Journal of Operations Management 30 (2012) 221–236 Contents lists available at SciVerse ScienceDirect Journal of Operations Management jo ur nal home page: www.elsev ier .co How to ce and pe Peter Let a RWTH Aachen b University of a r t i c l Article history: Received 29 Se Received in re Accepted 3 No Available onlin Keywords: Knowledge tra Learning Feedback Manufacturing task s rese edbac forma are t s of k on-fi r. Mo ning 1. Introdu Research has constituted that due to regular process changes and approaches in decentralization, responsibility is placed with individual workers to control work processes (e.g. Banker et al., 1993). This development results in a stronger demand for work- ers not only 1987) but t shop floor w conditions performanc 2003) when Ishikawa, 1 floor emplo ments – bot environmen Although of factors (e and Crossan the main co performanc relevant for nents have ∗ Correspon E-mail Marcus.Schwe (M. Zielinski). unde ception of the activities that a task requires, the sequence of the activities and the desired outcome. It helps the workers to acquire ‘know-why’ (Lapré et al., 2000) which is important in performing organizational processes and learning new tasks. This understanding is to a large extent driven by self-observation of 0272-6963/$ – doi:10.1016/j. to perform multiple tasks (e.g. Foster and Horngren, o continuously deal with new tasks. Learning among orkers and the acquisition of new knowledge are pre- for learning, continuous improvement and shop floor e (Torraco, 1999; Antonacopoulou, 2001; Enos et al., new tasks have to be carried out (Huang et al., 2008; 985; Upton and Kim, 1998). High learning rates of shop yees can translate into lower costs and quality improve- h highly relevant in today’s competitive manufacturing ts (Argote and Ingram, 2000). learning in an organization is stimulated by a number .g. Argyris and Schön, 1978; Fiol and Lyles, 1985; Bapuji , 2004), task understanding and skill development are mponents for learning a new task and improving task e (e.g. Baker, 1988). While both are regarded as being learning a new task, for an in-depth analysis the compo- to be disaggregated into their underlying mechanisms. ding author. addresses:
[email protected] (P. Letmathe),
[email protected] (M. Schweitzer),
[email protected] the shop floor workers through reflecting on their own working results (e.g. Manz and Sims, 1980), explicit knowledge transfer (e.g. Raelin, 1997; Nonaka and Takeuchi, 1995) and feedback which can lead to learning and continuous improvement (e.g. Aiello and Kolb, 1995; Chenhall, 1997). • Besides task understanding, employees require certain practical skills (‘know-how’, compare Lapré et al., 2000) to perform a task well. As can be seen from the literature (Nonaka and Takeuchi, 1995; Nonaka, 1991), these skills can be developed or learned over time. This learning might be the result of autonomous learn- ing processes (e.g. Lapré et al., 2000; Lapré and Van Wassenhove, 2001) or observing and imitating the skills of others, which are mechanisms of tacit knowledge transfer (Nonaka and Takeuchi, 1995). In this context, self-observation (stimulating task understand- ing) and autonomous learning (stimulating skill development) lead to automatically learning processes which cannot be deliber- ately controlled or influenced. Other components inducing learning (knowledge transfer and feedback) are based on concepts con- cerned with the formation of declarative and procedural knowledge that stimulate performance improvements (Bonner, 2008). see front matter © 2011 Elsevier B.V. All rights reserved. jom.2011.11.001 learn new tasks: Shop floor performan rformance feedback mathea,∗, Marcus Schweitzerb, Marc Zielinskib , Faculty of Business and Economics, Templergraben 64, 52062 Aachen, Germany Siegen, Hölderlinstr. 3, 57072 Siegen, Germany e i n f o ptember 2010 vised form 4 September 2011 vember 2011 e 20 November 2011 nsfer a b s t r a c t We investigate how learning and the knowledge transfer. Whereas previou edge transfer, self-observation and fe of these factors on learning and per reproducing manufacturing tasks that edge transfer is superior to other form in the form of cost information and n different forms of knowledge transfe significant performance effect on lear ction • Task m/locate / jom effects of knowledge transfer performance of individuals are affected by different forms of arch has proven the positive performance impacts of knowl- k mechanisms individually, we explore the cumulative effect nce. With the help of two laboratory experimental studies ypical for industrial production, we show that explicit knowl- nowledge transfer. Externally provided performance feedback nancial performance indicators has no effect on the order of reover, external feedback does not even have an additional new tasks irrespective of the type of knowledge transfer. © 2011 Elsevier B.V. All rights reserved. rstanding reflects a shop floor worker’s cognitive per- 222 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 With respect to the relevance of task understanding and skill development we build on previous research (e.g. Lapré et al., 2000) and compare the effects of tacit knowledge trans- fer and explicit knowledge transfer when autonomous learning and self-ob been regar knowledge environmen understand In particula effects of ou are helpful i back provid feedback ca the other ha can lead to fully occupi While so effects of kn 2003; Haas Sims, 1987) 1998), we t ing and ski manual tas terms of ‘kn The relation of special in standing as organizatio combinatio and skill de be stimulat manageme standing of can be used With res model, whi and test thi require the the tasks in ies we refin forms of kn refer to thre ing from kn forms of le Fig. 1. We will standing an new manu focus on th (1) Which autonom with re new tas (2) Does an the perf when le (3) Does an perform situatio observa Our resu most suitab load and li Performance ea rni oma ti rning nowle transf rning eedba edge learn the le ink i d kn tribu denc mou oor. rem disc ck on s hyp rform plica oret arnin prop mou serv ehav and S tand relation to a manufacturing process, self-observation in this t provides the basis for behavioral changes to meet perfor- goals and continuously improve (Manz and Sims, 1987). An yee observes himself/herself for example when filling in a on material consumption, required manufacturing time or . In this respect, the employee reviews the work process outcome. Reporting on one’s behavior and action during rk process enables employees to elaborate on experiential arn from the experience and reflect on their behavior as a dition for modifying work routines and for improving future ance (Chi et al., 1994). Along with other mechanisms, self- ation serves as an important strategy of self-management lf-leadership (Manz and Sims, 1980, 1981; Andrasik and erg, 1982). The literature states that self-leadership sup- he intrinsic motivation of an employee (Manz, 1986; Deci, 1975). Self-observation is pointed out as being a strategy blish intrinsic motivation as a key aspect for increasing nderstanding and task performance (Manz, 1986). This is servation are in place. As the role of feedback has ded as an effective instrument to induce declarative (Bonner, 2008) and is generally available in production ts, it is beneficial to study its influence to stimulate ing in the context of knowledge transfer mechanisms. r, it is beneficial to investigate the extent to which the tcome feedback (typically provided in manufacturing) n a setting of knowledge transfer. On the one hand, feed- es information for task understanding and unfavorable n help employees revise their actions and priorities. On nd, feedback information utilizes working memory and cognitive overload when cognitive capacity is already ed (Eppler and Mengis, 2004). me previous research has investigated the individual owledge transfer mechanisms (e.g. Edmondson et al., and Hansen, 2007), self-observation (e.g. Manz and or feedback (e.g. Banker et al., 1993; Sim and Killough, ake into account the cumulative effects of understand- ll development. This is appropriate as we refer to a k in manufacturing, requiring both understanding in ow-why’ and practical skills in terms of ‘know-how’. ship between skill development and understanding is terest because the formation of skills requires under- a promoter. The aim is to derive implications for the nal design of manufacturing and to state under which ns of the underlying mechanisms of task understanding velopment the learning of individual workers can best ed and which efforts are required from an organizational nt and human resources perspective. This better under- how knowledge transfer and feedback can be combined to strengthen and deepen learning processes. pect to individual workers we propose a task-learning ch is based on different forms of knowledge transfer, s model with two laboratory experimental studies that participants to learn a new task that is similar to many of manufacturing companies. With regard to other stud- e the learning theory as we base our model on typical owledge transfer in manufacturing. In the following we e learning mechanisms: learning automatically, learn- owledge transfer and learning from feedback. These arning form the basis of our model that is depicted in argue that not all components leading to task under- d skill development are equally important for learning facturing tasks and performance. In this context, we ree research questions: form of knowledge transfer in combination with both ous learning and self-observation is most beneficial spect to manufacturing performance when learning a k? additional performance feedback have an influence on ormance effect of different types of knowledge transfer arning a new task? additional performance feedback significantly increase ance when learning a new task in a manufacturing n where there is already knowledge transfer, self- tion and autonomous learning? lts clearly indicate that explicit knowledge transfer is le for learning new tasks. Due to information over- mitations of cognitive processing, combining explicit L aut Lea k Lea f knowl initial down that th tion an we con ing evi right a shop fl The section feedba derive and pe and im 2. The 2.1. Le As autono Self-ob own b (Manz unders ior. In respec mance emplo report defects and its the wo data, le precon perform observ and se Heimb ports t 1971, to esta task u Self-ob servati on Exp licit knowledge transfer External feedback Autono mou s lea rning Tac it kno wledge transfer Understand ing Skil l development ng ca lly fr om dge er fr om ck Fig. 1. Learning model. transfer with tacit knowledge transfer does not improve ing and additional feedback information can even slow arning process. This result is beneficial for many firms t is most beneficial to provide as much initial informa- owledge transfer as available. Since more is not better, te to the operations management literature by provid- e of the right type of knowledge transfer as well as the nt of information when new tasks are introduced to the ainder of this article is structured as follows. The next usses the effects of shop floor knowledge transfer and employees’ performance in more theoretical depth and otheses to be tested. Section three describes the design ance of our experimental studies. Research outcomes tions of our findings are discussed in the fourth section. ical basis and hypotheses development g automatically osed in our learning model, self-observation and s learning automatically stimulate learning processes. ation comprises the gathering of data referring to one’s ior and is a precondition for self-evaluation processes ims, 1980). With the help of such data an employee can the consequences but also the causes of one’s behav- P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 223 