Implementation of Quality Management in National Statistical Institutes

Transcription

1 1 CD-34-1 Implementation of Quality Management in National Statistical Institutes by Margarida Madaleno, Thomas Körner, Walter Radermacher, Max Booleman and Richard Van Brakel Summary Quality Management is implemented in National Statistical Institutes (NSIs) following decisions made by top management, as part of the organisation s overall strategy. It is clearly a top-down process, but the involvement of the whole organisation is also a necessary condition for success. It is a long-term commitment and implies a change in NSIs from product to customer orientation based on a new organisational culture and working on a set of core values such as continuous improvement, benchmarking, teamwork, innovation, and others. General guidelines for practical implementation of quality management can be outlined. Some resistance and other obstacles might emerge. All NSIs in the European Statistical System (ESS) should consider the adoption of a systematic approach to quality. 1. Introduction NSIs are nowadays confronted with a rapidly changing environment and facing major challenges such as new statistical requirements, new information systems, competition in the market, increase in and diversification of users, combined with stable or lower budgets. So it is essential to define their mission and vision, to outline strategies and policies in order to achieve good quality of their products and services. The LEG discussed different systematic approaches to quality and the links between them. The aim of quality management is the continuous improvement of quality and performance of the organisations. Most NSIs have ongoing quality improvement projects, but work is not always systematic. However, many NSIs have already opted for business excellence models and self-assessment tools to find strong and weak points, as a basis for improvement work. Yet the decision for a management model is just a first step towards systematic quality management. In order to achieve effective and suitable quality management, actions taken for its implementation are crucial. Top management commitment, staff involvement, a clear quality organisation as well as review and documentation have to go hand in hand. In the course of its work, the LEG outlined some general guidelines for the implementation of quality management, some of which in great detail, such as the quality framework, user orientation and documentation. The LEG report makes recommendations for future work in this field for all NSIs within the ESS. 2. Starting and managing a process for change The concern about quality is not a recent development in NSIs, what is recent is the systematic, integrated and organised way of working with quality issues. NSIs manage and improve a large number of processes involving products, services and user relationship. To

2 2 develop quality management, NSIs must identify and design the key processes needed to be worked, define sequences and take and measure actions, in order to enhance the satisfaction levels of users, both internal and external. It is important to focus on the processes, as product quality is achieved by means of process quality. But other aspects should be taken into account, such as planning (goals, timings and responsibilities) and the identification of all interested partners and their respective roles in this process. In NSIs the main purpose is continuous improvement of the quality of the products and services, and there are different ways to put this approach into practice, using various quality management frameworks. Empirical studies have shown that the implementation of quality management is a long-term commitment. It requires a change in the organisation s culture, supported by a new vision, based on a set of core values shared by all staff, such as: Leadership (defining a clear vision and mission, getting personally involved in quality work, investing in skills, innovation and creativity) User orientation (focusing on the needs of internal or external users, strengthening their relationship, improving accessibility of products and services) Process-orientation (focusing on the processes that create products, analysing and continuously improving the key-processes of the organisation) Teamwork (teambuilding, improving communication, stimulating teamwork as a motivating force) Benchmarking, continuous improvement and review (sharing good practices, learning from others experiences and measuring results). Each NSI can emphasise and concentrate work on different values in its own quality approach. However, all NSIs must have a clear perception and sense of their mission and goals, give priority to concrete results, and assess their results taking into account user satisfaction, aiming at efficiency and quality. The commitment that has to be assumed is a long-term investment on change. Organisation and structural changes must be introduced and sustained over the short and long term. It is not an easy and fast way, some changes can be put into practice immediately, but others face resistance and problems, namely the existent organisation s culture. Big and small changes have their own timing and need to be managed. Organisations normally take 5 to 10 years to progress and reach a higher level of quality. It demands continuous, patient and persistent work. But even then there are always new opportunities for improvement. Introducing quality management and making it work means implementing change in NSIs, and, as any process of change, is expected to face some constraints and resistance. Some examples are: Internal problems of motivation (goals of the organisation are not clear enough; the need for change is not well explained; fear of different working conditions or need of new skills, etc.) Low willingness to take risks on the part of management (commitment is missing, resources for assessment or benchmarking are lacking) Insufficient change management (the language of quality management is too formalised and abstract, it does not take the specific situation of the organisation into account) Lack of communication and co-operation (lack of external support, inefficient teamwork etc.).

