The percentage of time spent in each of the five training categories (didactic, discussion, simulation, participatory and non-training) based on the total time for the particular institute was summarized separately for each type of training institute. These included two education component types, mentor and teacher/mentor institutes, and five management component types, director, first management team, component coordinators, and second management team institutes, as well as the on-site follow-up for the subsample of programs that participated in this activity. Table 1 presents the mean percentage of time spent in each type of activity for each institute type. As seen in the table, among the four types of training activities, discussion tended to be the most frequently occurring type, with participatory activities being the second most frequent. Didactic activities tended to be the third most frequent, while simulation or role play activities were generally the least frequent category and occurred at a relatively low rate of frequency overall. Non-training activities occupied a substantial portion of the training day, and were generally the second most frequent category overall.

These results suggest that the training design incorporated a large amount of participation by trainees. Among the training activities, discussion and participatory activities were the most frequently occurring types on average across the various institutes, and together, generally comprised more than half the training time. Didactic activities, which offer little opportunity for active involvement by participants, tended to occur less frequently than either discussion or participatory activities. This contrasts with the more commonly expected version of training, which tends to focus heavily on trainer-directed didactic activities following the more traditional teacher-student model. In general, there was not a lot of variation across the various types of institutes in the distribution of training time, with this pattern holding true for the majority of institutes. These findings are in line with one of the overall purposes of the Teaching Center projects, which was to provide training designs which offered hands-on and individualized training.

The mean satisfaction ratings for the management component institutes are presented in Table 2, and for the education component institutes in Table 3. Sample copies of the rating forms, as well as separate means tables for each wave are contained in Appendix A. These scales were developed specifically for the Learning Center evaluation, with appropriate modifications for the particular training session (e.g., specification of the particular training sessions, inclusion of the specific facilities used). Items included both overall ratings of various aspects of the quality of the training and facilities, as well as ratings of each of the particular sessions within the individual institutes. Each item was rated on a 1-5 scale, from low to high. Average ratings have been examined for each institute as well as across institute types. In general, participants evidenced high satisfaction with the training, with mean scores averaging above 4 for nearly all items across both components. Participants considered the goals of the training and their roles in training to be clear. They reported that they gained much new information from training and saw the training as useful for further work to be completed at their home site, either for continuation of the education modules or working on the program project. They also reported that participation in the Learning Center would be helpful in making their jobs easier. They rated the Learning Center staff as very responsive to their needs or questions. On average, participants also gave high ratings for the usefulness of each institute overall as well as to the particular sessions within each institute. The vast majority of participants thought that the sessions within each institute were the appropriate length, with 10% or fewer of the ratings indicating that the sessions were either too short or too long. The facilities, both at the hotel and at the Learning Center, were also rated fairly high, although not quite as high as other aspects of the training. The one area that was rated low was teachers’ responses to being videotaped, both in terms of how helpful it was and how likely they were to do it again, with average ratings around 2. The videotaping was offered on an optional basis to teachers participating in the last half of the training waves.

The high degree of satisfaction with training expressed by participants’ ratings is echoed by the focus group discussions with the evaluator after each training institute and the qualitative data gathered from the written open-ended questions and the written comments on the satisfaction rating forms. Some consistently reported themes have been expressed by participants throughout the training waves. These represent the most salient issues for trainees regarding both the usefulness of training and barriers experienced within this Learning Center design. These data seem representative, since participants appeared candid in their responses, and reported both positive and negative aspects of training.

Overall, participants seemed pleased with all of the training institutes, and praised both the trainers and the training plan. The individual trainers were all described as very knowledgeable in their areas, and were highly regarded. Participants expressed how much they enjoyed training and how useful it was, in contrast to many other training programs. They made comments such as "they looked forward to each day" and "they weren't bored", contrasting this with their usual experiences with training. They stated that the overall structure and plan of the Learning Center seemed to work well for the various components of training. By participating in such a program, they carried out plans developed during training that they might not have implemented on their own.

The hands-on and participatory nature of training was praised by trainees in all components. Similarly, the individualization present within the training plan was also valued. Participants reported that the trainers were flexible enough to meet their needs for training, even when this deviated from the original plan. They felt that this training design contrasted with much of the more didactic and structured training that is often offered. Given that the participatory and individualized nature of training was a primary theme of the Learning Center, it is important that trainees recognized these elements as present in the training they received, and furthermore, that they valued these aspects of training.