especially relevant in a manufacturing environment where the employee has the opportunity to identify with the products of the enterprise and to develop a feeling of motivation for doing a meaningful task (Ariely et al., 2008). While th which enha to the mech matically o learning by cesses (Dut Wassenhov Both sel establish le specific res worker per 2.2. Learnin In contr to be stim Rajagopalan in intra-org induced lea as an impo Levin, 2000 has argued edge among performed 1995). Alth barriers in t 2002), the s the type of 1998; Dave Polanyi ( tacit knowl ficult to re capabilities of individu action and i 1995). Obse (Tsoukas, 2 ate instrum with abilitie addressee ( flow from o and imitate 2003; Nona edge transf ‘know-how human min observation behavior in 2002; Szula et al., 2003) of the know observation Explicit von Krogh, independen objective an in words, nu 2003; Dave facturing p individual w mation exc Ferdows, 2 regarded as being more predictable than the transfer of tacit knowl- edge. Because it depends on the sharing of written documents or pictures, it does not necessarily require any direct personal inter- action. As a result, several authors argue that explicit knowledge is and port rnin te a s ts th hi, 1 arnin olars erfo or in or to ing u ing m rs is n egard oduc ially ip be rnal the c havi hav aniel t the erfo 001) us on anc nufa eline ed to y res ck re and g pe anc prov Luck have rmat ck in form earch mea e or non proc or to nt (Jö nt fo In ge red tasks y un Draw ve ta h, 1 gnit ot m ng pe e concept of self-observation is a cognitive process, nces task understanding, autonomous learning refers anical repetition of work. Autonomous learning auto- ccurs during work activities and may be regarded as doing as a consequence of a repetition of work pro- ton and Thomas, 1984; Lapré et al., 2000; Lapré and Van e, 2001). f-observation and autonomous learning automatically arning processes as they do not necessarily require any ources or efforts, but take place while an individual forms manufacturing tasks. g from knowledge transfer ast to learning automatically, induced learning has ulated through specific efforts or resources (Li and , 1998; Lapré et al., 2000). Such efforts can be seen anizational knowledge transfer that is manifested in rning (Özkan et al., 2007). Knowledge transfer serves rtant driver of learning in the workplace (Goh, 2002; ; Tsai, 2001; Argote and Ingram, 2000). The literature that enterprises have problems in transferring knowl- their employees especially when new tasks have to be (e.g. Gupta and Govindarajan, 2000; Zander and Kogut, ough the literature has discussed different factors and his context (e.g. Ives et al., 2002; Szulanski, 1996; Tsai, uccess of knowledge transfer is strongly influenced by knowledge to be transferred (e.g. Inkpen and Dinur, nport and Prusak, 2000; Zander and Kogut, 1995). 1966) distinguishes between two types of knowledge: edge and explicit knowledge. Tacit knowledge is dif- port and consists of employees’ skills, routines and (Skyrme, 2002) being stored in the heads and hands als (Leonard and Swap, 2004). It is deeply rooted in nvolvement in a specific context (Nonaka and Takeuchi, rvation and imitation of the way others perform tasks 003; Ribeiro and Collins, 2007) serve as an appropri- ent for transferring tacit knowledge from a person s relevant to the workplace (knowledge sender) to an knowledge receiver). Specifically, tacit knowledge may ne employee to another while the receiver observes s the sender of the knowledge transfer (Schröder, ka, 1991). Considering our learning model, tacit knowl- er aims at developing manual skills, i.e., enhancing ’. As tacit knowledge essentially only resides in the d (Polanyi, 1966) and because its transfer requires al processes and the transformation of the observed to the receiver’s own actions (Sole and Edmondson, nski, 1996), its success is hard to predict (Edmondson . In particular, limited cognitive processing capabilities ledge receiver imply a constraint to this learning from (e.g. Bandura, 1973). knowledge has a “universal character” (Nonaka and 2009, p. 636) and constitutes the codified and context- t type of knowledge (Edmondson et al., 2003). It is the d rational type of knowledge which can be expressed mbers, or symbols (Alavi and Leidner, 2001; Schröder, nport and Prusak, 2000). With reference to manu- rocesses explicit knowledge may be transferred to an ith the help of documents, manuals, photos or infor- hange through the intranet or email (Schröder, 2003; 006). The transfer of explicit, codified knowledge is easier Daven our lea to crea and fac Takeuc 2.3. Le Sch work p behavi behavi enhanc Report worke with r and pr potent tionsh an exte about and be studies (e.g. D contex place p Lind, 2 we foc perform ing ma Tim report cholog feedba vation learnin perform and im 1990). (1999) of info feedba be per Res mance outcom versus While behavi ronme sufficie 1990). structu set of ized b action. objecti Gingric high co does n facturi faster to transfer than tacit knowledge (Hansen, 1999; and Prusak, 2000; Zander and Kogut, 1995). As far as g model is concerned, explicit knowledge transfer helps ound understanding of tasks as it provides information at induce cognitive processes and thinking (Nonaka and 995). g from performance feedback in psychology have stated that feedback related to the rmance of individuals is an important factor of human organizations, as it provides insights and helps to adjust meet goals while activating cognitive processes and nderstanding (Goldberg, 1968; Stone and Stone, 1985). anufacturing performance information to shop floor ecessary for learning and stimulates employees’ efforts to the continuous improvement of work processes ts (Banker et al., 1993). Overall, performance feedback enables workers to recognize and understand the rela- tween their behavior and manufacturing outcomes from point of view (Baker, 1988), by providing information orrectness, accuracy, and adequacy of work outcomes ors (Earley et al., 1990). Based on these insights, many e investigated the effects of feedback on performance and Reitsperger, 1991; Sim and Killough, 1998). In this influence of several feedback-related factors on work- rmance has been studied (Luckett and Eggleton, 1991; . With reference to the research questions of this paper the three feedback-related factors ‘Timeliness’, ‘Type of e measures’ and ‘Goals’, which are important for report- cturing performance to shop floor workers. ss describes how often performance information is workers (Chenhall and Morris, 1986). In general, psy- earch has revealed that more frequent performance ported to individuals increases both employee moti- workplace performance (Ilgen et al., 1979). From a rspective, a more frequent reporting of manufacturing e measures to workers helps increase their learning rate es their performance more quickly (Locke and Latham, ett and Eggleton (1991) as well as Frederickson et al. specified these insights and argued that the frequency ion should match the manufacturing time cycle with the formation available when the same type of task has to ed again. has intensively analyzed different types of perfor- sures. Apart from the discussion of the relevance of process feedback, the stimulating effect of financial -financial performance measures has been explored. ess feedback provides insights into adjusting individual improve performance or reach goals in a complex envi- nsson and Grönlund, 1988), outcome feedback is more r well-learned or well-structured tasks (Earley et al., neral, manufacturing tasks on the shop floor are well- and employees are required to perform a well-defined . In this respect, the tasks are basically not character- known or uncertain alternatives, or consequences of ing on the approach to define complexity in terms of sk qualities (e.g. March and Simon, 1958; Campbell and 986), a typical manufacturing task does not demand a ive effort from the individual, i.e., problem solving, and eet the criteria of a complex task. In practice, manu- rformance feedback therefore often contains outcome 224 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 components to a high extent. Research and practice have discussed the disadvantage of solely using financial performance measures over a longer period of time (cp. for a review for example Neely et al., 1995; Davila and Wouters, 2007) with the result that in connec- tion with th Total Qualit financial pe anchored in mance mea or waste) a times) com as a guidan 2005). Research are closely reveals the clear and m mance agai use of stati mance stan (e.g. McNai prevented ( Based on formance fe and improv ing should b time cycle, formance m goals which 2.4. Hypoth knowledge t Shop flo back inform above stimu of individua of time. W sideration t setting cont ment. In ea knowledge takes place receive add and non-fin means that material co In answe type of kno task in ma experiment of different whether ad sented abov knowledge in addition and self-ob The role feedback ar declarative cedural kno continuous focuses on that invent p. 387). Pro ments desc it refers to a method on how to produce a certain product. Proce- dural and declarative knowledge result from knowledge transfer and focus on the acquisition of ‘know-how’ (skill development due to tacit knowledge transfer) and ‘know-why’ (understanding due licit tive ew k e is i e is rep serv prov g tex nsfe tive truct the e int nal r rec ansf epth manu it kno d ind sed b edge ing a e obs d the te ac lls w ate ( par owl rning expli s, 20 the w tacit y stru rpret tter ation on th cesse hen Alth tand h ex t and r dis cogn tacit rbed edge is eff ng th g of cove whe . acit to ex e rising of management approaches like Just-in-Time or y Management, the reporting of both financial and non- rformance measures to shop floor workers has been many enterprises (Lind, 2001). Non-financial perfor- sures like quality information (e.g. defect rates, rework s well as information concerning time (e.g. throughput plement financial performance measures and function ce to detect areas of improvements (Maiga and Jacobs, has further emphasized that goal-setting and feedback connected (Earley et al., 1990; Erez, 1977). As feedback extent to which goals have been achieved, it is more otivating if an employee can compare his actual perfor- nst a given budget or goal (Earley et al., 1990). While the c budgets has been widely criticized, historical perfor- dards are regarded as a suitable option for setting goals r, 1990; Lind, 2001) as long as the Ratchet effect can be Indjejikian and Nanda, 1999). these findings, we can draw the conclusion that per- edback to shop floor workers that stimulates learning ement and that provides a basis for task understand- e provided frequently according to the manufacturing contain financial as well as non-financial outcome per- easures, and compare actual performance measures to are grounded in historical performance. eses on the performance effects of different types of ransfer or knowledge transfer, self-observation as well as feed- ation in accordance with the requirements discussed late learning effects and the continuous improvement ls resulting in increasing performance over the course ith respect to our research topics we take into con- wo experimental settings in manufacturing, with each aining components of understanding and skill develop- ch experimental study, we focus on a situation where is transferred to individuals and active self-observation . In the second experimental study, the participants itional performance feedback based on both financial ancial performance measures. Active self-observation all participants were required to record task time and nsumption to ensure a reflection of their own work. r to the first research question, i.e., to investigate which wledge transfer is superior to others