3 3 3. Guidelines for implementation of quality management in the ESS There are many ways to implement systematic approaches of quality work. NSI s are unique, with very distinctive strengths and opportunities, frameworks and resources. However, some collected experiences show that there are common important phases which may be followed: 3.1. Leadership Top management commitment As mentioned above, the implementation of quality management is, at the beginning, a topdown process. Top management decides univocally to start the implementation and to support it personally, through actions rather than statements. Top management has to set clear goals and priorities, break barriers and obstacles and stimulate continuous improvement. This commitment includes the willingness to provide the adequate human and financial resources for the implementation of the strategy and the system. The quality culture of the organisation should not be affected by a change in leadership. Strategic Orientation Top management has to make sure that implementation of quality management should be supported by the vision, mission and values of the NSI. Strategy formulation and policy making are clearly top management assignments which can hardly be delegated to other people in the organisation. Examples can be found in mission statements and quality declarations, such as Eurostat s corporate plan, TQM at Statistics Sweden, ONS 2000 change program in UK, NSI Portugal s Quality Charter. User Orientation Top management should communicate to all staff that the focal point of the activities of NSIs is the user of official statistics. The importance of involving users in the production and dissemination of statistics and improving their perceptions by creating added value has to be stressed during the implementation of quality management. Some NSIs assess and improve the levels of user satisfaction through regular customer satisfaction surveys. The user-producer dialogue should be strengthened. In most countries in Europe, the links between users and producers are established by Statistical Councils, whose main task is to review the statistical programs and set priorities Involvement of the staff The next step is the involvement and participation of all staff. Once initialised by top management, activities should be carried out as bottom-up processes (as far as possible). Important elements are information, participation and quality training. Communication and Information Good communication is essential for successful implementation of quality management. Effective two way communications should be achieved with the staff. People should be

4 4 informed of what is planned to make participation easier. Fears about the outcome of changes can be (partly) avoided only by open information. Communication between different levels and functions has to be assured. Benchmarking sessions can be convened within the organisation or with similar organisations. Recognition systems are adjusted to honour employee contributions to quality improvement through suggestion boxes, team participation, innovation and initiative. Participation of all Contribution to improving performance can come from every employee, regardless the respective position in the hierarchy, generating benefits for the NSIs. Change can come from individuals or small teams acting as the driving force for the change process. Examples are TQM pilots at Statistics Sweden or Quality poles in NSI Portugal, specially trained quality facilitators who assist participants in quality work and the use of quality tools. Mobilisation of employees for quality programs is a condition for the success of quality management. NSI Portugal and the Federal Statistical Office of Germany have recently undertaken so called Quality Meetings, aiming at involving all staff and giving information regarding its implementation. Quality Training Persons who are assigned for specific tasks within quality management should be competent and have a good mix of skills and knowledge on the basis of education, training and experience. Staff at all levels has to be trained in quality related tools, beginning with the top and working down. It is important that everyone, at all levels of the NSI, is actively involved in the implementation of quality management. Responsibility and accountability are assigned at the lowest level Quality organisation For the implementation of quality management some basic organisational requirements have to be fulfilled. Quality Management Organisation Management decides how the quality work has to be organised and makes sure that the necessary responsibilities and authorities are defined e.g., it decides whether or not a quality committee is to be created and what its mandate is to be, appoints the team responsible for promoting quality, as well as the quality facilitators to work on continuous quality improvement. Resources for Quality Management The implementation process of quality management is an investment, which needs human and financial resources. The commitment of the management has to include the willingness to pay for this task. Returns on investment have to be regularly evaluated and measured by quality and performance indicators.

5 5 Project Management In the beginning of the implementation, quality work will usually be organised as improvement projects and have to be carried out in accordance with the general rules of project management. Such projects can fall into different categories, some can be designed to bring improvements in timeliness, others to standardise methods and procedures or reduce production costs in survey processes, thus increasing efficiency. Quality Documentation Documentation is vital to ensure and improve quality, and also to better understand and use statistical data. However, it should never be considered as an aim in itself, but as an activity to support the implementation of quality management. Some documents describing NSIs strategies and policies, statistical production processes and data, and assessment tools are required for a systematic approach to quality. These documents should be controlled, reviewed and regularly updated Review and Continuous Improvement In order to ensure that quality management is effective, results should be monitored, measured assessed and analysed so that a process of continuous improvement is assured. Evaluation of Quality The process of identifying strengths and weaknesses and setting priorities is very important for many phases of the quality management implementation. One possible way is through self-assessment, an inventory to be carried out by the organisation. Within such an assessment a team is pressed to elaborate a common judgement of the status quo of the NSI. This assessment should be carried out in accordance with a framework, which helps organisations in this exercise. Different management models can be used as frameworks for self-assessment and quality improvement. The most commonly used are business excellence models, inspired by TQM principles, with similar components such as leadership, people, partnerships and processes, amongst others. The EFQM model is a well-known framework, which is used by some NSIs. This evaluation of quality should be repeated regularly in order to measure the progress (return on the investment), and allow benchmarking with similar organisations. Quality of Processes In NSIs a better control of processes will improve the final products - the statistics. Hence, the core of quality management is an optimisation of processes predominantly in production, but also in internal support and management processes. An evaluation of the most important processes should be a module of quality management. As product quality is achieved by means of process quality, NSIs must identify and manage a large number of processes. The key processes and their variables should be selected and controlled in order to introduce improvements and increase their efficiency. Process orientation is one of the basic elements of TQM.