The training seemed applicable to the real world for participants from all components. They felt that the training was broad enough that they could apply at least some parts of training in any Head Start program. Participants liked this form of training where they got to see the host program in action and in its normal routines. This made the training seem more realistic and applicable for them, suggesting that they would be more likely to implement these ideas in their own program.

Another major theme pertained to the usefulness of the training materials, particularly those specifically prepared for the education component, both during the training itself and for use at the home site. The time trainers spent reviewing the materials with participants was seen as both necessary and informative. The materials also provided the basis for the mentors to continue working with the mentee teachers in their own program. It is important to note that trainees found these materials relevant and helpful, given the variety of their needs and interests for training and the differences among programs in their organization and structure.

In general, participants reported that their roles and tasks during training were clear, and that the information presented by trainers was clear. In contrast to more traditional training plans which are often narrowly focused, highly structured, and conducted in a more pure training environment, the Learning Center training included diverse topics, an individualized nature, and was located in the real-world environment of a typical Head Start program. Given this design, it was especially important that trainees understand what they are supposed to do during training and why they are doing it.

Participants felt extremely welcome by the Learning Center staff and the host program. They reported feeling like they were at home and making friends with their colleagues here. The training staff and host program were seen as receptive and responsive to their needs, both professional and personal. This positive response suggests that trainees were more likely to be open to learning from training conducted in an environment where they felt accepted and welcome.

One recurring suggestion was to increase the amount of training, particularly by allowing participants to return either within their program or together with all the participants from their component. Trainees enjoyed the opportunity to network with colleagues from other programs and/or the host program, an opportunity which they feel is rare in Head Start. They often reported that they had established a bond with one another and would like the opportunity to maintain it through the Learning Center. They also felt that the trainees could learn from one another by this process, and could give the trainers ideas as well. While these findings suggest that trainees found training useful and innovative, it also suggests that they may need additional mechanisms for interacting with and learning from their peers in other programs, either through training opportunities such as the Learning Center or otherwise.

One suggestion for improvement concerned the availability of follow-up help once participants returned to their home site. While they felt that training was useful and gave them many ideas they planned to implement in their program, they also felt the need for additional resources and materials to help them follow through on this. Based on these responses, the Learning Center implemented an option for follow-up training at the trainees' home sites for later waves. Because of the small number of follow-up visits, and their occurrence at the various program sites, it was not possible to gather data to test the effectiveness of this particular addition. However, the expressed desire of participants to continue to carry out the work begun during training represents one measure of success for the training design.

Another suggestion for improvement was to provide a better explanation of the purpose and goals of the Learning Center training ahead of time to participants. While trainees reported that they all understood this by the end of training, they felt that they needed a better understanding ahead of time. For example, they did not always realize why the entire management team needed to be present for those institutes or why the training was not designed for new teachers, and therefore did not always choose to send the most appropriate individuals to training. Participants were in agreement with the training goals once they understood them, but would have liked this clarified before participating in training. While Learning Center staff continued to work on this issue based on this feedback, it remained an area of concern to participants. The individualized and participatory nature of Learning Center training is different from the often more familiar didactic model, and therefore it may be harder for participants to understand the more unfamiliar Learning Center model. While participants appreciate the additional efforts that were made, many still believed that further explanation was necessary.

The evaluation of the effectiveness of the management training measured changes in the working conditions or organizational climate of the work environment using the Early Childhood Work Environment Survey, ECWES, (Bloom, 1993). This survey was chosen because it is one of the few widely-used measures of organizational climate which is designed specifically for early childhood environments. The ECWES includes measures of 10 major dimensions: 1) collegiality; 2) opportunities for professional growth; 3) supervisor support; 4) clarity of procedures, policies, and responsibilities; 5) reward system; 6) decision-making structure; 7) goal consensus; 8) task-orientation; 9) physical setting; and 10) innovativeness. Three additional scales measure commitment to the organization, congruence of the work environment with the ideal, and decision-making influence, with the latter measured in two subscales, current level and desired level. The surveys were administered to all staff members in the participating Head Start programs who had been present for at least six months at the time of the survey administration. The first set of pre-training surveys were completed prior to the inception of training for the particular program, and the second set of post-training surveys were administered about one year later, following the completion of the training wave. The original plan for the Learning Center had incorporated only a pre-training survey, which was used by the management trainers as a training tool. During the planning year, a post-training climate survey was added to the project design in order to allow evaluation of the effects of management training on organizational climate changes. Because the pre-training survey results were used by the management trainers, we were concerned about whether this would result in any "teaching to the test" and therefore interfere with the objectivity of the evaluation findings. However, because these surveys are completed by a much larger number of program staff than those who attend training and the organizational climate dimensions covered are general nature, it seemed appropriate to use these data in the evaluation.

