 |
The Work Reorganization ProgramReport to the Senior Advisory Stakeholder CommitteeDecember 1, 1997 Wave One FindingsCivilian Labour Adjustment in National Defence Research Project
Preface The research team and sponsors of the CLAND project would like to acknowledge the generosity of the individuals who have agreed to participate in this study, during what for many is a turbulent period in their lives. The knowledge and understanding gained would not have been possible without their cooperation. This report presents the initial findings of the Civilian Labour Adjustment in National Defence (CLAND) study. The information contained in this report represents the preliminary baseline data for the study; future reports will present findings from four subsequent waves of data collection. The objective of the CLAND study is to investigate the impacts, over time, of the downsizing in the Department of National Defence on affected employees. Civilian workers offered the Department’s special departure incentive will be tracked over a four-year period to examine their socio-economic and labour market experiences. The impacts and experiences to be tracked focus around several theme areas. In the early rounds of data collection, special emphasis is given to individuals’ experience with the Civilian Reduction Program (CRP) and internal workforce adjustment processes used to administer the downsizing. As the study progresses, later periods will concentrate more on the external labour market and socio-economic experiences of affected employees as well as the many factors influencing those experiences. The research methodology adopted in the study incorporates a longitudinal design featuring multiple waves of data collection consisting of an initial baseline interview and regular contacts over the subsequent three and a half years. The study population includes all individuals who have received letters of offer between September 1994 and August 1996, regardless of departure decision. By tracking the experiences of both departing employees and those who remain, it is possible to compare the outcomes and experiences of the two groups. Two cohorts have been distinguished on the basis of the date upon which employees received their letter-of-offer from the Department. The first cohort consists of individuals affected prior to the July 15,1995 amendments to the Workforce Adjustment (WFA) Directive, whereas the second cohort involves those employees offered a departure incentive after July 15th. The extension of the data collection period for the second cohort beyond the March 1996 termination date of the Civilian Reduction Program led to the inclusion of a small proportion of individuals (7.4% of sample total) who were offered buyouts under Treasury Board’s ERI/EDI program applicable to designated federal departments. A further division is created within Cohort 1. The June 1995 launch date of the CLAND project prevented the researchers from obtaining three-month data for a large proportion of the Cohort 1 population. Rather than exclude this segment of affected employees completely, it was decided that individuals receiving letters of offer prior to March 1995 would be first contacted at nine months from time of offer. With the exception of Cohort 1a which skipped the three-month interview, all participants will be contacted a total of five times over a 45-month period. In the initial wave of data collection, interviews were conducted with a total of 2352 participants during a seventeen-month period extending from June 26,1995 to November 25,1996. Response to the study has been favorable, with 72 percent of individuals contacted agreeing to participate. A brief summary of the key findings from the first wave of data collection is presented below. More detailed information on these and other findings are contained in the main body of the report. For ease of review they are grouped by subject category.