6 6 Quality management will work better if built on sound statistical and scientific principles. For each statistical product it is possible to define and understand the production process and a measure of quality, and then improve the product through process improvement. Quality of Products A good process is a necessary but not sufficient condition for having good products. To be an effective organisation, the mix of products should be continuously optimised. The optimal mix is depending on user needs, costs, the available budget, the response burden and technical possibilities. The characteristics of the products should be described with the help of the underlying quality dimensions, such as accuracy, timeliness, comparability, accessibility and coherence. And during the production process all phases should be monitored to obtain, in advance, the specified characteristics of the process and its products. Improvement Projects Some NSIs undertake quality work in the form of improvement projects. Improvement is of course imperative not only in the statistics production, but for other processes and activities as well. Improvement projects should be initiated within all areas of NSIs. To support the work in these projects there are a number of proven quality tools that can be used. These include Flow-charts, Pareto Diagrams, Cause-Effect Diagrams, Control Charts, Brainstorming and more. Another powerful way to support the improvement activities within projects is to select and train "Quality Facilitators", with knowledge of effective team-work and the application of the mentioned tools. This approach has been taken in Finland, Norway, Portugal and Sweden with good results. 4. Conclusion NSIs face new challenges and operate in a fast-moving environment, so they need to adopt a systematic approach to quality work if they want to reach higher levels of quality of their products and services. This work must clearly be carried out as both top-down and bottom-up approaches. Frameworks such as business excellence models inspired by TQM principles (EFQM or others) can be used by NSIs to provide statistical data and services with high quality and enable continuous quality improvements. This systematic quality approach is difficult to implement and demands a lot of effort and this is the reason why so much emphasis should be put on the procedures. The LEG report provides some general guidelines for the implementation of quality management. EUROSTAT and many NSIs within the ESS have already started quality work on a systematic basis. Benchmarking and learning from each other s experiences should be stimulated through increased co-operation in quality issues at the European level. The search for Quality has to be permanent and pursued all the time by all NSIs. REFERENCES Process Management to Quality Improvement: The way to Design, Document and Reengineer Business Systems BORN, Gary. West Sussex, England: Wiley, Survey Measurement and Process Quality LYBERG, Lars; BIEMER, Paul; COLLINS, Martin; DE LEEUW, Edith; DIPPO, Cathryn; SCHWARZ, Norbert; TREWIN, Dennis. New York: John Wiley & Sons, INC,1997.

7 7 The EFQM Excellence Model EUROPEAN FOUNDATION FOR QUALITY MANAGEMENT. Brussels: EFQM,1999. The Management of Quality and its Control TAPIERO, Charles S. Paris: Chapman & Hall,1996.

8 CD-34-2 KEYS TO SUCCESSFUL IMPLEMENTATION OF CONTINUOUS QUALITY IMPROVEMENT IN A STATISTICAL AGENCY David A. Marker and David R. Morganstein Westat 1. Introduction Westat has worked with statistical offices in Sweden, Finland, Norway, the Netherlands, and Denmark to improve quality using a more coherent planned approach than has been used in the past (Morganstein and Marker, 1997). In the United States Westat has assisted the Census Bureau, National Center for Education Statistics, and other Federal agencies. For 20 years we has assisted private sector organizations throughout North America and Australia in similar quality improvement work. Six key factors have emerged as vital to the success of continuous quality improvement (CQI) efforts in statistical agencies. Top management commitment; The structure of the organization Self-supporting operation; Improved communications; Team work; and, Continuous improvement. 2. Top management commitment Far and away, the most successful improvement work has had top management s commitment to and personal involvement in continuously improving quality. While it is possible to make selective improvements, driven by groups of committed individuals or middle managers, continuous improvement can not be sustained without top management making such efforts a priority. Top management is also needed to break down barriers between units, barriers that impede progress. Management sets the priorities on how their staff should spend their time. If the priority is production deadlines, or explanations for problems with previous production, then the 1