The analyses include data from 21 programs which completed both the pre- and post-training surveys. (One program in Wave A1 did not complete the post-training surveys and therefore was not included in the analyses.) A total of 893 pre-training surveys and 934 post-training surveys were completed from these programs. Approximately 25% of the surveys were completed by individuals who directly participated in the Learning Center training (management and/or education components), while the remainder were completed by program staff who did not attend any of the Learning Center training institutes.

For the measurement of change related to the education component, observations were conducted of the classrooms of each participating teacher before and after training. Independent observers were recruited and trained by the evaluator to conduct observations in the classrooms of teachers participating in the Learning Center. Different observers gathered the pretest and posttest data in a given classroom. The pretest classroom observations were conducted prior to any Learning Center training for the program (including management training), and the posttest observations were conducted approximately 3-4 months after teachers completed training. Both pre-training and post-training data were gathered from 77 classrooms in 21 programs. Only pre-training data were available for an additional 25 classrooms, either because the teacher did not end up attending the Learning Center training after the pre-training data had already been gathered or because the teacher left the Head Start program before the post-training observations were conducted. These 25 observations were not included in the analyses.

In addition, the same set of observational measures was gathered in the host program classrooms which served as demonstration sites during each of the four years training took place (which represent the second through fifth years of the Learning Center project). For the host program, these data were gathered in both the fall and the spring during the first two years of training, and were gathered only in the spring during each of the last two years. Observations were conducted in 12 different teachers’ classrooms over the four year period. During that period of time, the host program maintained 8 classrooms each year. In year 1, data were gathered from 9 different teachers’ classrooms, with both fall and spring observations in 5 of these. In year 2, both fall and spring data were collected in 4 classrooms. Year 3 observations were conducted in 4 classrooms, and year 4 observations in 5 classrooms. Of the 12 classrooms, 6 had data over at least two different training years. Four classrooms had both year 1 and year 2 data, two additional classrooms had both year 1 and year 3 data, and two of these six had data from three different years.

Five measures of classroom quality were used during a 3-4 hour observation in each participating teacher’s classroom: 1) the Early Childhood Environment Rating Scale, ECERS, (Harms & Clifford, 1980) which provides a global measure of numerous aspects of the classroom environment; 2) the Caregiver Interaction Scale, CIS, (Arnett, 1989), which measures teacher involvement; 3) the Interacting Scale of the Assessment Profile for Early Childhood Programs, Profile, (Abbott-Shim & Sibley, 1992) which provides a rating of teacher-child interaction; 4) the Anti-Bias Environment Checklist, ABC, (Peisner-Feinberg, 1993) which examines the extent to which the child care environment is multicultural, non-biased, and individualized for the children; and 5) the Classroom Overview (Peisner-Feinberg, 1993) which provides a description of the organization of the classroom. These measures were selected to cover a wide range of characteristics of quality in early childhood settings, as well as to cover the variety of domains potentially affected by participation in training.

The ECERS provides a measure of seven areas of classroom environments: 1) personal care routines, 2) furnishings and display, 3) language and reasoning, 4) fine and gross motor activities, 5) creative activities, 6) social development, and 7) adult needs. A slightly modified version of the scale was used, based on the modifications used in other large-scale studies. A list of the modifications is included in Appendix A. The ECERS consists of 37 items, scored on a 1-7 scale (low to high), with descriptors for scoring at the odd number scores. Mean item scores of 5.0 or above are considered in the range of developmentally appropriate practices, with scores between 4.5 and 4.9 considered "close to developmentally appropriate practices". In addition to a total score and total child items score (consisting of all items except provisions for exceptional children and adult needs), two factor scores are used to examine different components of classroom quality. The developmentally appropriate activities factor consists of items relating to materials, schedule, and activities. The appropriate caregiving factor includes items pertaining to adult-child interactions and supervision. The two factor scores are more internally consistent and therefore are considered more valid than the subscale scores. The ECERS has become one of the most widely used measures of early childhood environments in both practical and research settings (e.g., Bryant, Clifford, & Peisner, 1991; Cost, Quality, and Outcomes Team, 1995; Goelman & Pence, 1986; Whitebook, Howes, & Phillips, 1989).