The CLAND project was originally conceived in the spring of 1994, following announcement of the drastic personnel reductions scheduled for the Department of National Defence (DND). Downsizing targets of 13,500 civilian personnel were contained in consecutive federal budgets: 8,400 in 1994 and a further 5,200 in 1995. The projected reductions were to represent almost half of the 1993 Departmental indeterminate staffing level of 28,553 civilians. The reduction program signaled not only a steep escalation of downsizing efforts within National Defence but also the introduction of a special departure incentive, known as the Civilian Reduction Program (CRP). As expected, the National Defence initiative became a precursor to the subsequent 45,000 reductions to the broader federal public service achieved through the application of a similar departure incentive. The Union of National Defence Employees, the union representing most of the affected employees, identified a need to track the individual-level outcomes of affected workers over a number of years in order to be able to objectively assess the impacts of the buyout program. Although the general issue of worker displacement has been the subject of some research, the magnitude, breadth, and occupational diversity found in the DND context were felt to merit special investigation. To launch the project, the union solicited the involvement of other key stakeholder organizations. In early July 1995, a Memorandum of Understanding outlining the terms of reference for a multi-partite research project was signed with senior representatives of the Department of National Defence, Treasury Board Secretariat (TBS), Human Resources Development Canada (HRDC), and the Public Service Alliance of Canada (PSAC). Summary of the civilian reduction program in DND As part of the 1994 federal budget, the Department of National Defence, in conjunction with Treasury Board, introduced a special one-time departure incentive package for its indeterminate employees. Under the Civilian Reduction Program (CRP), affected employees were to be given a letter-of-offer outlining the terms of their individualized incentive package and a defined departure date. Employees were given 60 days in which to reach a decision on whether to accept the offer. The CRP consisted of a financial component and a training option. The financial component had four pieces: a lump-sum payment equivalent to 12 months of regular pay; severance pay at the lay-off rate stipulated in the applicable collective agreement (usually calculated at two weeks pay for the first year of service and one week of pay for each additional year); a CRP supplement of up to six weeks of regular pay based upon a combination of age and years of service; a CRP departure allowance of one week of pay for each year of continuous service (to a maximum of 15 weeks) for employees entitled to an annuity under the federal government’s superannuation act. The combination of these elements could not exceed a ceiling equivalent to 104 weeks of regular pay. In addition, the training option provided for up to $7000 over four years for retraining towards employment outside the federal public service. The CRP distinguished among four types of departure offers. Type A, known as directed offers, were received by employees occupying positions specifically targeted for elimination through the discontinuance of a function or the closure of a unit. Type B, known as group offers, were delivered to a set of employees occupying similar positions in a particular unit and/or occupational category from which a fixed proportion of positions were to be eliminated. Type C refers to positions being relocated while Type D covered those employees on leave without pay at the time of receiving a letter-of-offer. The latter two categories were negligible in size, together representing less than one per cent of all employees offered the CRP. Officially put into operation in May of 1994, the CRP program remained in effect for approximately two years, ending on March 31,1996. During that period, however, amendments made to the federal Workforce Adjustment Directive (WFA) in the summer of 1995 would have direct impacts on the delivery of the CRP. First, employees declining directed offers (Type A) lost their entitlement to a reasonable job offer from the employer. Although the federal government committed itself to making an effort to accommodate surplus employees, job guarantees were explicitly removed. Secondly, the more voluntary group offers (Type B) were discontinued on July 15,1995, the date at which the amendments to the WFA took effect. Finally, upon termination of the CRP program, subsequent reductions to the DND civilian workforce became subject to the terms of the twin Early Retirement Incentive (ERI) and Early Departure Incentive (EDI) programs operating across the federal public service. Modelled in general on the CRP, the ERI/EDI programs were made available to indeterminate (full-time) employees who were declared surplus within designated, "most affected" federal departments, including National Defence. In almost all instances, the ERI/EDI provides individuals with a financial package inferior to the compensation offered under the CRP. Like its predecessor, the EDI includes financial and training components. The primary difference resides in the lower magnitudes and limits imposed by the EDI; the maximum payout under EDI is 44 weeks of pay for employees with less than five years service, and 70 or 90 weeks pay for employees with more than five years (depending on their entitlement to an unreduced pension). The ERI was made available to surplus employees between the ages of 50-59 eligible for a federal government pension. In addition to their regular severance, these employees could retire early without having the normal pension reductions applied and, in certain instances, qualify for a special separation benefit (to a maximum of 15 weeks pay). The objective of the CLAND study is to investigate the longitudinal impacts of the downsizing in the Department of National Defence. Civilian workers affected by the reductions will be tracked over a four-year period to examine socio-economic and labour market outcomes and experiences. The principal research themes and priorities identified through consultation with stakeholder representatives are as follows:
Summary of research methodology In developing the questionnaire to be used in the survey, the research team reviewed existing literature in the area of labour adjustment as well as examples of instruments used in comparable studies. Extensive consultation was also conducted with stakeholder representatives, academic advisors, and personnel from Ekos Research (the firm hired to administer the CLAND survey) to ensure an efficient instrument that would meet the information needs of the respective funding organizations and allow for comparison with findings emerging from other studies. The diverse interests inherent in the multi-partite undertaking and the time constraints imposed by telephone interviewing posed a challenge in the construction and editing of the questionnaire. Pretests were performed in both official languages with a small proportion of recipients of letters-of-offer drawn from the DND administrative database representing the study population. The purpose of the pretest was to review the effectiveness of the instrument prior to actual field use by clarifying response categories, confirming proper skip logic, and testing the validity of questions. Based on the pretest results, minor revisions were made, primarily to shorten the questionnaire to allow for an average completion time within 30 minutes. The survey instrument utilizes, almost exclusively, closed-ended questions in which respondents choose from specified response categories. The advantages of closed-ended questions are that they are more efficient to administer and analyze, reduce the coding burden, and provide a more reliable process for capturing factual information. Open-ended questions are found in only a few instances such as the identification of adjustment services not offered during reductions. The substantive content of the instrument was formulated to advance the specified research objectives. The body of the questionnaire was organized around several theme areas:
The occupational categories referred to throughout the report represent the principal groupings found in the federal public service classification system:
The analysis of respondent information obtained through the interview process is further supplemented by linking to the CRPIS administrative database supplied by National Defence. The merger with select fields of CRPIS data yields information on years of service, union membership, occupation, compensation, organizational unit, language, age, and CRP characteristics. The research methodology adopted in the CLAND study incorporates a longitudinal design featuring five waves of data collection. A process of continuous sampling allows for the entry of respondents into the sample on a staggered basis, at a prescribed duration from the date they receive their CRP offer. Following an initial baseline interview at three months, participants will subsequently be contacted at nine months, twenty-one months, thirty-three months, and forty-five months from date of offer. Two distinct cohorts have been designated, based on the receipt date of their letter-of-offer. The first cohort consists of individuals affected prior to the July 15,1995 amendments to the Workforce Adjustment Directive whereas the second cohort involves those employees receiving letters after the amendments came into force. The July 15th cleavage is critical since it marks a fundamental change to the discretionary complexion of the CRP. The elimination of job guarantees contained in the WFA amendments severely limit employees’ alternatives to acceptance of the CRP. While the CRP takers in the first cohort have a bonafide opportunity to remain in the public service, their colleagues in cohort two encountered much more restrictive options. A further division is created within Cohort 1. The June 1995 launch date of the CLAND project prevented the researchers from obtaining three-month data for a large proportion of the Cohort 1 population. Rather than exclude this segment of affected employees completely, it was decided that individuals receiving letters of offer prior to March 1995 would be first contacted at nine months from time of offer. These respondents will only participate in four waves of data collection. The target population for the CLAND study was all civilian employees of DND who received a special departure incentive offer. An administrative database maintained by DND, the Civilian Reduction Program Information System (CRPIS), constituted the sampling universe for our purpose. The CRPIS is the Departmental record system for the administration of the CRP program. It contains information on individuals offered the departure incentive including personal identification, select demographic characteristics, occupational data, unit, location, and specific CRP elements such as total payment. Excluded from the sample at the outset were management category employees, individuals for whom there was