their favourite learning activities.This finding can be corroborated with the previous finding that mock trials and simulations are not as popular in initial training as probably expected.
'Legalistic' versus 'case-based' sentencing training methods.If one looks at preferences across types of training (initial and continuous), there seem to be two clusters of respondents based on their preferences for training methods.The first cluster is composed of respondents who prefer 'legalistic' methods, which consist in analysing and discussing legal regulations pertaining to sentencing.In this sense, a very high correlation was found between respondents who prefer legal analysis of current law and those who prefer analysing procedural regulations.Moreover, fans of 'legalistic methods' are less likely to prefer 'case-based' methods, such as discussing issues encountered by judges, case studies, sentencing exercises, or landmark cases.Preference for discussing landmark cases is split among judges and prosecutors, suggesting a more hybrid nature of the method.Conversely, respondents who prefer one 'case-based' method, i.e. a method who makes use of specific cases or examples to teach sentencing, are also more likely to also prefer the other suggested case-based methods ('discussionbased' methods are particularly correlated: case studies and landmark cases and real issues encountered by judges).E-learning methods.All respondents were asked what an e-learning session should contain so that it would engage them throughout.There were no significant differences between the types of respondents.The only point where judges and prosecutors differed significantly was in their take on the existence of a forum where magistrates could exchange ideas and approaches, with judges considering the forum more useful than prosecutors.

Figure 9: Favorite e-Learning Methods
Overall, all e-learning methods suggested were considered useful by at least 86% respondents, so we recommend their implementation on an e-learning platform dedicated to judicial training.The methods were ordered first and foremost by the ratios of respondents considering them essential, with direct access to relevant legislation being the top option (66% essential), followed by multimedia materials provided by trainers (54% essential) and live Q&A sessions with trainers (53% essential).
3.4.E-learning and experience.Surprisingly, years in post did not seem to make any difference in respondents' attitudes towards elearning methods overall.But by looking in more detail at judges' vs. prosecutors' experience, it seems having judicial experience does affect judges and prosecutors differently.
On one hand, it affects judges, but not prosecutors, in appreciating live Q&A sessions and access to legislation.First, judges with judicial experience are much more likely to appreciate live Q&A sessions with trainers than inexperienced judges.This is further reinforced and clarified if we look at judges with experience in sentencing; they are more likely to require live Q&A sessions with trainers where they could ask questions than those without experience.This is specifically true for judges but not for prosecutors.This could be because having experience in sentencing multiplies the issues and the problems a judge would address in an online Q&A session.This was anecdotally observed in live Q&A sessions as well.The Q&A sessions were the only e-learning method correlated with existence of judicial or sentencing experience.
Second, the amount of judicial experience seems to make judges more appreciative of given direct access to relevant legislation: namely, the more judicial experience judges have, the more likely they are to appreciate direct access to legislation.This is very much in line with the previous findings pointing out that more experienced judges look for ways to keep in touch with and discuss the law itself rather than worry about their practical experience.On another hand, experience affects prosecutors, but not judges, in appreciating forums with magistrates.Among prosecutors with experience, those with more criminal experience and/or with more prior legal experience are significantly less likely to think a magistrates' forum is a useful e-learning tool.

Judicial Craft -Can it be Taught?
This findings report concludes with the responses received to a question regarding the possibility and effectiveness of judicial skills training."Judicial craft" is typically defined as being separate from content-oriented training, focusing specifically (and exclusively) on the skills judges need to develop. 21To date, the Romanian National Institute of Magistracy opposes the idea that standalone judicial craft sessions can be useful and effective in isolation from legal content.Under this view, the Institute has not organised specific judicial craft seminars, but it incorporates skills training in its seminars and mock trials.This question was meant to test if and how do the opinions of the trainees themselves differ from what the NIM management thinks.But beyond that, it raises the more general question whether specialised skills-training can be regarded as feasible in the absence of legal content and separate from practical experience. 22dges and prosecutors undergoing continuous training were asked if they think they were offered enough opportunities to develop their "judicial craft", in addition to the content-oriented sessions.The meaning of the expression was inferred through an enumeration suggesting what would count as relevant judicial skills (writing decisions, sentencing, dealing with expert witnesses).This enumeration was slightly adapted for prosecutors to reflect relevant prosecutorial skills which would make sense to them.
The answer choices offered aimed to reveal two aspects: 1) if the respondent thinks judicial skills can only be gained through experience, or judicial skills training would also help; 2) if the respondent thinks judicial craft can be trained, is he/she happy with the amount of training received so far.Out of 282 total respondents envisaged, 95% (267) responded to this question.