The CIS includes four subscales measuring different aspects of teacher involvement: sensitivity (warmth, attentiveness, engagement), harshness (critical, threatens children, punitiveness), detachment (low levels of interaction, interest, and supervision), and permissiveness (doesn’t exercise control or firmness). The CIS consists of 26 items which are scored on a 1-4 scale according to how frequently each of the items occurs, from not at all to very much. More positive teacher involvement is represented by higher ratings on the sensitivity subscale and lower ratings on the harshness, detachment, and permissiveness subscales. For the total score, ratings on the latter three scales are reversed, so that higher scores represent more positive teacher involvement. The CIS has been used in a number of studies examining the quality of early childhood environments (e.g., Cost, Quality, and Outcomes Team, 1995; Whitebook et al., 1989), as well as in studies examining the effectiveness of teacher training (e.g., Arnett, 1989).

The interacting scale of the Profile includes four subscales for rating different areas of teacher-child interaction: 1) teacher-initiated positive interactions, 2) teacher responsiveness to children, 3) positive behavior management, and 4) child happiness and involvement. This scale includes a total of 15 items, which are scored yes/no according to the dominant style or mode of classroom interactions. Overall and subscale scores are converted into the proportion of yes ratings (0-1.00), with more yes scores representing higher quality. This scale has been widely used in a number of studies of the quality of early childhood environments and Head Start in particular (e.g., Bryant, Burchinal, Lau & Sparling, 1994).

The ABC measure examines the extent to which the child care environment is multicultural, non-biased, and individualized for the children. This measure was included since the provision of diverse and multicultural environments was an important focus of the education component training. The ABC was designed specifically for this evaluation because no other appropriate measures were found, and none of the other measures included in the observations adequately represented this area. The measure was based on the anti-bias curriculum work of Derman-Sparks (1989), which is one of the leading curricula in this area. This instrument examines three aspects of the setting: 1) visual/aesthetic environment; 2) materials and activities; and 3) interactions. A total of 41 items are rated on a 4-point scale, ranging from none/never to many/often, with higher scores indicating more diverse environments.

The Classroom Overview provides a description of the child-centeredness and organization of the classroom. The purpose of this measure was to gather basic information about the general structure of the classroom not included in the other instruments. We have used similar measures in other studies of early childhood environments, and found the information useful in providing a more complete picture of classroom quality (e.g., Bryant et al., 1991). The Classroom Overview includes rankings of group time (large group, small group, individual); the role of the teacher and child in choosing activities, choosing peers, and determining completion of activities; ratings of the type of activities (work and play); ratings of teacher tolerance of non-task-oriented behavior; and ratings of child attention. All but the last two variables involve rank-ordered or categorical ratings. The last two items were rated on a 1-5 scale (low to high), with higher ratings representing a more positive situation.

Each of these evaluation measures was determined prior to the inception of training. However, in contrast to the organizational climate surveys used in the management component, the classroom observation data were used only for evaluation, and not training, purposes. Because the education component evaluation data were used at the individual level, rather than at the program level as in the management component data, it was decided not to share pre-training evaluation data with the trainers. Further, because of the specific nature of the items on the classroom observation measures and the fact that they were linked to a particular individual, different measures were used as training and evaluation tools for the education component. Additionally, two minor modifications were made to the original evaluation plan during the course of training. First, trainers were asked to report data on teacher module selection during training because it was discovered that teachers did not reliably report this information.

Two sets of analyses were conducted using hierarchical linear models. The first set included two independent variables: time, which examined whether there were differences in the ratings before and after training, and participation, which examined whether there were differences in the ratings between those staff members who directly participated in the training (from all components) and those who did not. In addition, the interaction between time and participation was included to test whether participants were more likely to perceive greater changes in organizational climate before and after training than non-participants. In the second set of analyses, the program involvement ratings completed by the management trainers were included to examine whether programs that were more highly involved in the training and follow-up activities scored higher than those that were less involved. In addition, the interactions of both participation and involvement with time were included, with nonsignificant interaction terms deleted from the final model. Each of the 10 dimensions and four additional scales and subscales was tested separately, with time (pre-training vs. post-training), participation (direct participant vs. non-participant), and program involvement (in the second set) included as fixed-effect variables, and the Head Start center included as a random-effect variable. Center was included to partially adjust for the effects of repeated measures over time from the same individuals as well as non-independence of ratings from individuals within the same center.