9 CD-34-2 staff will quickly see that extra effort spent on prevention is neither valued nor rewarded. As an example, consider an automated procedure for determining which maps to review during the production of millions of maps for a large-scale in-person data collection study. When it was suggested that review of this procedure and the development of flow charts might identify ways to make it even more efficient, and that other parts of the organization might benefit from learning how this automated procedure worked, the response from the lead person was how will I benefit? There was no sense that this person s manager had conveyed a value that improvements to any part of the organization would be beneficial to all. They perceived no reward for identifying how to continually improve what they do or that their manager considered this an integral part of their job responsibility. Top managers must demonstrate their personal involvement. They must understand the goals, methods, and cultural change required to become an organization focused on continuous improvement. The best way to demonstrate this involvement is for Directors General to use these methods in their own work. Re-examine your work operations for possible activities to improve, participate in relevant quality improvement teams, make sure you understand your customers needs. Most importantly, change the questions you ask of your staff. One very successful Director General was known for stopping managers and nonmanagers in the hallway to ask, how are your quality improvement efforts going? Have you mapped your key processes? Have you identified obstacles to improvement? These types of questions show a level of knowledge and commitment that cannot be communicated by simple slogans or speeches. New questions that show an understanding of continuous improvement indicate that top management is focusing on the processes that determine the quality of products and services. Another Director General was a regular attendee at staff meetings to review proposed updated current best methods (standards). His active participation in this effort helped move forward their adoption. He also demonstrated to other staff the importance of their participation in this process. Management s support for CQI must be consistent. An important mechanism for this is to incorporate quality leadership into the evaluation/reward structure. Most organizations review their staff s performance annually (or at the conclusion of major projects). Does your evaluation value participation in quality improvement efforts? Does it reward cooperative 2

10 CD-34-2 activities, like teamwork, that will improve the overall quality of the organization, or does it force units within the organization to compete against each other? One type of reward worth special mention is the establishment of quality awards. These awards can either be supportive of CQI, or undermine such efforts. If an award is given to the best team each year, the message sent to your staff is not to share your success (at least initially) because then another team might use these ideas to beat your team for the award. This sets up barriers to communication that actually slow down improvement efforts. Alternatively, awards can be given to all outstanding teams deserving of recognition. This implies that some years there can be multiple winners, and other years there may be no winners. This process demonstrates a better recognition by management of the need to reward all worthy efforts. There are many forms such rewards can take. Many national statistical offices are constrained from changing staff salaries. Alternative rewards can include: promotions, participation in desirable projects, opportunity to attend conferences on interesting topics or in interesting locations, recognition at staff meetings, etc. Management sets the priorities. One priority should be finding new areas to improve. It is neither possible nor desirable to attempt to improve everything in an organization simultaneously. Priorities must be assigned. While some processes are being investigated, others should be identified for future effort. Managers must plan for the resources necessary to continue these efforts. Staff who have completed participation in a quality improvement team are usually anxious to use these tools again to improve other parts of their operations. Unfortunately, management is often not ready for follow on activities as soon as their staff are ready. Obviously such activities have to be balanced with other job responsibilities, but management s lack of advance planning is often a strong impediment to building the self-sustaining effort that is required. One obstacle faced in almost every organization (and described in Deming s 14 Points) is the barriers between departments. This is particularly true of governmental organizations with stove pipe organizations (as described by Dillman 1996). For-profit companies eventually recognize that if they share successful ideas the entire company can become more profitable. It is not so obvious why employees of a governmental statistical organization should share their successes. Only top management can convince their staff that the reputation of each part of the organization affects all other parts. Breaking down these barriers is vital to improving the communication and efficiency of the entire organization. Staff from 3