Figure 1211: Perceptions on the possibility and effectiveness of judicial craft sessions
The most popular answer among respondents was that judicial skills can only be cultivated through experience (42%), so training is irrelevant and no amount of specialised skills training can help.In contrast, 57% respondents seem to think judicial training can have a bearing on 'judicial craft'.Among these, 22% would welcome supplementary skills sessions; 21% consider the training already supplied the necessary skills; and 14% consider the training was insufficient but no further training is necessary as the initial period of practice supplemented the training.
There are some significant differences between how the experienced judges and prosecutors regard 'judicial craft'; while judges are more likely to think that 'judicial craft' sessions should exist and are much less likely to think the current training was sufficient, prosecutors think the training was enough to develop their skills.
The amount of judicial experience seems to influence the attitude towards judicial craft as well: more-experienced respondents are more likely to think that judicial skills can only be cultivated through experience; lessexperienced respondents are more likely to believe the training was insufficient, but they managed to compensate during their initial practice.The difference between the two choices is that the less-experienced respondents seem to believe that the training provided them some important skills, while the more-experienced respondents seem to deny any merits of the training and claim these skills come exclusively through experience.
The amount of criminal experience seems to have no influence on judges, but it does nevertheless seem to make prosecutors more sceptical with regards to judicial craft training.More-experienced prosecutors claim judicial craft cannot be trained and can only be learned through experience, while less-experienced prosecutors claim the training was insufficient, but they managed to compensate during their initial period of practice.

Conclusion -Summary of Findings
This article is meant to present some of the findings resulted from a European study on judicial attitudes towards judicial training.So far, one jurisdiction was involved and although the study enquires on a diversity of matters, ranging from judicial role, factors in sentencing up to informal peer support, this paper focused on the findings pertaining to the main theme, judicial training.
The key insights to remember: 1. Judges are overall appreciative of the initial judicial training they received, with 93% considering it useful or very useful.This insight confirms that judicial training is seen as a useful practice and every jurisdiction should consider maintaining (or implementing) it.2. Judicial experience seems to generally have a positive effect on the respondents' appreciation of initial training; more-experienced respondents tend to appreciate initial training more than inexperienced respondents.This suggests that training programmes that are not evaluated longitudinally or evaluated by only asking recent trainees about their current experience of the training might underestimate the impact the training has throughout the judges' careers.3. Similarly, the more experienced respondents are, the more they appreciate their university law studies.This suggests, again, that a study of perceived value of law degrees need to be done longitudinally and/or needs to involve alumni of various levels of professional experience to get an accurate sense of its value.4. While for experienced judges and prosecutors NIM seminars and courses were the top two preferred initial training methods, for initial and induction trainees, the court internship was the most important initial training element.On one hand, inexperienced trainees are very keen on getting practical experience as soon as possible, and this often supersedes their appreciation for other training methods.On the other hand, the more judicial and/or sentencing experience respondents have, the more likely they are to appreciate the value of the past NIM seminars on criminal law/procedure on their subsequent activity.This difference in attitudes needs to be taken into account in both the design and the evaluation of training sessions.5. Mock trials are unpopular among Romanian judicial trainees.Two possible interpretations: either mock trials are seen as less authentic than actual court experiences, or only a small fraction of trainees actually get to 'play' the relevant judicial role.The design of mock trials might have to be reanalysed.More qualitative research needs to be done to enquire on the reasons for why mock trials are unpopular.6. 70% induction trainees would like the opportunity to shadow a judge before beginning their sentencing activity.
This should be seriously taken into account in the design of induction-like (or crash-course) trainings.7.During continuous training sessions, experienced judges and prosecutors prefer discussing the law, while inexperienced judges need to hear about other judges' challenges.In addition, respondents with prior advocacy/policing experience care about procedures more than those without legal experience.This could suggest that different continuous training sessions could be devised for judges of different levels of experience; if they are common, they should pay attention to integrating all these activities in order to cater to all needs.8. Judicial experience does not seem to make a difference in the respondents' attitude towards the usefulness of various e-learning methods.The only exception is that experienced judges seem to need more online Q&A/feedback sessions with trainers; moreover, as they get significantly more judicial experience, judges also tend to appreciate direct online access to legislation.It is recommended that all these e-learning activities should be implemented on an online platform, as 90% of respondents agreed that they all are useful or very useful.9. Finally, 57% respondents think judicial training can help with the development with judicial craft, while 42% think it can only be cultivated through experience.Experienced respondents are more likely to think that judicial skills can only be cultivated through experience.This supports the idea that a judicial institute must strive to offer judicial craft courses which would introduce the judicial trainees to the basics of judicial skills, while also emphasising the importance of practice in cultivating and improving the relevant skills in the long term.
The findings of this study would be all the more powerful and relevant if conducted in several jurisdictions.Therefore any opportunity to run the study again in new jurisdictions is welcome.Also very welcome would be any reference towards similar research conducted recently that might further inform the current findings, or discussions on the external validity of the findings as well as the degree to which they can be extrapolated to other jurisdictions.What hopefully can be concluded from this article is that this type of research can genuinely contribute to the design and assessment of judicial training programmes, by offering a rigorous and independent perspective on how they can be improved in the future.
by Profession.Out of the 470 responses considered valid and taken into account in the data analysis, prosecutors are distinguished from judges.For continuous training; this was straightforward as judges and prosecutors had different versions of the survey.For induction training, the 2014 generation only had judges.For initial training, trainees