The least squares mean ratings for each of the 14 outcome variables based on the first set of analyses are presented in Table 4 for pre- and post-training scores and in Table 5 for participant and non-participant scores, with significant main effects indicated with asterisks. Scores on each of these variables range from 1 to 10, with higher scores representing a more positive work climate. On the first set of analyses, significant differences between pre-training and post-training scores were found for four of the dimensions, collegiality, professional growth, goal consensus, and task-orientation. All of the changes were in the predicted direction, with higher scores after training than before, suggesting that management team training was related to improvements in the overall organizational climate of the Head Start program.

There were also significant differences on seven dimensions between participants and non-participants: professional growth, clarity, reward system, decision-making structure, goal consensus, setting, and innovativeness. In addition, the scales measuring commitment to the organization, congruence with the ideal, and decision-making influence also showed significant differences between participants and non-participants. Participants rated all of these outcome variables higher than non-participants, suggesting that they had more positive views of these particular aspects of organizational climate within their programs. There were also significant time x participation interactions on two of these dimensions, goal consensus and physical setting. These interaction effects indicate that participants, compared to non-participants, perceived greater improvement after training in goal consensus and the physical setting within their program. These results suggest that direct participation in the training positively influenced perceptions of change in organizational climate.

Separate analyses were conducted for each of the five observational measures of classroom quality used: 1) the Early Childhood Environment Rating Scale, ECERS, (Harms & Clifford, 1980); 2) the Caregiver Interaction Scale, CIS, (Arnett, 1989); 3) the Interacting Scale of the Assessment Profile for Early Childhood Programs, Profile, (Abbott-Shim & Sibley, 1992); 4) the Anti-Bias Environment Checklist, ABC, (Peisner-Feinberg, 1993); and 5) the Classroom Overview (Peisner-Feinberg, 1993). Only data from participants with both pre-training and post-training assessments were included.

Three series of analyses were conducted using repeated measures of. analyses of variance (ANOVAs) to test whether there were changes in scores over time, and whether these changes were related to the training module or to the level of program involvement. In all of these analyses, time was included as an independent variable with two levels (pre-training and post-training) and the pre-training and post-training scores on each of the observational measures were the dependent variables. The first set of analyses tested whether there had been significant improvement in scores over time (pre-training and post-training). The second series examined whether the particular module selected during training was related to the amount of change in scores over time, with module selection included as an independent variable with three levels (individualizing, language, and relationships modules). The third set of analyses for each measure of classroom quality examined whether trainer ratings of program involvement in the education component training predicted the amount of change in scores over time, with program involvement ratings included as a continuous independent variable.

The principal measure of global classroom quality was the ECERS. The mean ECERS scores for the pre- and post-observations are reported in Table 6. For the first set of analyses, significant differences were found on both the developmentally appropriate activities factor and the appropriate caregiving factor, as well as the total score and the total child items. In addition, four of the subscales showed significant differences, furnishings/display, language/reasoning, creative activities, and social development, and one subscale, personal care, was close to significance. The two scales that did not show improvement, fine/gross motor and adult needs, were less likely a focus of the specific training received by teachers and mentors. As shown in Table 6, all of the scores changed in the direction of improvement following Learning Center training. These findings indicate that following training, the classrooms of participating teachers were higher overall in the quality of materials, activities, and schedule provided for the children.

The second observational measure, the CIS, looked at the teacher’s involvement with the children. For the first set of analyses testing changes over time, there was one significant difference on the sensitivity subscale, with a trend toward significant improvement for the total score, as shown in Table 7. However, the three other subscale scores did not change. Higher scores on the total score and the sensitivity subscale indicate better quality, while lower scores on the harshness, detached and permissiveness subscales correspond with higher quality. Sensitivity tends to be the strongest subscale on this measure, and improvement in this area suggests that the training was having a generally beneficial effect on the nature of teachers’ involvement with the children in their classrooms.

The third source of observational data consisted of ratings of the extent to which the classroom environments showed improvement in the area of diversity, using the Anti-Bias Environment Checklist. For the first series of analyses, significant differences in scores before and after training were found for the total score, as well as for the three subscales, visual/aesthetic environment, materials and activities, and interaction, as shown in Table 8. These findings indicate that the teachers did a better job of providing a multicultural and non-biased environment, and were generally less biased and more individualized in their interactions with the children following Learning Center training.

Changes in teacher-child interactions related to training were examined based on ratings from the interacting scale of the Profile. Table 9 reports the average proportion of yeses for the total score and for each of the subscales. For the first set of analyses examining changes over time, significant improvement was found for the responsiveness subscale following training, while the total score and the initiations subscales showed a trend toward improvement. It should be noted that the mean scores for the subscale measuring children’s happiness was near the ceiling of 100% both before and after training. These results indicate improvement in teacher-child interactions after the Learning Center training, particularly in terms of teachers being more responsive to children’s initiations for interaction.