in learning a new nufacturing, we derive identical hypotheses for both al settings. In this respect, we assume that the effects forms of knowledge transfer are identical regardless of ditional performance feedback is given or not. As pre- e, we additionally have to take into consideration that transfer leading to induced learning always takes place to learning automatically from autonomous learning servation. of learning from knowledge transfer and learning from e based on the concepts concerning the formation of and procedural knowledge. Both declarative and pro- wledge are regarded to stimulate induced learning and performance improvements. Declarative knowledge facts and “state a description, such as the information ory is equal to a 100 books” (Kogut and Zander, 1992, cedural knowledge on the other hand consists of state- ribing a process (Kogut and Zander, 1992), for example to exp declara their n practic Practic to task self-ob The readin The tra declara edge s When ing th additio transfe edge tr an in-d into a Tac tion an discus knowl observ that th ior an concre the ski or imit Com tacit kn the lea when Ferdow use of when alread to inte new pa inform taxing be pro vant w 1990). unders throug of taci transfe due to tion of is distu knowl sate th Usi learnin which of time output H1. T pared knowledge transfer). People gradually integrate new knowledge with existing knowledge while encoding nowledge. Declarative knowledge is best acquired when n place and knowledge is instructional (Bonner, 2008). provided with the help of learning automatically due etition from autonomous learning (Reber, 1995) and ation. ision of instructional knowledge is best achieved from ts or diagrams, i.e., the instructions should be codified. r of explicit knowledge is beneficial for the formation of knowledge as it can directly provide a desired knowl- ure through an outline of the structure (Bonner, 2008). transferred knowledge contains information concern- eraction of the learner with the task to be executed, procedural knowledge is acquired by the knowledge eiver (Woolfolk, 1998; Gredler, 2005). Explicit knowl- er is beneficial here because explicit knowledge induces task understanding which can be directly transferred facturing action. wledge is largely transferred by observation and imita- uces observational learning, which has been primarily y Bandura (1973). In this respect, the receiver of the transfer has to process the information provided while model, i.e., the sender. Observational learning requires erver has the capacity to reproduce the model’s behav- ability to transform the information provided into tion (Nadler et al., 2003). This is a critical process as hich constitute tacit knowledge are difficult to pass on Ferdows, 2006; Kogut and Zander, 1992). ing the transfer of explicit knowledge to the transfer of edge with respect to learning a new manufacturing task, process is faster and manifests in a higher performance cit knowledge is transferred (Zander and Kogut, 1995; 06). The learning of new tasks in particular makes full orking memory’s capacity which is more demanding knowledge is transferred. Whereas explicit knowledge ctures information items and provides patterns of how information, tacit knowledge involves the formation of ns and it is more difficult to prioritize different items of . Therefore, tacit knowledge transfer is generally more e working memory and initially not all information can d properly. An information overload can be more rele- explicit and tacit knowledge are combined (Schick et al., ough tacit knowledge transfer can lead to a deeper task ing in the long run, we predict that learning a new task plicit knowledge transfer is superior to a combination explicit knowledge transfer because tacit knowledge turbs the concentration on explicit knowledge transfer itive capacities required for the observation and imita- knowledge transfer. When explicit knowledge transfer and task-understanding slows down, additional tacit transfer inducing skill development does not compen- ect when learning a new task. e previous arguments and considering the workplace shop floor workers, we derive the following hypotheses, r the individuals’ total performance for a given period n the individuals manufacture a pre-defined amount of knowledge transfer induces a lower performance com- plicit knowledge transfer P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 225 (a) when self-observation and autonomous learning take place. (b) when self-observation and autonomous learning take place and external performance feedback is given. H2. Tacit k pared to a c transfer (a) when se (b) when se externa H3. A com edge trans explicit kno (a) when se (b) when se externa We also knowledge no transfer rience auto we assume autonomou formance c alone due sophisticate resulting fro et al., 2000 autonomou working m to improve transfer. H4. Auton performanc self-observa (a) when n (b) when ex H5. Auton performanc self-observa (a) when n (b) when ex H6. Auton performanc self-observa knowledge (a) when n (b) when ex All hypo given perio priate, as in may interru cient length knowledge ruptions. In improveme insights into task learning processes. Empirical research on learn- ing curves has proven that autonomous and induced learning result in continuous improvements over the course of time (e.g. Adler and Clark, 1991; Mukherjee et al., 1998; Lapré et al., 2000). To pro- rthe mou nt ty type e co perfo atio wled ck in onti addit ed. addit ed an addit ed an addit ted nsfer onti addit ed an addit ed, ta dbac add ula man addit ted, nsfer espo exter actur lf-ob n th edb anu e fe indi form feed and r, 2 d be utcom then ue t Balze yees eting ve lo ove ult t ue to nowledge transfer induces a lower performance com- ombination of tacit knowledge and explicit knowledge lf-observation and autonomous learning take place. lf-observation and autonomous learning take place and l performance feedback is given. bination of tacit knowledge transfer and explicit knowl- fer induces a lower performance compared to sole wledge transfer lf-observation and autonomous learning take place. lf-observation and autonomous learning take place and l performance feedback is given. explore the impact on the performance improvement of transfer compared to a manufacturing situation where of knowledge takes place and employees only expe- nomous learning and self-observation. In this respect, that any kind of knowledge transfer in addition to s learning and self-observation induces a higher per- ompared to autonomous learning and self-observation to the fact that knowledge transfer stimulates more d cognitive learning processes than learning processes m learning-by-doing (Dutton and Thomas, 1984; Lapré ). From the cognitive perspective, we predict that s learning and self-observation do not fully utilize emory. The remaining cognitive capacity can be used performance due to explicit and/or tacit knowledge omous learning and self-observation induce a lower e compared to a combination of autonomous learning, tion and tacit knowledge transfer o external performance feedback is provided. ternal performance feedback is given. omous learning and self-observation induce a lower e compared to a combination of autonomous learning, tion and explicit knowledge transfer o external performance feedback is provided. ternal performance feedback is given. omous learning and self-observation induce a lower e compared to a combination of autonomous learning, tion as well as tacit knowledge transfer and explicit transfer o external performance feedback is provided. ternal performance feedback is given. theses formulated refer to the cumulative output of a d of time when work interruptions occur. This is appro- almost all work situations, interruptions occur and pt learning processes if there is a break of any suffi- (Nembhard and Osothsilp, 2001). A repeated transfer of may overcome forgetfulness resulting from work inter- this context, it is also interesting to explore gradual nts over the course of time, as these provide in-depth vide fu autono differe which over th about combin of kno feedba H7. C when (a) in lat (b) in lat (c) in lat (d) in ula tra H8. C when (a) in lat (b) in lat fee (c) in stim for (d) in ula tra To r tional manuf and se sions o from fe In m outcom mance in the that if ments (Bonne require from o rect ‘if- rules d noise ( emplo interpr cogniti mation is diffic load d r insights into continuous improvements, we analyze s learning effects combined with learning generated by pes of knowledge transfer. This enables us to compare of knowledge transfer induces stronger learning effects urse of time. In line with our theoretical consideration rmance differences, we formulate hypotheses for each n of autonomous learning in addition to different forms ge transfer and either sole self-observation or external addition to self-observation. nuous improvements over the course of time occur ion to autonomous learning, self-observation is stimu- ion to autonomous learning, self-observation is stimu- d tacit knowledge is transferred. ion to autonomous learning, self-observation is stimu- d explicit knowledge is transferred. ion to autonomous learning, self-observation is stim- and a combination of tacit and explicit knowledge is red. nuous improvements over the course of time occur ion to autonomous learning, self-observation is stimu- d external performance feedback is provided. ion to autonomous learning, self-observation is stimu- cit knowledge is transferred and external performance k is provided. ition to autonomous learning, self-observation is ted, explicit knowledge is transferred and external per- ce feedback is provided. ion to autonomous learning, self-observation is stim- a combination of tacit and explicit knowledge is red and external performance feedback is provided. nd to the third research question, i.e., whether an addi- nal feedback significantly increases performance in a ing situation where there is already knowledge transfer servation, we derive a hypothesis on the basis of discus- e formation of declarative and procedural knowledge ack. facturing, feedback is generally provided in terms of edback. Basically, feedback compares actual perfor- cators to budgets. The provision of feedback manifests ation of declarative knowledge. The literature argues back is to be of importance for continuous improve- learning processes it must be cognitive and procedural 008). To achieve this, process feedback is generally cause it is difficult to acquire procedural knowledge e feedback alone. The reason is that people infer incor- ’ rules from outcomes as they permanently change their o varying outcomes or the outcomes are affected by r et al., 1989). When outcome feedback is provided, the are requested to make behavioral adjustments while the feedback which is difficult. This induces a high ad on working memory potentially resulting in infor- rload (Hahn et al., 1992; Sutcliffe and Weick, 2008) that o transform into manual work action. Information over- feedback specifically occurs when the utilization of 226 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 the working memory is high. This is relevant for our study since learning a new task, i.e., establishing a procedural structure that is not already automated, occupies the working memory to a high degree. The literature shows that feedback is more beneficial in sit- uations whi Although w therefore re beneficial a This is o induce addi in place. Th directly foc that by kno structured i ory which Atkinson e observation for knowled as a suppor Addition tion overlo feedback ca sic motivat because of additional formance i self-efficien helps stimu We ther manufactur is best supp learning ph practice do tribute to th Taking o pose that s mainly driv learning an of output fr the perform performanc H9. The p without ou (a) in