Figure 1 :
Figure 1: Distribution of types of training per judicial role Out of 453 respondents, 307 (68%) underwent initial training, while 146 (32%) did not.For purposes of this research, induction training is considered initial training.While within NIM, the induction training is organised by the department of continuous education; we consider induction training a kind of initial training because it is dedicated to trainees with no prior judicial experience, just like initial trainees.Those who underwent initial training began their training between 1992 and 2013.Ninety-three of those who have undergone initial training now consider it useful or very useful for their career.This suggests that initial judicial training is seen as a useful aid in preparing judges and prosecutors for their careers.Perhaps more interesting than the aggregated results are the correlations that were found between the respondents' perception of initial training and other characteristics such as prior experience and further exposure to training.Figure2summarises these correlations.

Figure 2 :
Figure 2: Factors correlating with appreciation for initial training

Figure 4 :
Figure 4: Most appreciated initial training activities

Figure 5 :Figure 4 :
Figure 5: Characteristics correlating with prioritising the 2nd year court internship as a useful initial training activity

Figure 7 :
Figure 7: Amount of continuous training

Figure 8 :
Figure 8: Most useful activities in a continuous training sessions on sentencing

Figure 10 :
Figure 10: Characteristics correlating with preference for online Q&A sessions Initial training seems to have an important bearing on the respondents' attitudes towards judicial craft training.Those who benefitted from initial training are more likely to think that part of their judicial skills were provided by the training and part by initial practice while being less likely to think judicial craft can only be developed with experience.In contrast, respondents who have not benefitted from initial training are less likely to admit training played a role and more likely to claim judicial craft can only be developed with experience.

Figure 12 :
Figure 12: Characteristics correlating with the belief that judicial craft can only be cultivated through experience " (Working Group on 'Professional Judicial Systems', 'Enhancing Judicial Reform in the Eastern Partnership Countries: Training of Judges' (Council of Europe 2012) 7.) 7 David A Kolb, Experiential Learning: Experience as the Source of Learning and Development (Prentice-Hall 1984); Virginia L Lester, 'Life Stages and Learning Interests.'http://eric.ed.gov/?id=ED159945 accessed 29 September 2014; John Dewey, Experience And Education (Simon and Schuster 2007); John Dewey, How We Think: A Restatement of the Relation of Reflective Thinking to the Educative Process (DC Heath and company 1933).

Table 1 : Stratified Sample in Romania Case study: ROMANIA Experience Training
The study employs a mix of qualitative and quantitative methods.In the exploratory phase, observations of different types of judicial training sessions on sentencing were conducted as well as interviews with trainers and trainees of different types.Three interview guides were formulated for interviews with the three types of trainees.Since the NIM trainers are the same for all three types of training, only one interview guide was needed which included questions about all three types of training.The exploratory phase took place between January and February 2014.