The fifth observational measure was the Classroom Overview, which provided a description of the basic organization of the classroom. Because of the nature of the ratings on these items, statistical analyses were not conducted on the items which were rank-ordered or categorical. Descriptive data indicating the percentage of classrooms rated in each category are reported in Table 10 for these items. Examination of these percentages before and after training suggests a pattern of change in the direction of better quality. Following training, somewhat more classrooms were rated as having more small group and less large group activities. The percentage of classrooms in which children primarily chose the activities increased quite a bit, and the percentage in which teachers primarily chose decreased similarly after training. The activities were rated by observers as more like play and less like work after training. There was little change in teacher vs. child choice of peers, however. The number of classrooms in which children primarily decided when activities were completed was somewhat larger after training than before, while the number in which teachers made some or all of the decisions was somewhat lower. There were no significant differences in the two items which were rated on 1-5 scales and could be analyzed, teacher tolerance for non-task oriented and non-disruptive behavior and child attention during teacher-led activities. Means for these items are reported in Table 10

The second series of analyses tested whether the training module selected (individualizing, language, or relationships) was a significant predictor of the amount of change in classroom quality exhibited over time. There were no differences in the pattern of change related to module selection. For every measure tested, module selection was not significantly related to the amount of classroom quality improvement. These findings suggest that improvements in classroom quality related to the Learning Center training were not dependent on the particular module selected, but rather, were similar across the three modules.

In general, there was a clear trend toward improvement in classroom quality within the host program over the four years the Learning Center was in operation across the various areas measured. As seen in Table 11, the average ECERS scores tended to increase throughout this period of time. Average scores were generally within close range of developmentally appropriate practices (mean scores of 4.5-4.9), with several areas meeting the criteria for developmental appropriateness by year 4 (mean scores of 5.0 or above). A similar trend toward improvement was found for teacher involvement scores on the CIS, as shown in Table 12. For both the total score and the sensitivity subscale, scores increased more than one standard deviation from year 1 to year 4, with improvement seen each year. This pattern was somewhat less strong for the three negative subscales, harshness, detachment, and permissiveness, although there was some tendency toward improvement (i.e., lower scores). Similarly, some improvements were found in the area of cultural diversity in the classroom, as seen by the scores on the ABC measure in Table 13. The improvement seems most apparent for the environment subscale, with a somewhat less consistent pattern of improvement for the other areas. As seen in Table 14, general improvement was also found in teacher-child interactions, as measured by the interacting scale of the Profile. Overall, teachers were more positive in their interactions with children over time. Similarly to the findings for the participant outcomes, scores on the child happiness subscale were near the maximum each year. The pattern of change in the classroom organization, based on the Classroom Overview data shown in Table 15, is less consistent. There were some aspects in which these classrooms became more child-centered, such as child choice of activities and peers, type of activities, teacher tolerance of non-task oriented behavior, and child attention during teacher-led activities, although many of the changes were differences between year 1 and the other years.

In general, the training design incorporated a large amount of participation by trainees. The training activities most frequently consisted of discussion, with participatory activities being the next most frequent, both of which involve high levels of trainee participation. Didactic activities, the more traditional teacher-directed approach which involves very little participation by trainees, tended to occur less frequently. Activities involving simulation or role play were very infrequent in this particular training model, which may be a function of the emphasis on actual participatory activities. Non-training activities tended to occupy a significant amount of time, generally being the second most frequent category. Much of this time involved activities which may have enhanced the training either through more indirect learning and development opportunities such as classroom tours, or by creating a more welcoming environment such as providing lunch. Some of the non-training time did detract from the training time, however, such as participants arriving late.

Based on their own reports, participants viewed training very positively overall. Participants in both education and management components had average satisfaction ratings above 4 on a 5-point scale (with 1 low and 5 high) for nearly all aspects measured. The data from participant focus groups and written comments echoes these findings. Key themes from these data suggest that participants viewed particular characteristics of the training as strengths of this model: knowledgeable trainers, hands-on and participatory nature of training, applicability to the real world, usefulness of the training materials, clear roles and tasks for training, and welcoming attitude of the Learning Center staff. Given that the Learning Center utilized a very different approach from the standard workshop training often encountered, one which involved longer-term involvement by the program staff, greater emphasis on participatory activities, and more individualization, it is important that participants recognized these characteristics as strengths. Two major areas were noted by participants as needing improvement,: wanting follow-up help after returning to their home sites and needing a better understanding of the training model beforehand. Both of these issues were addressed by the Learning Center throughout the project and resulted in some changes to the training design. Trainees also requested more training, both additional networking opportunities with staff in similar positions from other programs and training for additional staff within their programs. While it was not feasible to alter the training design in response to these issues, it suggests that such training is highly valued by Head Start staff both in terms of the formal content and the informal networking opportunities offered.