addit lated. (b) in addit lated an (c) in addit lated an (d) in addit ulated a transfer 3. Design o To test t mental stud of using stu 1946; Pete as an altern field study, ling extran validity. Fu tions mana are supposed to hold in the laboratory (Bendoly et al., 2006). Differences between students and experienced employees occur when expert knowledge (i.e., the recognition of patterns for faster information processing) accelerates learning and improves perfor- . Wit oor e fect i ons t Erics gle ansf n lea t. cond s stud 0 und ariou w, bu ontro ents ot ad h exp ed al sim we nd w unica lly, i , tran ad to ad to as eq h ele , this l ind ptio ditio edge imp ring, dent ation quir ad to l shee he co xpec g th gram lity. T rticip ual p perim fir mou les. T pend ablis in a nic c on th oups er g ubje ch are not overloaded with information (O’Reilly, 1980). e only test learning in the initial learning phase, it is asonable to assume that feedback information is more fter the basic task procedure has been internalized. ne reason why additional outcome feedback does not tional performance when knowledge transfer is already rough knowledge transfer the employees attention is used on the work processes to be learned which means wledge transfer the relevant information is basically n a manner that reduces the load in the working mem- is of high importance (e.g. Ward and Sweller, 1990; t al., 2000). The literature has also shown that self- as a strategy of self-influence is an important driver ge acquisition (e.g. Politis, 2005) that can be regarded ting instrument in the context of knowledge transfer. ally, as employees are overtaxed due to informa- ad under a situation of additional feedback, external n place limitations on self-influence and reduce intrin- ion (Ilgen et al., 1979). If self-influence is limited external performance feedback, we do not expect any benefit from outcome feedback for learning or per- mprovement, when self-observation that stimulates cy (e.g. Manz and Sims, 1980) is already in place and late intrinsic motivation. efore propose that for learning a manual task on the ing shop floor the formation of procedural knowledge orted by a transfer of explicit knowledge. In the initial ase, additional outcome feedback as often provided in es not induce any further benefits as it does not con- e formation of procedural knowledge. ur learning model into consideration, we generally pro- hop floor learning and continuous improvement are en by the three factors: self-observation, autonomous d knowledge transfer. With respect to the total amount om a given period of time we therefore do not expect ance to change significantly when additional outcome e feedback is provided. erformance effects are the same in situations with or tcome performance feedback when ion to autonomous learning, self-observation is stimu- ion to autonomous learning, self-observation is stimu- d tacit knowledge is transferred. ion to autonomous learning, self-observation is stimu- d explicit knowledge is transferred. ion to autonomous learning, self-observation is stim- nd a combination of tacit and explicit knowledge is red. f laboratory experimental studies he hypotheses, we performed two laboratory experi- ies with undergraduate students. The appropriateness dents has been discussed for a long time (e.g. McNemar, rson, 2001), and field studies have been suggested ative with respect to external validity. Conducting a however, is associated with the problems of control- eous disturbances and would imply reduced internal rthermore, the use of students is accepted in opera- gement to test general theories because accurate ones mance shop fl The ef situati 1992; of a sin edge tr focus o relevan We nomic with 4 from v ness la were c of stud were n Eac perform identic session gram a comm manua diodes they h they h pant w for eac ments severa interru precon knowl The ufactu depen observ time re They h specia it and t were e meetin cuit dia in qua that pa individ 3.1. Ex The autono variab the de we est to fill electro Based four gr dents p were s hout a doubt, such expert knowledge is relevant when mployees learn tasks that are similar to previous ones. s mainly faster learning and better decision making in hat a subject is very familiar with (Mackay and Elam, son and Charness, 1994). However, we are not aware study which has shown that different forms of knowl- er are less favorable for experts. In our experiment, we rning a new task where expert knowledge is even less ucted the first experiment with 28 undergraduate eco- ents (16 males, 12 females) and the second experiment ergraduate economics students (27 males, 13 females) s economic disciplines (business administration, busi- siness and engineering, economics). Both experiments lled in terms of the previous manufacturing experience , gender, age and the subject of study. Repeating subjects mitted. erimental study comprised of four sessions, which were biweekly. In each session the students had to build four ple electronic circuits in a serial production. After each collected all the material provided, e.g., the circuit dia- ork instructions, which could have served as a basis for tion between the sessions. The students had to work .e., to solder five electronic components (light emitting sistors, resistors, diodes, battery clips) onto pins, which place on a foam board beforehand. To perform the task, make use of a soldering iron and tweezers. Each partici- uipped with the same amount of soldering wire (60 cm) ctronic circuit. In terms of complexity and skill require- task is similar to many manufacturing procedures in ustries. With such a realistic task structure and work ns, we created a setting with high external validity as a n for learning from the real effects of different types of transfer and for testing our hypotheses. ortant indicators of individual performance in man- product quality and manufacturing time, served as variables in both experiments. To ensure an active self- , all participants had to measure the manufacturing ed as well as the remaining amount of soldering wire. report on these performance measures and fill out a t for each electronic circuit on the time required to build nsumption of soldering wire. Additionally, the students ted to self-observe their performance with respect to e quality standards of their boards. Because of the cir- provided it was possible to self-observe any deviations he measurement and documentation process ensured ants dealt actively and in a standardized way with their erformance outcomes. ental design of the first study st experiment included active self-observation, s learning and knowledge transfer as independent o evaluate the impacts of the independent variables on ent variables (output quality and manufacturing time), hed four groups. For this purpose the students had questionnaire asking for their previous knowledge on omponents and circuits as well as their manual skills. e results of the survey, the students were assigned to by parallelization and randomization, with seven stu- roup. Depending on the group assignment, the students cted to different forms of knowledge transfer, while all P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 227 participants had to actively self-observe their task performance. In particular they had to check their manufactured circuit boards for quality defects and control the required manufacturing time. Prior to the start of the experiments, the students received basic information as video-instructions on their task within the exper- iment and on how to solder the components. Additionally, they were instructed that each participant had to build four electronic circuits per week and 16 circuits in total. They were told that they had to follow a lexicographical order of goals, i.e., quality first, time second. Furthermore, the test persons were told to strictly carry out the instructions: Test persons in group (S,A) had to per- form active self-observation and acted as the control-group, they did not receive any kind of knowledge transfer. They simply had to use a circuit diagram (Appendix A) with a legend representing the components’ graphical symbols (Appendix B) and the related forms of the components. With the help of the legend, the test per- sons were able to autonomously learn about the correct place for each component. In addition to autonomous learning (use of the circuit diagram with the legend), the participants of the second group (S,A, allowing th their group and imitati resulting fr and (S,A,T), use a writte represented cuit diagram of where to tronic circu of documen ded in quali group (S,A,T (S,A,T) and group (S,A,T and explici knowledge codification explicit kno the knowle only consid imitation b not spend a the design o At the en ated by the following c penalty poi wrong place nent with c for each wi Table 1 Experimental Group (S,A) (S,A,T) (S,A,E) (S,A,T,E) assigned. Following the scope of this experiment, the participants were not given any feedback on their performance. 3.2. Experim The desi that of the fi feedback w (S,A,F), (S,A included bo provided co in the diffe tion conten firms (Fishe over, the fe beneficial i nization lite In this y poi y po nic c uct cuit gned y of f wir unte ition ring t as a sis o orke , the ering the b dget actur defe 00D. time and d wa he b ment his o of qu wire l as ced onta ced o dere sold lity, m mpa ailed form viou uld b nform quen ntain d no ment T) were able to make use of tacit knowledge transfer em to observe and to imitate the other test persons of . We limited tacit knowledge transfer to observation on to control for and rule out any confounding effects om any other interaction. In contrast to groups (S,A) the students of the third group (S,A,E) were allowed to n document (construction manual, Appendix C) which a transfer of explicit knowledge. In addition to the cir- and the legend, the manual contained a detailed chart place the components, a sample-photo of the elec- it and a short description of the work steps. This type tation resembles work instructions, which are embed- ty management systems. The test persons of the fourth ,E) were requested to share tacit knowledge like group to use the written document. The participants of the ,E) therefore experienced a combined transfer of tacit t knowledge. For all groups the time required for the transfer was included in the manufacturing time. As the of the knowledge was already done, the time to transfer wledge comprises the time the receiver needs to extract dge. With respect to tacit knowledge transfer, we also ered the receiving process in terms of observation and ecause the sender (i.e., the observed participant) does ny time on the transfer process. Table 1 summarizes f the experimental study. d of the experiment each electronic circuit was evalu- instructors in terms of product quality, applying the riteria: If a circuit board was free from defects, no nts were assigned. For each component mounted in the , a penalty of four points was assigned. For each compo- onnection wires mixed up, a penalty of two points and re not mounted correctly, a penalty of one point was design of experimental study 1. Form of knowledge transfer Knowledge transfer instrument Self-observation, autonomous learning, no transfer of knowledge – Self-observation, autonomous learning, transfer of tacit knowledge Observation and imitation Self-observation, autonomous learning, transfer of explicit knowledge Construction manual Self-observation, autonomous learning, transfer of tacit and explicit knowledge Observation and imitation, construction manual penalt penalt electro of prod If a cir be assi penalt nection not mo Add ufactu served the ba floor w Finally of sold On the bu manuf tion of was 0. turing 2.03D, a boar 2.13D. At t experi about terms dering as wel introdu back c mispla not sol tion of to qua and co