11 CD-34-2 multiple divisions, both customers and suppliers, can be included in the same quality improvement effort. Training programs and other activities can involve staff across departmental boundaries. Overcoming existing barriers is an excellent opportunity to improve quality throughout your organization. Dale and Lascelles (1994) discuss the evolution of an organization through stages of the quality improvement process. They describe six easily recognizable levels which they describe as follows: 1. "Uncommitted. Not yet started. 2. Drifters. Quality improvement perceived as a program, rather than a process. The management has no plan for deployment throughout the organization. Teamwork is superficial. 3. Tool-pushers. Employ a selection of quality management tools. Looking for the latest panacea. Meeting output targets is management s key priority. Emphasis on solving current, rather than future problems. 4. Improvers. Understand TQM involves long-term cultural change. Members of senior management have demonstrated commitment through understanding and leadership. Policy deployment and infrastructure in place, a proactive quality system. Long-term education and training. Process improvement activities throughout the organization. 5. Award-winners. Leadership culture that is not dependent on the commitment and drive of a limited number of individuals. Effective crossfunctional management processes. Strategic benchmarking practiced at all levels. A participative organizational culture. Powers of decision making relinquished by management to people at lower levels. 6. World Class. Total integration of quality improvement and strategy to delight the customer. Probably taken ten years to reach this level. Management focuses on enhancing competitive advantage by increasing the customer s perception of the organization and its products and services. Total quality is a way of life." Notice that a critical discriminator at all of Dale s levels is the degree and nature of management involvement in changing the culture of the organization. Two final comments are worth considering. First, communications must be twoway. For example, it is a good idea to ask your staff for suggestions. However, you must be prepared to be responsive to reasoned suggested improvements. This doesn t mean that you must 4

12 CD-34-2 agree with all suggestions. But if you wait a year before responding to suggestions, staff will quickly determine that you don t really want my opinion and will stop participating. Second, don't view CQI as something done in "Quality Projects." It is true that when you begin a CQI effort, it is necessary to demonstrate the process of improvement through quality improvement teams working on specific projects. But this is an interim solution, the goal is to use these tools and procedures widely and make them the way of working in your organization. Always be on the look-out for opportunities to use these tools in your other activities. In this way you will eliminate any confusion that quality work is different from regular work. 3. The Influence of Organizational Structure on Quality and Innovation The structure of the organization, the patterns of control and communication, the centralization or decentralization of critical skills, and the reward system, all effect the speed at which the organization can evolve. Changes in these components change the resulting output. In his 1996 JOS paper, Dillman argues that there are important barriers to innovation in a statistical organization: differences in the dominant value systems of operations and of research cultures; and, the difficulty of resolving those differences in a hierarchically-oriented organization. His four step solution is:! "The presence of professionals with training in the theories relevant to defining, identifying and resolving measurement and nonresponse error issues;! The capability for understanding and working to overcome measurement and nonresponse error into the operations, as well as, the research culture;! Dealing effectively with organizational structure; and, finally,! Increasing people s understanding of the multidimensional nature of survey error (people both within and without of the organization)." There are, of course, other factors that affect the pace of innovation, even among the most educated of scientists (see, for instance, Kuhn 1970 and Goleman 1985). Dillman s observations are important. Notice the theme of differences between the research culture with its historical focus in modeling, measuring, and reducing errors, and the operations culture where the 5

13 CD-34-2 budget and schedule may have priority over error control or reduction. Deming (1982) reminded us to "break down barriers between staff areas." The cultural differences mentioned by Dillman are some of the barriers to which Deming referred. Hierarchical organizational structures are likely to inhibit or reduce communications. Decentralized services and geographically isolated units are likely to cause this, as well. Reward systems that pit one group against another discourage cooperation, "If the other units win a bigger slice of the budget pie, then our slice will be smaller," and create the "notinvented-here" syndrome. These attitudes discourage the development and use of effective, efficient, standardized software. Hierarchical organizational structures require that approvals for change traverse up and down a chain of command. They may be well suited for static on-going surveys where consistency with previous practice was paramount. However, they do not encourage cooperation across departments and, therefore, are not as adept at dealing with rapidly changing environments, such as the harmonization process being faced by many European statistical agencies. Gross and Linacre (1997), in their discussion of Australian Bureau of Statistics efforts to harmonize estimates across business surveys, identified three requirements for quality management, all of which are made more complicated by stove pipe organizations. The staff in these hierarchical organizations know well how to repeat the same survey they have conducted year after year. However, they are not as prepared to make wide-scale changes in survey instruments, sample designs, and delivery schedules. How then can these barriers be reduced? In a discussion about Quality Improvement in the Australian Tax Office, Godfrey (1994) reviewed his experiences with the problem of typical government organization structures and presented some solutions. It was clear to him that the hierarchy-inhibited change and that team work, as we discuss in more detail later, would help bring down the barriers. He focused on the training process as a key to changing attitudes and behavior. "We recognized that we needed to re-create training and development as one part of a whole system of policies (job design, rewards and promotion, work force planning, career planning) which might help us change attitudes and behavior. The most successful approaches, we have found, begin with whole teams." We suggest that a better balance is needed between an organization that conforms to regulations, rules, and policy dictates (based on the "continental European" view, derived from 6