The results of the outcome evaluation pertaining to the management component indicate that the Learning Center training is related to improvements in organizational climate, with even stronger positive effects in some cases for those who directly participated in the training compared to those who did not. Significant improvements following training were found on four dimensions, collegiality, professional growth, goal consensus, and task-orientation, all of which seem relevant to the content of the training provided. In addition, staff members at programs which were more involved in the training and follow-up activities also perceived the work climate more positively overall than those at less involved programs, although there were no differences in the amount of improvement seen over time. The survey results were based on information from all program staff, while only about one-quarter participated directly in training, suggesting that this type of training more broadly affected the overall quality of the work environment. Furthermore, concepts such as organizational climate entail widespread issues which would be expected to change at a relatively slow rate. The finding of significant differences over a one-year period on several of the aspects of organizational climate that were measured seems meaningful, given both the limited exposure of program staff directly to the training and the wide variation among programs in this area. While these results offer positive evidence for the success of the Learning Center training, it may be that longer-term follow-up would provide even greater opportunity for change.

Similar to the findings for the management component, participation in the Learning Center training was related to quality improvements in the education component of Head Start programs. Overall, the pattern of results from the outcome evaluation pertaining to the education component indicates improvement in the quality of teaching practices on a number of dimensions following participation in the Learning Center. Positive changes were found in the developmental appropriateness of the care and education practices, teacher sensitivity and responsiveness to children, the multicultural and individualized nature of the classroom environment, and the child-centeredness and structure of the classroom following training. These findings were based on independent and comprehensive observations of the classroom environment, including the materials, activities, schedule, interactions, and organization. It is notable that significant results were found across all of the domains of classroom quality that were examined. Furthermore, the specific pattern of changes found within and across the various measures seems consistent with the goals and objectives of the training. The changes that were found were in the direction expected given the type of training that was provided (e.g., developmentally appropriate activities, teacher involvement and responsiveness), while the areas in which change did not occur seemed less relevant to the training design (e.g., fine/gross motor activities, adult needs, behavior management). The consistent pattern of findings in all areas measured suggests that the Learning Center training is effective in improving the quality of participating teachers’ Head Start classrooms across a wide variety of domains. Furthermore, the improvements in classroom quality following training were not related to differences in the training modules selected, suggesting that such training is more broadly useful in improving quality.

There are three potential limitations to the interpretation of the evaluation findings. First, these results are based on a pre-training / post-training repeated measures design, with data gathered from the same individuals and/or programs over a limited period of time. The originally proposed evaluation design also included an independent comparison group derived from programs on the waiting list for the next round of training; however, once the Learning Center began to operate, there was almost never a waiting list with programs continuously recruited for each wave of training. While results based on the changes before and after training are appropriate for this type of evaluation, and further, allow for greater power in detecting differences, they do not allow for causal attributions of the reasons for change in the same way a comparison group design would allow. Second, there was a relatively short interval between the completion of training and the collection of the post-training data. While the interval chosen was designed to allow a sufficient amount of time to implement changes related to training, it did not allow for examination of the long-term effects of training. While differences were found in both the management and education component outcomes evaluations, change in both of these areas is likely to be a longer-term process which may not have been completed at the time of the post-training data collection. Third, the high levels of staff turnover in the participating Head Start programs made collection, and to some extent, interpretation, of the post-training data problematic. Especially in the case of the classroom observations, it was not possible to gather post-training data for nearly 25% of the classrooms that had been included in the pre-training data collection due to staff turnover. Turnover was less of an issue for the organizational climate survey data collection, since these data were analyzed at the program level, not the individual participant level. While some of the turnover occurred before the training began and therefore would not affect the evaluation results, turnover that occurred among individuals who participated in the training but left the program before the post-training data collection could have an effect. Since only classrooms with both pre-and post-training data could be included in the analyses, these findings may provide a less precise estimate of the effects of training than if data on the complete sample attending training could have been included.