in a det back in the pre that co evant i the fre and co cial an experi ental design of the second study gn of the second experimental study was the same as rst study, except that additional outcome performance as given. Therefore, we had another four groups, namely ,T,F), (S,A,E,F) and (S,A,T,E,F). The feedback information th financial as well as non-financial information and mprehensive information on the subjects’ performance rent sessions. The type of indicators and the informa- t is typical for shop floor feedback in manufacturing r, 1992; Fry et al., 1998; Ittner and Larcker, 2002). More- edback design that we implemented is considered to be n the management accounting and management orga- rature (Lillis, 2002). context, the participants were instructed about the nts assigned for the defects as well as the price of a int that was 4.00D. The students were told that each ircuit would be evaluated by the instructors in terms quality whereby the following criteria would be used: board was free from defects, no penalty points would . For each component mounted in the wrong place, a our points was assigned. For each component with con- es mixed up, a penalty of two points and for each wire d correctly, a penalty of one point was assigned. ally, the students were informed that in terms of man- ime the benchmark to be achieved was 3.5 min, which historical standard. They were also informed that on f estimated mean costs per hour of 35.00D for a shop r the price per minute of manufacturing time was 0.58D. y were informed that the standard for the consumption wire was set at 10 cm, with 1 cm priced at 0.01D. asis of this information the participants were told about s that were set for each electronic circuit board to be ed: in terms of the primary objective, i.e., a minimiza- cts, the budget costs for each electronic circuit board For the secondary objective (minimization of manufac- ) the budget costs for each electronic circuit board were the budget costs for the soldering wire to be used for s 0.10D. Therefore the total budget for each board was eginning of the second, third and fourth sessions of the each student received a feedback report (Appendix D) r her manufacturing performance of the last event in ality, manufacturing time, and the consumption of sol- . Financial and non-financial information was provided the participants differences compared to the budgets before the experiment started. The non-financial feed- ined information about the amount of components n the circuit board, the amount of wires mixed up or d properly, the manufacturing time and the consump- ering wire. Additionally, the actual costs with reference anufacturing time and soldering wire were presented red against the benchmark. The feedback was provided way as the feedback report contained not only the feed- ation for each board but also the cumulative output of s event. The feedback was presented in such a manner e quickly and easily understood and contained the rel- ation for the participants (see Appendix D). Moreover, cy of the feedback matched the manufacturing cycle ed both types of performance measures, namely finan- n-financial ones. Table 2 summarizes the design of the al study. 228 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 Table 2 Experimental design of experimental study 2. Group Form of knowledge transfer Knowledge transfer instrument (S,A,F) (S,A,T,F) (S,A,E,F) (S,A,T,E,F) 4. Results 4.1. Descrip Descript quality and and Table 4 in the week of 7.1% that missing dat imputed. Descript improved t ity as well Figs. 2 and (circuits 1– 13–16) in t amounted t group (S,A, 20.0% [39.2 ferences be of knowled example th compared t 96.5% in the When co ence betwe the course o gaps are sti group (S,A,E the first ses weakest gro the ranking quality and all sessions The resu the second total perfor well as in te paring the fi by 75.9% [4 70.0% [48.2 by 88.1% [4 In accord explicit kno In relative t the weakes to quality [ fourth sessi 4.2. Discussion of results To test our hypotheses, we first compared the groups within udy. As far as the first study was concerned, significant per- ce d (�2 = ng ti nces sts w all 0.091 .960, an in h ex ed cted the of kn st stu udy. not ased pure ies, t mpor ). Th t kn con mo eses in h edge ielde r (hy of ex actur utpe eses r of e tra trans a hig red t er, a owle tuati acit 6a. first nmen r of wled Self-observation, autonomous learning, no transfer of knowledge, outcome feedback – Self-observation, autonomous learning, transfer of tacit knowledge, outcome feedback Observation and imitation Self-observation, autonomous learning, transfer of explicit knowledge, outcome feedback Construction manual Self-observation, autonomous learning, transfer of tacit and explicit knowledge, outcome feedback Observation and imitation, construction manual and discussion tive statistics ive statistics with respect to the performance indicators manufacturing time are reported in Table 3 (first study) (second study). In the first study we had missing data before the last session of 3.5% and in the last session was statistically imputed. In the second study we had a from the last session of 15%, which was statistically ive statistics for the first study indicate that all groups heir total performance measured in terms of qual- as in terms of manufacturing time (compare also 3). When comparing the first session of experiments 4) with the fourth session of experiments (circuits erms of quality [time], the improvement of group (S,A) o 56.8% [38.0%], group (S,A,T) achieved 50.5% [58.9%], E) improved by 66.7% [47.7%] and group (S,A,T,E) by %]. Although all groups improved on the whole, the dif- tween the groups are pronounced with the availability ge transfer as a major driver of quality performance. For e relative advantage of the best group (S,A,E) was 95.6% o the weakest group (S,A) in the first session and even fourth session. mparing the mean times in absolute terms, the differ- en all groups from the first experiment decreased over f time. However, when we consider relative differences, ll evident. Whereas the relative advantage of the best ) compared to the weakest group (S,A,T) was 28.6% in sion, the relative advantage of (S,A,E) compared to the up (S,A,T,E) in session four was still 18.8%. Comparing and the absolute performance measures in terms of time, group (S,A,E) outperformed the other groups in of the study. lts of the first study are supported by the outcomes of each st forman found facturi differe man te among (�2 = 2 (�2 = 9 For for bot perform condu shows forms the fir ond st priori were b 2007). For activit high i (S,A,E) explici finding edge is hypoth results knowl only y transfe vance manuf edge o hypoth transfe than th bined induce compa howev and kn to a si when t ses 4a– The enviro transfe of kno experimental study. Again, all groups improved their mance measured in terms of quality (see also Fig. 4) as rms of manufacturing time (see also Fig. 5) when com- rst session with the fourth one. Group (S,A,F) improved 5.6%] with respect to quality [time], group (S,A,T,F) by %], group (S,A,E,F) by 72.2% [46.7%] and group (S,A,T,E,F) 0.3%]. ance with the first study, the group that solely received wledge outperformed the other groups in all sessions. erms for example, the advantage of group (S,A,E,F) over t group (S,A,F) amounted to 91.96% [19.7%] with respect time] in the first session and to 90.7% [21.4%] in the on. of the seco edge transf The results of hypothe learning, se autonomou different re the group r nificant her is strongly time. While the second knowledge ifferences for the performance indicator quality were 11.029, **p = 0.012) as well as for the indicator manu- me (�2 = 8.486, **p = 0.037). These outcomes prove that between all groups of the first study existed. The Fried- ith respect to the second study also revealed differences groups of this study, both for the indicator quality , ***p = 0.000) and for the indicator manufacturing time **p = 0.019). -depth analysis, we tested hypotheses 1–6 separately perimental studies. For each performance indicator we a Friedman test. In additional non-parametric tests we pair-wise group comparisons for each study. Table 5 results of the Wilcoxon tests comparing the different owledge transfer and reflection by paired groups of dy, while Table 6 presents the outcomes of the sec- Against the background of the hypotheses specified a requiring ˛-adjustment (Bortz et al., 2008), the tests on a ‘per comparison risk level’ of ˛ = 0.05 (Rasch et al., autonomous learning and self-observation during work he results for the performance indicator quality reveal tance for the transfer of explicit knowledge (group e results for hypotheses 1a–3a show that transferring owledge outperforms tacit knowledge transfer. This firms our argument in Section 2.4 that tacit knowl- re difficult to transfer than explicit knowledge. Taking 4a–6a into consideration, a transfer of knowledge igher performance compared to a situation where no transfer takes place. However, significant results were d for the group (S,A,E) receiving explicit knowledge potheses 5a). Furthermore, the results show the rele- plicit knowledge transfer for the performance indicator ing time. As before, the transfer of explicit knowl- rforms the transfer of tacit knowledge (see results for 1a and 3a). Furthermore, we can conclude that the tacit knowledge takes up more manufacturing time nsfer of explicit knowledge. As a consequence, a com- fer of tacit and explicit knowledge transfer does not her performance in terms of manufacturing time when o the transfer of explicit knowledge only. Interestingly combination of autonomous learning, self-observation dge transfer predicts significant advantages compared on of autonomous learning and self-observation only knowledge alone is transferred (see results for hypothe- experimental study showed that in a manufacturing t with persons actively self-observing their work, the explicit knowledge basically outperforms other forms ge transfer. Compared to the first study the results nd study support the superiority of explicit knowl- er because the findings are even slightly stronger here. for hypotheses 1b–3b are in accordance with those ses 1a–3a. The advantage of combined autonomous lf-observation and knowledge transfer compared to s learning, and self-observation alone shows somewhat sults. In connection with additional feedback, results for eceiving tacit and explicit knowledge transfer are sig- e with respect to quality. Furthermore, hypothesis 5b confirmed for the performance parameters quality and the results for both studies are basically in accordance, study reveals an even stronger importance of explicit transfer. P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 229 Table 3 Descriptive results of experimental study 1. Groups Session 1 circuits 1–4 Session 2 circuits 5–8 Session 3 circuits 9–12 Session 4 circuits 13–16 (S,A) MQ: 4.71 (4) MQ: 2.18 (4) MQ: 2.11 (4) MQ: 2.04 (4) MT: 12:46 (2) MT: 9:38 (2) MT: 8:56 (2) MT: 7:55 (3) (S,A,T) MQ: 3.82 (3) MQ: 1.82 (3) MQ: 1.32 (3) MQ: 1.89 (3) MT: 17:30 (4) MT: 11:30 (4) MT: 9:32 (4) MT: 7:11 (2) (S,A,E) MQ: 0.21 (1) MQ: 0.18 (1) MQ: 0.14 (1) MQ: 0.07 (1) MT: 12:30 (1) MT: 9:05 (1) MT: 7:29 (1) MT: 6:32 (1) (S,A,T,E) MQ: 0.89 (2) MQ: 0.71 (2) MQ: 0.96 (2) MQ: 0.71 (2) MT: 13:14 (3) MT: 11:05 (3) MT: 9:12 (3) MT: 8:03 (4) S: self-observation, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit knowledge, MQ: average penalty points per circuit board, MT: average manufacturing time per circuit board reported in minutes:seconds. The numbers in brackets provide the rankings of the groups in the categories ‘quality’ and ‘manufacturing time’. Table 4 Descriptive results of experimental study 2. Groups Session 1 circuits 1–4 Session 2 circuits 5–8 Session 3 circuits 9–12 Session 4 circuits 13–16 (S,A,F) MQ: 5.60 (4) MQ: 2.98 (4) MQ: 1.23 (3) MQ: 1.35 (4) MT: 14:33 (4) MT: 10:31 (3) MT: 8:37 (3) MT: 7:55 (4) (S,A,T,F) MQ: 3.58 (3) MQ: 1.40 (3) MQ: 1.58 (4) MQ: 1.08 (3) MT: 13:55 (3) MT: 11:18 (4) MT: 8:26 (2) MT: 7:13 (2) (S,A,E,F) MQ: 0.45 (1) MQ: 0.10 (1) MQ: 0.33(1) MQ: 0.13 (1) MT: 11:41 (1) MT: 7:50 (1) MT: 6:39 (1) MT: 6:14 (1) (S,A,T,E,F) MQ: 1.68 (2) MQ: 0.55 (2) MQ: 0.35 (2) MQ: 0.20 (2) MT: 12:58 (2) MT: 10:03 (2) MT: 9:01 (4) MT: 7:45 (3) S: self-observation, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit knowledge, F: feedback, MQ: average penalty points per circuit board, MT: average manufacturing time per circuit board reported in minutes: seconds. The numbers in brackets provide the rankings of the groups in the categories ‘quality’ and ‘manufacturing time’. -6 -5 -4 -3 -2 -1 0 16151413121110987654321 M ea n Q ua lit y Quality Group S,A Group S,A,T Group S,A,E Group S,A,T,E Fig. 2. Comparison of mean quality (experimental study 1). 