14 CD-34-2 the Napoleonic Code) and one that sees the purpose of public administration as creating quality outcomes delivering a service (the "Anglo-Saxon" view, based on English "public/civil service" tradition). Godfrey used a combination of training, the use of teamwork, and a refocus of the reward system on those who can lead the changes needed. These steps will help overcome the inertia of a static hierarchical structure that is good at following the rules but may lack innovation and the skills and motivation to evolve. 4. Self supporting A CQI culture must become self-supporting if it is to be truly continuous. In the initial stages, there may be a need for outside consultants to train management and key staff, to provide additional skills as needed, and to help with benchmarking and other activities. But the goal should be to develop the in-house skills necessary to continuously improve, with less and less dependence on outside resources. All managers need to be trained so that they can be supportive of the quality improvement efforts. However, they must do more than provide verbal support. Their decisions should be consistent with a change in culture. A key to making this self-supporting is to train internal facilitators who will be knowledgeable in tools for quality improvement, guide quality improvement teams, look for opportunities to use these tools in other settings, and generally serve as spokespersons for the CQI effort. Their personal involvement helps persuade other staff of the usefulness of this effort. Their commitment breaks down barriers of suspicion that might result if the effort is only being pushed by management. These facilitators need to be good communicators, respected by other staff, and comfortable with statistical concepts (although they do not need to be statisticians). The other key to self-sustaining CQI is the development of a Guidance Team. The goal of the Guidance Team is to provide the resources and mid-course corrections often needed to successfully implement the CQI effort. They must carefully select internal facilitators and areas for experimenting with team work. By regularly communicating to both top management and to those actively involved in CQI, the Guidance Team helps break down barriers between management and staff. The Guidance Team should regularly seek feedback from teams and internal facilitators on the progress, or lack thereof, to date. Based on this information, they can decide when additional resources, training, or policies should be implemented. 7

15 CD-34-2 For a Guidance Team to succeed, it needs to have the confidence of both management and other staff. Thus its membership should include representatives of top management, middle management, and facilitators. Someone who reports directly to the Director General, and is present at all regular senior management meetings should be part of the Guidance Team. Wide representation helps to assure that the Team s recommendations will be addressed by those with the power to modify existing practices. 5. Improve communications Improving communications is key to CQI and is the gold mine of improvement opportunities. Communications can be improved between managers and staff; between units in the same departments or across departments; between people and the process they are responsible for; and between experts and novices. There is nothing as depressing as doing the wrong thing well. If we do not communicate clearly what we want someone else to do, we should not be surprised when we don t get it. To demonstrate the need for improved communications, we often ask staff from one operation to identify their customers. We ask them to define what would make an excellent quality product/service for them to deliver to that customer. Then we ask the customer to define what an excellent quality product/service would be for them to receive from their supplier. Comparing these two definitions can be surprising. When you realize that the supplier does not understand what is truly important to their customer, it is easy to understand why we are always redoing work that was not done right the first time. There is an American expression that there is never time to do a pretest, but there is always time to redo the work. Communication is so fundamental that one of the first current best methods (CBMs) (Morganstein and Marker, 1997) that should be developed in any organization should deal with this topic. A CBM documents what is currently the best method known to the organization. While often called standard operating practices, we prefer the term CBM because it recognizes that these are living documents, requiring periodic update. They reflect knowledge of the organization, and are therefore not written by outside experts. CBMs should be short, frequently including a checklist of actions that even the most experienced staff member would benefit from reviewing. The rest of the CBM gives brief explanations for the items on the checklist, aimed at less experienced staff members, who need to understand why such a step is recommended. 8