In sum, these findings offer evidence for the effectiveness of the North Carolina Head Start Learning Center training in enhancing Head Start program quality at both the organizational level and the classroom level. Given the systems approach taken by this Learning Center, it is important that quality improvements were found at both levels. While it is not possible to determine whether the systems approach resulted in greater differences than would be found by working with management or education components alone, these findings do suggest that this approach was successful in generating positive changes in each. Further, trainees viewed their experiences at the Learning Center quite positively overall, which provides additional evidence for the success of the Learning Center training. These positive perceptions may also have served to enhance their responsiveness to training and thereby its effectiveness. In addition, the existence of the Learning Center did not interfere with the quality of the host Head Start program, and may have had some beneficial effects.

There are several significant features of the Learning Center design which distinguish it from many other training models, and which may be important for the success of any future replication efforts. First, the training generally maintained an emphasis on individualized and participatory activities, in line with adult learning theory approaches. While this focus was one of the key goals of the Teaching Center demonstration projects, it is not typical of much of the training that is generally available. The training design also involved a relatively long-term plan, with each program participating over a 6-month period. This long-term relationship facilitated individualization of the training as well as offered greater opportunity to implement changes related to the training.

Second, the training staff attempted to meet the needs of participants not just in terms of learning opportunities, but also as individuals, again in accord with the principles of adult learning theory. This resulted in positive attitudes toward training by participants, and possibly in greater receptivity and responsiveness to the training. The facilities used supported the training activities, and further contributed to the welcoming attitude perceived by participants by providing adequate meeting space, meals and snacks, and overnight accommodations.

Third, participants seemed to appreciate the opportunity training provided to network with colleagues from other programs, particularly for component coordinator staff. This design allowed both formal and informal opportunities for learning, with participants learning both from one another and from training staff, and vice versa as well. The networking opportunities may have enhanced trainees’ positive attitudes toward training, as well as offered unforeseen educational interactions.

Fourth, by linking the Learning Center with a host Head Start program, participants were provided the opportunity to train in real-life settings. Further, the quality of the host classrooms tended to be higher than the participant classrooms, thereby allowing for good demonstration classrooms. However, these were not necessarily "model" classrooms, which meant that there were opportunities for observing both good and poor practices based on developmentally appropriate standards, and the training needed to use both as relevant learning opportunities. It is also important to note that the host Head Start program maintained its usual operations during the project period. This made the learning opportunities more realistic for participants, but added an additional responsibility to the host site. Accordingly, the Learning Center training had to be flexible in response to the changes and needs of the host site. For example, the teachers whose classrooms were used as demonstration sites changed over the course of the training. In some cases, this was related to turnover in the teaching staff at the host site, and in others to the desires of individual staff members to be included or excluded as a demonstration class.

Arnett, J. (1989). Caregivers in day-care centers: Does training matter? Journal of Applied Developmental Psychology, 10, 541-552.

Bryant, D. M., Burchinal, M., Lau, L. B., & Sparling, J. J. (1994). Family and classroom correlates of Head Start children’s developmental outcomes. Early Childhood Research Quarterly, 9, 289-309.

Bryant, D. M., Clifford, R. M., & Peisner, E. S. (1991). Best practices for beginners: Developmental appropriateness in kindergarten. American Educational Research Journal, 28, 783-803.

Cost, Quality, and Child Outcomes Study Team (1995). Cost, quality and child outcomes in child care centers. Denver, CO: Economics Department, University of Colorado at Denver.

Derman-Sparks, L. (1989). Anti-bias curriculum: Tools for empowering young children. Washington, DC: National Association for the Education of Young Children.

Goelman, H., & Pence, A. R. (1987). Children in three types of day care: Daily experiences, quality of care and developmental outcomes. Early Child Development and Care, 33, 67-76.

Harms, T., & Clifford, R. M. (1980). The Early Childhood Environment Rating Scale. New York: Teachers College Press.

Jorde Bloom, P. (1993). Early childhood work environment survey. Evanston, IL: Early Childhood Professional Development Project, National-Louis University.

Jorde Bloom, P., & Sheerer, M. (1992). The effect of leadership training on child care program quality. Early Childhood Research Quarterly, 7, 579-594.

Peisner-Feinberg, E. S. (1993). Anti-Bias Environment Checklist. Unpublished rating scale. Chapel Hill, NC: University of North Carolina at Chapel Hill.

Peisner-Feinberg, E. S. (1993). Classroom Overview. Unpublished rating scale. Chapel Hill, NC: University of North Carolina at Chapel Hill.