0:00:00 0:03:36 0:07:12 0:10:48 0:14:24 0:18:00 0:21:36 0:25:12 16151413121110987654321 M ea n T im e Time Group S,A Group S,A,T Group S,A,E Group S,A,T,E Fig. 3. Comparison of mean time (experimental study 1). 230 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 -1 0 16151413121110987654321 Quality Table 5 Comparison o Hypothesis 1a 2a 3a 4a 5a 6a S: self-observa ** Significant *** Significant To comp and therefo to Tables 7 the Friedma course of t tor ‘quality that group autonomou assigned ta -5 -4 -3 -2 M ea n Q ua lit y -8 -7 -6 Fig. 4. Comparison of mean quality (experimen 0:00:00 0:03:36 0:07:12 0:10:48 0:14:24 0:18:00 0:21:36 0:25:12 1110987654321 M ea n T im e Time Fig. 5. Comparison of mean time (experimen f the different forms of knowledge transfer. Group comparison Dependent variable “quality” Z p (one (S,A,T) ≺ (S,A,E) −2.201 0.014* (S,A,T) ≺ (S,A,T,E) −2.023 0.022* (S,A,T,E) ≺ (S,A,E) −1.947 0.026* (S,A) ≺ (S,A,T) −0.338 0.368 (S,A) ≺ (S,A,E) −2.371 0.009* (S,A) ≺ (S,A,T,E) −1.183 0.119 tion, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit know on a level of � = 0.050. on a level of � = 0.010. are the gradual improvements over the course of time re in order to respond to hypotheses 7 and 8, we refer and 8. With reference to the first experimental study, n test indicates high significant improvements over the ime for group (S,A) only for the performance indica- ’. This result can be explained against the background (S,A) received no transfer of knowledge and performed s learning and self-observation. When performing the sks the first time round, the test persons of this group concentrate did not foc a conseque changed ov routine. Th ual, but no group (S,A, (S,A,E) and performanc Group S,A,F Group S,A,T,F Group S,A,E,F Group S,A,T,E,F tal study 2). 1615141312 Group S,A,F Group S,A,T,F Group S,A,E,F Group S,A,T,E,F tal study 2). Dependent variable “time” -tailed) Z p (one-tailed) * −2.197 0.014** * −1.183 0.119 * −2.197 0.014** −1.859 0.032** ** −1.014 0.155 −0.169 0.433 ledge, (≺): is inferior to. d on putting the components in the right place and us their attention on securely attaching the wires. As nce, these test persons received high penalties. This er the course of time as the task became more of a e results for the groups (S,A,T) and (S,A,E) show grad- significant improvements, while the performance of T,E) did not change at all. We observed that groups (S,A,T,E) assembled the circuits with a high level of e the whole time. The transfer of explicit knowledge P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 231 Table 6 Comparison of the different forms of knowledge transfer in combination with self-observation and additional feedback. Hypothesis Group comparison Dependent variable “quality” Dependent variable “time” Z p (one-tailed) Z p (one-tailed) 1b 0.003 2b 0.011 3b 0.008 4b 0.384 5b 0.003 6b 0.011 S: self-observa know ** Significant *** Significant Table 7 Continuous im Hypothesis tailed 7a * 7b 7c 7d S: self-observa know *** Significant Table 8 Continuous im Hypothesis -tailed 8a ** 8b * 8c 8d ** S: self-observa ** Significant *** Significant therefore in in situation the mistake the course With respe the Friedma Based on th course of ti ongoing im The resu the results introduced, (S,A,F), but feedback sp pared to a si is in place. I ments part conclude th tacit and e improveme ments over knowledge test persons beginning w As far a concerned, the first stu The chi-squ of the seco (S,A,T,F) ≺ (S,A,E,F) −2.807 (S,A,T,F) ≺ (S,A,T,E,F) −2.295 (S,A,T,E,F) ≺ (S,A,E,F) −2.402 (S,A,F) ≺ (S,A,T,F) −0.296 (S,A,F) ≺ (S,A,E,F) −2.805 (S,A,F) ≺ (S,A,T,E,F) −2.296 tion, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit on a level of � = 0.050. on a level of � = 0.010. provements (experimental study 1). Group Dependent variable “quality” �2 p (two- (S,A) 34.957 0.002** (S,A,T) 16.610 0.343 (S,A,E) 12.522 0.639 (S,A,T,E) 6.673 0.966 tion, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit on a level of � = 0.010. provements (experimental study 2). Group Dependent variable “quality” �2 p (two (S,A,F) 70.134 0.000* (S,A,T,F) 29.288 0.015* (S,A,E,F) 17.949 0.265 (S,A,T,E,F) 34.849 0.003* tion, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit know on a level of �= 0.050. on a level of � = 0.010. duces a high product quality from the beginning while s where no transfer of explicit knowledge takes place s made are quite high at the beginning, declining over of time, mainly due to effects of ‘learning-by-doing’. ct to the performance indicator manufacturing time, n tests indicate highly significant results for all groups. is result, we can draw the conclusion that over the me, strong learning effects take place, which result in provements concerning the factor time. lts of the first experimental study are consistent with of the second study. Interestingly when feedback was gradual improvements did not only occur in group also in groups (S,A,T,F) and (S,A,T,E,F). An additional eeds up learning and continuous improvement, com- tuation where knowledge transfer and self-observation n this respect, feedback supports performance improve- icularly when tacit knowledge is transferred. We can at feedback can partly fill the performance gap between xplicit knowledge transfer with reference to gradual nts. Group (S,A,E,F) showed no incremental improve- the course of time proving that with the help of explicit transfer, autonomous learning and self-observation the of this group already reach a very high quality from the hich can hardly be improved. s the performance indicator manufacturing time is the Friedman tests are consistent with the results of dy indicating highly significant results for all groups. are values are not remarkably higher for the groups nd study compared to their counterparts of the first study excep Group (S,A improveme the descrip of experime indicating t advantage t Finally, hypothesis were able t with refere formed in e comparison parisons (e two-tailed p difference b (S,A,F) with We can ment, self- autonomou performanc Through se transfer, em the cause-a Although ad ufacturing period of ti if tacit kno for the perf *** −2.701 0.004*** ** −0.561 0.288 *** −2.090 0.019** −0.255 0.400 *** −2.599 0.005*** ** −0.357 0.361 ledge, F: feedback, (≺): is inferior to. Dependent variable “time” ) �2 p (two-tailed) 52.680 0.000*** 81.040 0.000*** 69.249 0.000*** 69.370 0.000*** ledge. Dependent variable “time” ) �2 p (two-tailed) 94.457 0.000*** 83.405 0.000*** 78.864 0.000*** 74.192 0.000*** ledge, F: feedback. t for those groups receiving no transfer of knowledge. ,F) outperformed group (S,A) with respect to gradual nts in the manufacturing time required. Referring to tive analysis, however, we find that in the last session nts the manufacturing time was equal for both groups, hat an additional feedback seems to induce a short-term hrough learning a new task faster. in answer to the second research question and to test 9, we conducted Mann–Whitney-tests (Table 9). We o directly compare the matching groups of both studies nce to the cumulative output of the four sessions per- ach experiment. Again, the tests were based on a ‘per risk level’ ˛ = 0.05. The results show for all group com- xcept one) that we cannot reject the hypotheses. The -values indicate that there is no statistically significant etween matched groups except for groups (S,A) and respect to manufacturing time. conclude that in a typical manufacturing environ- observation and knowledge transfer in addition to s learning effects induce learning effects and increase e when the employees have to learn a new task. lf-observation and in particular through knowledge ployees receive the information needed to understand nd-effects of their actions and therefore to improve. ditional feedback speeds up the learning of new man- tasks, referring to the total performance for a given me, feedback is only effective on a low significant level wledge transfer is in place, and in this respect only ormance indicator manufacturing time. On the whole, 232 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 Table 9 Comparison of both experimental studies. Hypothesis Group Comparison Dependent variable “quality” Dependent variable “time” d) 9a 9b 9c 9d S: self-observa know * Significant we have ve effects of au respect to a In general, a higher perf is transferr autonomou in enterpris encourage a comes in or that learnin formance fe performanc ments after time. 5. Conclus The outc transfer of learning an observation with respec assembly ti transfer ind tialization o creation of also shown cally learnin and feedba with the fin able effects where no k any signific for hypothe of performa When know that are alr actions and induce fast mance effe conclude th mainly driv while outco and continu Consiste studies hav task, workin knowledge back can le learning pr knowledge explicit kno ation bett hich ation s du se re o the o far that edge difie d by nstru redu duce t kno ed w ve au en n nal r nsfe stem mot ors a uous mpli perim nufa expe equa of pr ironm ions ry, w ation ecen be c oug anc t kno ded. , mo serv lore Z p (two-taile (S,A) = (S,A,F) −1.175 0.270 (S,A,T) = (S,A,T,F) −0.147 0.887 (S,A,E) = (S,A,E,F) −0.550 0.601 (S,A,T,E) = (S,A,T,E,F) −0.293 0.813 tion, A: autonomous learning, T: transfer of tacit knowledge, E: transfer of explicit on a level of � = 0.100. rified the proposed effects concerning the cumulative tonomous learning, self-observation and feedback with dditional knowledge transfer and learning new tasks. n additional feedback does not automatically induce a ormance in a manufacturing setting when knowledge ed and employees self-observe their actions and learn sly. This implies that decision makers and managers es should establish knowledge transfer activities and ctive self-observation of their work behavior and out- der to improve performance. As we were able to show g new tasks does not require sophisticated external per- edback, it should be further analyzed whether or not e feedback is helpful to achieve incremental improve- a task has been implemented for a longer period of ions omes of the experimental studies have shown that the explicit knowledge in combination with autonomous d self-observation or with autonomous learning, self- and additional outcome feedback is most beneficial t to manufacturing performance in terms of quality and me when a new task is introduced. Explicit knowledge uces a high performance level directly after the ini- f a new manufacturing task because it stimulates the ‘know-why’, i.e., task understanding. The findings have that knowledge transfer in combination with automati- g processes as well as automatically learning processes ck result in strong gradual improvements. Comparable dings from Arzi and Shtub (1997) we found no remark- of work interruptions in this context. Even in the groups nowledge transfer took place, the data did not show ant effects