16 CD-34-2 CBMs have been written for statistician programmer communications; project staff - statisticians/programmers communications; and, project staff - field staff communications. Communication issues include what to specify regarding data files, what output to check to confirm accuracy, what documentation to provide; standardizing result codes from field operations, and much more. A senior manager at a statistical agency complained that 50 percent of the tables I requested are wrong when first delivered. He then explained that this was not the fault of those preparing the tables, rather it was due to incomplete specification in the request. The appropriate weights, subpopulations to include, file versions, check totals for comparison, etc. were not clearly stated in the specifications. Worse, this happened repeatedly. They had not taken the time to develop a CBM for such requests, but had instead suffered repeated delays and potential uncaught errors, over and over. A different communication issue was demonstrated when the part of an NSO responsible for preparation of an annual database improved the database at the request of one if its users. Instead of dropping units that were no longer in business, they were retained on the database, but with a code indicting their ineligible status. This allowed for easier longitudinal analyses. Unfortunately, other users were not made aware of this change without reading the database documentation. They conducted analyses of the entire database assuming it hadn t changed. Luckily the resulting estimates were so surprising that the mistake was caught just before publication. The need for metadata that accurately communicates changes like this is becoming more and more vital as we make data available to a wider array of users. 6. Team work We have found that many of the most efficient quality improvements come from teams established to review and improve processes. The teams have a facilitator trained in CQI, include customers and suppliers of the process, and novices as well as experts. In this way people knowledgeable in the process inputs and outputs, along with those involved in the process itself, can select the best method for their processes. The team is encouraged to share its findings with others involved in the process. This improved communication provides valuable insight while gaining cooperation and trust from those not on the team. Such activities increase the likelihood that everyone will be supportive of the recommendations assembled by the team. 9

17 CD-34-2 Teams generally begin by understanding and sometimes sharpening the goals for their activity. These goals may be established by their managers, or through discussion with customers. The existing process is then mapped, to better understand the current way the process operates. Key processes are identified and measured, if possible. Improvements are made until a stable, reliable process is developed that meets the expectations that have been established. Quality assurance procedures are established to assure continued compliance with these expectations. Language and culture can have a major effect on the success of any program. We have determined this to be particularly true concerning teamwork. Some cultures are accustomed to working in teams, while others traditionally work individually. Also, the word team can be translated many ways in some languages. For example, in Finnish, the word tiimi implies a group where everyone s responsibilities are taken over by the team, with the individuals losing their personal responsibility. Alternatively, in a rhuma the individuals maintain responsibility, but work together to achieve their goals. It was important to translate team as a rhuma to connote the type of teamwork encouraged by this section. 7. Continuous improvement Continuous quality improvement means moving away from reliance on checking quality after delivery of products or services, to focusing on improving the processes that affected the resulting quality. There are many ways to focus on these processes. Many organizations make use of some of these, but few take advantage of the full range of opportunities that exist. Many official statistics are produced annually or monthly. Take the time to debrief project staff at the end of a cycle and prepare a short memo summarizing lessons learned. Are there things that could be done better, or activities that should be avoided? Such discussions can be a vital source of improvement efforts and provide excellent documentation for future staff. There are many stories of a new staff member taking over and receiving output from a process. Unfortunately, there is no documentation saying why this output is important and what should be reviewed. The new staff have to relearn everything, recreating the same errors discovered by previous staff. 10

18 CD-34-2 In the spirit of better communication and team work, often it is important to involve field staff. Many organizations do not regularly ask their interviewers for suggestions for improvements. Few activities are repeated more often by governmental statistical agencies than the interaction between interviewer and respondent. There is a tremendous opportunity for higher response rates and better cooperation rates through improved question wording, contact procedures, collection of information on nonrespondents, etc. Many key processes can be monitored to determine how they might be improved. These can vary from management budgeting procedures, to telephone interviewing information, to register updating procedures, to time and errors involved in producing reports. To the extent that such monitoring can be tracked numerically, tools such as control charts can be used to measure the effectiveness of improvements and accurately predict likely outcomes in the future. It is worth noting that monitoring, by itself, will not result in continuous improvements. Procedures must be established for responding to this information in a manner consistent with the numerical results. It is important to periodically develop and revise CBMs for key processes. As quality improvement teams determine best methods for achieving quality, these should be discussed, documented, and the relevant staff trained in their use. This will spread the knowledge that has been gained and encourage everyone to adopt these policies. As technology changes and other improvements are implemented, it becomes necessary to periodically re-examine old practices. As an example, at Westat we had excellent procedures for implementing the Mitofsky- Waksberg method for conducting random digit dialing (RDD) telephone surveys. However, improved lists of telephone numbers have become available, making alternative list-assisted approaches preferable to Mitofsky-Waksberg in most situations. We have therefore had to revise our CBM for conducting RDD surveys. 8. Summary Successful implementation of continuous quality improvement in a statistical agency involves a coordinated effort led by top management who demonstrate their commitment through their personal actions. They must establish a self-supporting structure that is flexible and values innovation so that the organization can continue to improve. Communications between levels of employees and between departments must be improved, so that barriers can be overcome and successes spread throughout the organization. Team work must be encouraged and rewarded so 11