from the work interruptions. The results ses 9a–9d provide interesting insights into the effects nce improvement of an additional outcome feedback. ledge is continuously transferred to those employees eady learning autonomously and self-observing their behavior, an additional outcome feedback does not er learning of a new task or other significant perfor- cts. Taking our learning model into consideration, we at task understanding among shop floor employees is en by self-observation and explicit knowledge transfer, me feedback does not contribute to learning a new task inform always ture w autom pattern The cable t (and s show knowl this co learne work i design and re explici provid and ha Wh nizatio and tra back sy should behavi contin these i our ex the ma of the are ad terms an env operat contra codific and a d should Alth perform explici is nee cerned self-ob to exp ous improvement to a high extent. nt with our theoretical reasoning, our experimental e demonstrated that for learning a new semi-complex g memory is absorbed by self-observation and explicit transfer. Additional tacit knowledge transfer or feed- ad to information overload and can even slow down ocesses. Furthermore, our results indicate that tacit transfer is more taxing on working memory than wledge transfer. These findings are consistent with the Our analysi manufactur in combina prove whet One limitat observe lea tal techniq about whic how inform Z p (two-tailed) −0.293 0.813 −1.952 0.055* −1.561 0.133 −0.488 0.669 ledge, F: feedback, (=): is equal to”. overload theory which has shown that ‘more is not er’. Therefore, firms should design an information struc- is adjusted to the limitations of working memory, the of task execution over time and the ability to recognize e to past experience. sults show that cognitive processing theories are appli- manufacturing shop floor with some highly relevant often neglected) practical implications. Our findings managers should encourage the creation of explicit on the manufacturing shop floor and the transfer of d knowledge, in particular when new tasks have to be the employees. Due to potential information overload, ction should be short, step-by-step and precise. This ces the time to process the explicit knowledge provided s the utilization of working memory. More detailed wledge as well as tacit knowledge and feedback can be hen employees have learned the task’s basic structure tomated parts of the task execution procedure. ew tasks have to be learned on the shop floor, orga- esources should therefore be allocated to the creation r of explicit knowledge instead of implementing feed- s. Our findings additionally reveal that decision makers ivate their employees to actively self-observe their work nd outcomes in order to learn new tasks faster and ly improve. We have to take into consideration that cations particularly apply to a stable environment. In ental settings we created such a stable environment as cturing environment was not changed over the course riment. In this case manuals and written documents te to transfer codified knowledge (Ferdows, 2006). In actical implications, Ferdows (2006) shows that in such ent central staff should codify the knowledge, develop manuals or incorporate it in technical systems. On the hen the production ‘know-how’ changes quickly, the and transfer processes have to be carried out quickly tralized approach for a systematic change of knowledge hosen (Ferdows, 2006). h our findings imply a high relevance for individual e in manufacturing and point out the importance of wledge transfer and self-observation, further research As far as individual shop floor workers are con- re research on self-management aspects such as ation is needed. In particular, it would be interesting the influence of different levels of task complexity. s has focused on a well-structured and semi-complex ing task and revealed the benefits of self-observation tion with explicit knowledge transfer. Research could her these results also hold in a more complex setting. ion of our experiment is that we could not directly rning processes which would require other experimen- ues, i.e., fMRI studies that could provide information h brain areas are used during learning processes and ation overload might disturb learning. The effects of P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 233 group effects and in this respect the analysis of performance obser- vation and knowledge transfer processes in work-teams might be another interesting field for future research. Although the transfer of tacit knowledge through social interaction did not outperform explicit knowledge transfer with respect to individual performance in our experiments, in a team setting, social interaction and tacit knowledge transfer may be more beneficial. In this context, the improvement effects of employees’ workshops that emphasize per- sonal interaction could be investigated and the behavioral motives of the employees could be studied, for example on the basis of further experimental studies in combination with survey measures. Appendix A. Circuit diagram Appendix B Appendix C. Manual com ins o olde com pin com ove t the d. . Legend • Bend • Put p • Put s • Hold • Heat • Place • Rem • Hold geale ponents (with/without tweezers). n the foam board. r on the pins. ponent at the pin. with the help of your soldering iron. ponent with the help of tweezers into the solder. he soldering iron. component with the tweezers until the solder is con- 234 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 Appendix D. Feedback report Person Primary Board 1 Board 2 Board 3 Board 4 Sum Seconda Board 1 Board 2 Board 3 Board 4 Sum Tertiary Board 1 Board 2 Board 3 Board 4 Sum Total cos Board 1 Board 2 Board 3 Board 4 Sum References Adler, P.S., Cla process. M Aiello, J.R., Ko text: impa 339–353. Alavi, M., Leidn agement s (1), 107–1 Andrasik, F., H (Ed.), Hand 219–247. Antonacopoul between 327–350. Argote, L., Ing tage in fir 150–169. Argyris, C., Sch Ariely, D., Ka case of L 671–677. Arzi, Y., Shtub comparati Atkinson, R., instruction cational Re Baker, E.M., F.M. (Eds pp. 10.1–1 Balzer, W.K., D performan Bandura, A., 1 Cliffs. Banker, R.D., P measures Research 5 Bapuji, H., Cro research. M objective: minimization of defects Components at wrong place Wires mixedup Wires not mounted correctly Other defects Penalty points Price per penalty point Actual costs Budget costs Difference 4.00D 0.00 D 4.00D 0.00D 4.00D 0.00D 4.00D 0.00D 0.00D ry objective: minimization of manufacturing time. Actual time [Min] Price per minute Actual costs Budget costs Difference 0.58D 2.03D 0.58D 2.03D 0.58D 2.03D 0.58D 2.03D 8.12D objective: minimization of material consumption. Actual consumption [cm] Price per cm Actual costs Budget costs Difference 0.01D 0.10D 0.01D 0.10D 0.01D 0.10D 0.01D 0.10D 0.40D ts Defects Time Soldering wire Total actual costs Total budget costs Difference 2.13D 2.13D 2.13D 2.13D 8.52D rk, K.B., 1991. Behind the learning curve: a sketch of the learning anagement Science 37 (3), 267–281. lb, K.J., 1995. Electronic performance monitoring and social con- ct on productivity and stress. Journal of Applied Psychology 80 (3), er, D.E., 2001. Review: knowledge management and knowledge man- ystems: conceptual foundations and research issues. MIS Quarterly 25 36. eimberg, J.S., 1982. Self-management procedures. In: Frederikson, L.M. book of Organizational Behavior Management. Wiley, New York, pp. ou, E.P., 2001. The paradoxical nature of the relationship training and learning. Journal of Management Studies 38 (3), ram, P., 2000. Knowledge transfer: a basis for competitive advan- ms. Organizational Behavior and Human Decision Processes 82 (1), ön, D.A., 1978. Organizational Learning. Addison-Wesley, Reading. menica, A., Prelec, D., 2008. Man’s search for meaning: the egos. Journal of Economic Behavior & Organization 67 (3–4), , A., 1997. Learning and forgetting in mental and mechanical tasks: a ve study. IIE Transactions 29 (9), 759–768. Derry, S., Renkl, A., Wortham, D., 2000. Learning from examples: al principles from the worked examples research. Review of Edu- search 70 (2), 181–214. 1988. Managing human performance. In: Juran, J.M., Gryna, .), Juran’s Quality Control Handbook. McGraw-Hill, New York, 0.61. oherty, M.E., O’Connor Jr., R., 1989. Effects of cognitive feedback on ce. Psychological Bulletin 106 (3), 410–433. 973. Aggression: a social learning analysis. Prentice-Hall, Englewood otter, G., Schroeder, R.G., 1993. Reporting manufacturing performance to workers: an empirical study. Journal of Management Accounting , 33–55. ssan, M., 2004. From questions to answers: reviewing organizational anagement Learning 35 (4), 397–417. Bendoly, E., Donohue, K., Schultz, K.L., 2006. Behavior in operations management: assessing recent findings and revisiting old assumptions. Journal of Operations Management 24 (6), 737–752. Bonner, S.E., 2008. Judgment and Decision Making in Accounting. Pearson Education, Upper Saddle River. Bortz, J., Lienert, G.A., Boehnke, K., 2008. Verteilungsfreie Methoden in der Biostatis- tik, 3. Auflage. Springer, Heidelberg. Campbell, D.J., Gingrich, K., 1986. The interactive effects of task complexity and participation on task performance: a field experiment. Organizational Behavior and Human Decision Processes 38 (2), 162–180. Chenhall, R.H., 1997. Reliance on manufacturing performance measures, total qual- ity management and organizational performance. Management Accounting Research 8 (2), 187–206. Chenhall, R.H., Morris, D., 1986. The impact of structure, environment, and inter- dependence on the perceived usefulness of management accounting systems. Accounting Review 61 (1), 16–35. Chi, M.T.H., de Leeuw, N., Chiu, M.H., Lavancher, C., 1994. Eliciting self- explanations improves understanding. Cognitive Science 18 (3), 439–477. Daniel, S.J., Reitsperger, W.D., 1991. Linking quality strategy with management control systems: empirical evidence from Japanese industry. Accounting, Orga- nizations and Society 16 (7), 601–618. Davenport, T.H., Prusak, L., 2000. Working Knowledge: How Organizations Manage What They Know. Harvard Business School Press, Boston. Davila, T., Wouters, M., 2007. Management accounting in the manufacturing sector: managing costs at the design and production stages. In: Chapman, C.S., Hop- wood, A.G., Shields, M.D. (Eds.), Handbook of Management Accounting Research, vol. 2. Elsevier, Amsterdam, pp. 831–858. Deci, E.L., 1971. Effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology 18 (1), 105–115. Deci, E.L., 1975. Notes on the theory and metatheory of intrinsic moti- vation. Organizational Behavior and Human Performance 15 (1), 130–145. Dutton, J.M., Thomas, A., 1984. Treating progress functions as a managerial oppor- tunity. Academy of Management Review 9 (2), 235–247. Earley, P.C., Northcraft, G.B., Lee, C., Lituchy, T.R., 1990. Impact of process and out- come feedback on the relation of goal setting to task performance. Academy of Management Journal 33 (1), 87–105. P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 235 Edmondson, A.C., Winslow, A.B., Bohmer, R.M.J., Pisano, G.P., 2003. Learning how and learning what: effects of tacit and codified knowledge on perfor- mance improvement following technology adoption. Decision Sciences 34 (2), 197–223. Enos, M.D., Kehrhahn, M.T., Bell, A., 2003. Informal learning and the transfer of learning: h Quarterly Eppler, M.J., M literature disciplines Erez, M., 1977 relationsh Ericsson, K.A., American Ferdows, K., 2 Operation Fiol, C.M., Lyl Review 10 Fisher, J., 1992 ment 6 (1) Foster, G., Hor Manageme Frederickson, effects of p feedback f Fry, T.D., Steel tems in m 503–525. Goh, S.C., 200 work and 23–30. Goldberg, L.R., judgment. Gredler, M., 20 Education Gupta, A.K., Go rations. St Haas, M.R., Ha productivi agement J Hahn, M., Law load on de Hansen, M.T., knowledge (1), 82–11 Huang, X., Kris cess imple Manageme Ilgen, D.R., Fis behavior i Indjejikian, R., Journal of Inkpen, A.C., D joint ventu Ishikawa, K., 1 Englewood Ittner, C.D., La worker inc Ives, W., Torre Morey, D. Classic and Jönsson, S., G and mana 512–532. Kogut, B., Zand replication Lapré, M.A., M curve: link (5), 597–6 Lapré, M.A., V for produ 1311–132 Leonard, D., Sw Levin, D.Z., 20 tigation of Li, G., Rajagop Manageme Lind, J., 2001. Manageme Locke, E., Latha New York. Lillis, A.M., 20 Accountin Luckett, P.F., E of research ety 16 (4), Mackay, J.M., Elam, J.J., 1992. A comparative study of how experts and novices use a decision aid to solve problems in complex knowledge domains. Information System Research 3 (2), 150–172. Maiga, A.S., Jacobs, F.A., 2005. Antecedents and consequences of quality perfor- mance. Behavioral Research in Accounting 17, 111–131. .C., 19 esses 600. .C., Si cial le 367. .C., S on or 113. .C., Sim ership 106–1 ., Simo C.J., onsibi stry 4 ar, Q., 374. jee, A. rovem ., Thom owle ., Greg ature uction rd, D. Trans I., 199 04. I., Tak panie . I., vo rovers nizati .F., G uct an Colleg C.A., ssaril , R.A. hts fr 450–4 M., 19 .D., 20 ings o 458. .A., 19 578. ., Ver en un ., 1995 . R., Co s of ac .G., Go untin r, K.A. senstr s. De ., Kill een m anage D.J., 2 ip. In agem 3. Edmo sh Jou .L., S favor racy. 185. , K.M. Starb ing. O i, G., pract R.J., of wo 270. ow managers develop proficiency. Human Resource Development 14 (4), 369–387. engis, J., 2004. The concept of information overload: a review of from organization science, accounting, marketing, MIS, and related . Information Society 20 (5), 325–344. . Feedback: a necessary condition for the goal setting-performance ip. Journal of Applied Psychology 62 (5), 624–627. Charness, N., 1994. Expert performance: its structure and acquisition. Psychologist 49 (8), 725–747. 006. Transfer of changing production know-how. Production and s Management 15 (1), 1–9. es, M.A., 1985. Organizational learning. Academy of Management (4), 803–813. . Use of nonfinancial performance measures. Journal of Cost Manage- , 31–38. ngren, C.T., 1987. JIT: cost accounting and cost management issues. nt Accounting 68 (12), 19–25. J.R., Peffer, S.A., Pratt, J., 1999. Performance evaluation judgments: rior experience under different performance evaluation schemes and requencies. Journal of Accounting Research 37 (1), 151–167. e, D.C., Saladin, B.A., 1998. The use of management accounting sys- anufacturing. International Journal of Production Research 36 (2), 2. Managing effective knowledge transfer: an integrative frame- some practice implications. Journal of Knowledge Management 6 (1), 1968. Simple models or simple process? Some research on clinical American Psychologist 23 (7), 483–496. 05. Learning and Instruction: Theory into Practice, 5th ed. Pearson , Upper Saddle River. vindarajan, V., 2000. Knowledge flows within multinational corpo- rategic Management Journal 21 (4), 109–122. nsen, M.T., 2007. Different knowledge, different benefits: toward a ty perspective on knowledge sharing in organizations. Strategic Man- ournal 28 (11), 1133–1153. son, R., Lee, Y.G., 1992. The effects of time pressure and information cision quality. Psychology and Marketing 9 (5), 365–378. 1999. The search-transfer problem: the role of weak ties in sharing across organization subunits. Administrative Science Quarterly 44 1. tal, M.M., Schroeder, R.G., 2008. Linking learning and effective pro- mentation to mass customization capability. Journal of Operations nt 26 (6), 714–729. her, C.D., Taylor, M.S., 1979. Consequences of individual feedback on n organizations. Journal of Applied Psychology 64 (4), 349–371. Nanda, D., 1999. Dynamic incentives and responsibility accounting. Accounting and Economics 27 (2), 177–201. inur, A., 1998. Knowledge management processes and international res. Organization Science 9 (4), 454–468. 985. What is Total Quality Control? The Japanese Way. Prentice-Hall, Cliffs. rcker, D.F., 2002. Determinants of performances measure choices in entive plans. Journal of Labor Economics 20 (2), S58–S90. y, B., Gordon, C., 2002. Knowledge sharing is a human behavior. In: , Maybury, M., Thuraisingham, B. (Eds.), Knowledge Management: Contemporary Works. MIT Press, Cambridge, pp. 99–129. rönlund, A., 1988. Life with a sub-contractor: new technology gement accounting. Accounting, Organizations and Society 13 (5), er, U., 1992. Knowledge of the firm, combinative capabilities, and the of technology. Organization Science 3 (3), 383–397. ukherjee, A.S., Van Wassenhove, L.N., 2000. Behind the learning ing learning activities to waste reduction. Management Science 46 11. an Wassenhove, L.N., 2001. Creating and transferring knowledge ctivity improvement in factories. Management Science 47 (10), 5. ap, W., 2004. Deep smarts. Harvard Business Review 82 (9), 88–97. 00. Organizational learning and the transfer of knowledge: an inves- quality improvement. Organization Science 11 (6), 630–647. alan, S., 1998. Process improvement, quality, and learning effects. nt Science 44 (11), 1517–1532. Control in world class manufacturing – a longitudinal case study. nt Accounting Research 12 (1), 41–74. m, G., 1990. Goal Setting Theory and Task Performance. Prentice Hall, 02. Managing multiple dimensions of manufacturing performance. g, Organizations and Society 27 (6), 497–529. ggleton, I.R.C., 1991. Feedback and management accounting: a review into behavioural consequences. Accounting, Organizations and Soci- 371–394. Manz, C proc 585– Manz, C a so 361– Manz, C ing 105– Manz, C lead (1), March, J McNair, resp Indu McNem 289– Mukher imp Nadler, J of kn Neely, A liter Prod Nembha IEEE Nonaka, 96–1 Nonaka, Com York Nonaka, cont Orga Özkan, G prod ogy, O’Reilly, nece Peterson insig (3), Polanyi, Politis, J ceed 449– Raelin, J 563– Rasch, D such Reber, A York Ribeiro, type Schick, A Acco Schröde Wis team Sim, K.L betw of M Skyrme, ersh Man 61–8 Sole, D., Briti Stone, D back accu 167– Sutcliffe G.P., Mak Szulansk best Torraco, role 249– 86. Self-Leadership: toward an expanded theory of self-influence in organizations. Academy of Management Review 11 (3), ms Jr., H.P., 1980. Self-management as a substitute for leadership: arning theory perspective. Academy of Management Review 5 (3), ims Jr., H.P., 1981. Vicarious learning: the influence of model- ganizational behavior. Academy of Management Review 6 (1), s Jr., H.P., 1987. Leading workers to lead themselves: the external of self-managing work teams. Administrative Science Quarterly 32 28. n, H., 1958. Organizations. Wiley, New York. 1990. Interdependence and control: traditional vs. activity-based lity accounting. Journal of Cost Management in the Manufacturing , 15–24. 1946. Option-attitude methodology. Psychological Bulletin 43 (4), S., Lapré, M.A., Van Wassenhove, L.N., 1998. Knowledge driven quality ent. Management Science 44 (11), S35–S49. pson, L., van Boven, L., 2003. Learning negotiation skills: four models dge creation and transfer. Management Science 49 (4), 529–540. ory, M., Platts, K., 1995. Performance measurement system design: a review and research agenda. International Journal of Operations and Management 15 (4), 80–116. A., Osothsilp, N., 2001. An empirical comparison of forgetting models. actions on Engineering Management 48 (3), 283–291. 1. The knowledge creating company. Harvard Business Review 69 (6), euchi, H., 1995. The Knowledge-Creating Company: How Japanese s Create the Dynamics of Innovation. Oxford University Press, New n Krogh, G., 2009. Tacit knowledge and knowledge conversion: y and advancement in organizational knowledge creation theory. on Science 20 (3), 635–652. aimon, C., Kavadias, S., 2007. Knowledge management strategies for d process design teams. Working paper, Georgia Institute of Technol- e of Management. 1980. Individual and information overload in organizations: is more y better? Academy of Management Journal 23 (4), 684–696. , 2001. On the use of college students in social science research: om a second-order meta-analysis. Journal of Consumer Research 28 61. 66. The tacitdimension. Routledge and Kegan Paul Ltd., London. 05. Dispersed leadership predictor of knowledge acquisition. In: Pro- f the 2nd European Conference on Knowledge Management, Bled, pp. 97. A model of work-based learning. Organization Science 8 (6), dooren, L.R., Gowers, J.I., 2007. Planung und Auswertung von Ver- d Erhebungen, 2. Auflage. Oldenbourg, München. . The Penguin Dictionary of Psychology, 2nd ed. Penguin Books, New llins, H., 2007. The bread-making machine: tacit knowledge and two tion. Organization Studies 28 (9), 1417–1433. rdon, L.A., Haka, S., 1990. Information overload: a temporal approach. g, Organizationsand Society 15 (3), 199–220. , 2003. Mitarbeiterorientierte Gestaltung des unternehmensinternen ansfers: Identifikation von Einflussfaktoren am Beispiel von Projekt- utscher Universitäts-Verlag, Wiesbaden. ough, L.N., 1998. The performance effects of complementarities anufacturing practices and management accounting systems. Journal ment Accounting Research 10, 325–346. 002. Developing a knowledge strategy: from management to lead- : Morey, D., Maybury, M., Thuraisingham, B. (Eds.), Knowledge ent: Classic and Contemporary Works. MIT Press, Cambridge, pp. ndson, A., 2002. Situated knowledge and learning in dispersed teams. rnal of Management 13 (Special issue), S17–S34. tone, E.F., 1985. The effects of feedback consistency and feed- ability on self-perceived task competence and perceived feedback Organizational Behavior and Human Decision Processes 36 (2), , Weick, K.E., 2008. Information overload revisited. In: Hodgkinson, uck, W.H. (Eds.), The Oxford Handbook of Organizational Decision xford University Press, Oxford, pp. 56–75. 1996. Exploring internal stickiness: impediments to the transfer of ice within the firm. Strategic Management Journal 17, 27–43. 1999. Integrating learning with working: a reconception of the rkplace learning. Human Resource Development Quarterly 10 (3), 236 P. Letmathe et al. / Journal of Operations Management 30 (2012) 221–236 Tsai, W., 2001. Knowledge transfer in intraorganizational networks: effects of network position and absorptive capacity on business unit innovation and per- formance. Academy of Management Journal 44 (5), 996–1004. Tsai, W., 2002. Social structure of “coopetition” within a multiunit organization: coordination, competition, and intraorganizational knowledge sharing. Organi- zation Science 13 (2), 179–190. Tsoukas, H., 2003. Do we really understand tacit knowledge? In: Easterby-Smith, M., Lyles, M. (Eds.), The Blackwell Handbook of Organizational Learning and Knowledge Management. Blackwell, Oxford, pp. 410–427. Upton, D.M., Kim, B., 1998. Alternative methods of learning and process improve- ment in manufacturing. Journal of Operations Management 16 (1), 1–20. Ward, M., Sweller, J., 1990. Structuring effective worked examples. Cognition and Instruction 7 (1), 1–39. Woolfolk, A., 1998. Educational Psychology, 9th ed. Pearson Education, Upper Saddle River. Zander, U., Kogut, B., 1995. Knowledge and the speed of transfer and imitation of organizational capabilities: an empirical test. Organization Science 6 (1), 76–92. How to learn new tasks: Shop floor performance effects of knowledge transfer and performance feedback 1 Introduction 2 Theoretical basis and hypotheses development 2.1 Learning automatically 2.2 Learning from knowledge transfer 2.3 Learning from performance feedback 2.4 Hypotheses on the performance effects of different types of knowledge transfer 3 Design of laboratory experimental studies 3.1 Experimental design of the first study 3.2 Experimental design of the second study 4 Results and discussion 4.1 Descriptive statistics 4.2 Discussion of results 5 Conclusions Appendix A Circuit diagram Appendix B Legend Appendix C Manual Appendix D Feedback report References