19 CD-34-2 that everyone can understand how their actions fit into the larger organization, and will use that knowledge to produce better products and services. Finally, the goal must be to continuously improve. CQI is not meant to be a program, it should be the basic philosophy of how an organization learns. 9. References Dale, B. and Lascelles, D.M. (1994) "Levels of Total Quality Management Adoption," Managing Quality, edited by Barrie G. Dale, Prentice Hall, 2nd edition, 117. Deming, W. E., (1982) Quality, Productivity, and Competitive Position, MIT CAES, Cambridge, MA., Dillman, D. A. (1996). Why innovation is difficult in government surveys. Journal of Official Statistics. 12, Godrey, B., (1994) "Can Large Government Learn?" The Fifth Discipline Handbook, Doubleday, Goleman, D., (1985) Vital Lies and Simple Truths, Simon and Schuster, 234. Gross, W. and Linacre, S. (1997) "Improving the Comparability of Estimates Across Business Surveys." In Survey Measurement and Process Quality, ed. L. Lyberg et al., John Wiley & Sons., Kuhn, T., (1970) The Structure of Scientific Revolution, Chicago Press. Morganstein, D.R. and Marker, D.A. (1997). Continuous Quality Improvement in Statistical Agencies. In Survey Measurement and Process Quality, ed. L. Lyberg et al., John Wiley & Sons,

20 Improving Process Quality - From Control to Motivation Contribution to the International Conference on Quality in Official Statistics, May 14-15, 2001 in Stockholm Session Implementation of Quality Management III by Walter Radermacher and Thomas Körner (Federal Statistical Office Germany) 1. Introduction General Background The Method of Implementation: From Control to Motivation The Case of the Process Analyses in the Federal Statistical Office Germany Concluding Remarks Introduction The improvement of statistical products requires in most cases changes in the processes with which these products are produced. In this sense, improving quality almost in every instance has to make sure that processes are in good shape and continuously adapted to the requirements of the users, the agency, and the staff. But process improvements are important in still another respect: Most National Statistical Institutes (NSIs) have to take more assignments while the number of the staff is decreasing. These requirements necessitate process optimisation and rationalisation. Process optimisation often was not fully successful because of wrong strategies for implementation. Too much stress was put on central control and examination and too less room was given to the innovation and creativity of the staff. Improving the process quality of NSIs does not only involve documentation and standardisation. Also a climate of change and openness has to be created. This important function sometimes has been neglected by traditional tools for process optimisation. After some some general remarks on the current situation of official statistics and methods for the implementation of process optimisation schemes, we will show a new concept of process optimisation. The case in point used is a new scheme for analysing and improving processes developed by the Federal Statistical Office Germany. Finally, some lessons learned from the operation of the scheme will be named.

21 2. General Background Most European NSIs face the problem of having to do more things with less resources. On the one hand, budgets of NSIs are stable or (in most cases) decreasing. On the other hand, the demand for statistical information and for higher and higher quality standards introduces numerous new fields of activity. New assignments and challenges include: an increasing demand for statistical information increasing quality requirements for statistics (timeliness, accuracy, comparability etc.) rapidly changing fields of interest to be covered by official statistics (e.g. new demands concerning areas as the information society, e-business, biotechnology etc.) an increasing demand for European statistics (sometimes in addition to national requirements) new technological challenges of IT-use for production and dissemination of statistics a new demand for the use of the internet (a whole range of web-based solutions for surveys and information services) For these new activities arising for NSIs, new resources are needed. In most countries an increase of the budget will not be possible. Accordingly, NSIs will have to rationalise the fulfilment of the existing tasks in order to save money that could then be spent on the new challenges. One way to reach this aim is process optimisation. By an assessment of the Processes both an improvement of quality and a reduction of costs can be attained. 3. The Method of Implementation: From Control to Motivation As a first summary, we can state that for most European NSIs process optimisation is becoming a more and more important task. However, the implementation of such projects requires special attention. Besides the improvement of quality, rationalisation is a main task. In order to obtain good results and to put those results successfully into practice a clear strategy for implementation is indispensable. Within such a strategy, elements like decentralised responsibility and authority as well as motivation and participation of all members of the staff are essential elements. Let s briefly sketch such a strategy of implementation. First of all, process analysis and optimisation have to be carried out in a decentralised way. The units responsible for a task should be responsible for the analysis and improvement of the respective processes as well. Only with a decentralised responsibility, a culture of open-