insufficient contact information, individuals offered the CRP prior to September 1994, and individuals whose records lacked an open date (the date they received their letter-of-offer). The management category personnel, also referred to as the executive level in the federal government, were excluded since they were offered a different departure incentive and represented very small numbers. The high level of attrition due to these exclusions necessitated a census approach to sampling in which attempts were made to contact all cases with sufficient tracking information. The original research design called for 3000 interviews to be completed in the initial wave of data collection, divided into two equal-sized cohorts of 1500 respondents. Those targets were developed on the basis of information provided by the Department of National Defence regarding the estimated volume and scheduling of their civilian personnel reductions. Unfortunately, the generation of an adequate sampling frame from which study participants could be recruited was plagued with difficulties arising from the incomplete nature of the CRPIS. A large proportion of individual records (almost half) present in the database were missing the necessary information to permit interviewers to establish contact. In addition to the absence of critical contact information, a sharp and unanticipated reduction in the volume of reductions occurred during the period of data collection for the second cohort. A number of steps were taken to compensate for the missing contact information. A second DND database, the Training/Education Tracking System used to monitor access of the CRP training allowance, was merged with the CRPIS in an effort to augment the available contact information. The Ekos survey administrators also contacted personnel coordinators at both the unit and headquarters level to gain additional information. Finally interviewers accessed telephone directory assistance to obtain numbers in instances where an address was provided. Despite these measures, the combined effect of poor contact information and the low volume of eligible cases rendered the project unable to fully complete the second cohort within a practical time-frame. As a result, a decision was taken in November 1996 to halt the initial formation of the second cohort at 800 cases. The sampling periods (time periods in which letters-of-offer were issued) and magnitudes for the respective cohorts are as follows: cohort 1a (904
cases) September 5, 1994 - February 20, 1995 The response rate has been encouraging. Among those individuals whom the interviewers were able to contact, the response rate was 72 percent. Only about one-quarter of the employees contacted refused to participate in the study. The high rate of participation also provides a measure of assurance against potential non-response bias. total sample (meeting
eligibility criteria)= 4,674 As was mentioned previously, the study population was defined as all civilian employees of DND who have received a CRP letter-of-offer. In order to detect possible bias in the formation of our sample, selected sample characteristics are compared against characteristics of the study population. To assess the representativeness of the sample we are aided by the demographic information contained in the administrative database (CRPIS). Table 1 provides a comparison between the two groupings. Table 1: Comparison of Unweighted Sample to the CRP Population and the Total DND Civilian Workforce Population Demographic variable CRPIS database CLAND sample DND Civ.Workforce
% n (1) % n % n (2)
Gender Male 64.5 (7320) 66.1 (1551) 64.8 (18278) Female 35.5 (4033) 33.9 (796) 35.2 (9948) Occupation Admin/FS 8.0 (905) 6.2 (146) 9.3 (2615) Admin Support 27.3 (3100) 24.0 (564) 25.9 (7304) Operational 46.8 (5307) 51.5 (1211) 48.5 (13677) Scientific/Professional 6.1 (687) 4.9 (116) 6.5 (1841) Technical 11.8 (1335) 13.3 (313) 9.5 (2690) Region Atlantic 21.4 (2427) 18.5 (424) 24.1 (6789) Quebec 19.2 (2180) 17.4 (399) 17.5 (4929) Ontario 35.9 (4065) 33.1 (760) 33.9 (9570) Prairies 10.8 (1219) 11.8 (270) 13.3 (3849) British Columbia 12.8 (1447) 19.2 (440) 11.1 (3141) 1complete CRPIS database for period ending August 1,1996 2as of March 31,1994 Weighting formulas were developed
by the survey administrator to correct for stratification and response bias so
as to ensure the data-set was representative of the population from which it was
drawn: DND employees listed on the CRPIS who received letters-of-offer between
September 5,1994 and August 26, 1996. After comparing the completed file of
cases with the CRPIS population along key demographic variables, weights were
required for occupation and region only. All interviews in the CLAND study were conducted by Ekos Research Associates under contract to the Union of National Defence Employees. Ekos employed a computer-assisted telephone interviewing (CATI) system to perform the interviews. The CATI system produces high quality data through its process of instantaneous data entry and its capability to follow the complex skip patterns embedded in the questionnaire. The number of callbacks (repeated attempts to contact respondents) was specified as ten, a level that exceeds industry standards. The callback rate was set high in order to minimize attrition. The telephone interviews averaged 28 minutes in length, two minutes under the allocated time. The research project also anticipated a requirement to respond to concerns that participants might express regarding any aspect of the adjustment process. Field staff were instructed to direct respondents to different project resources, depending on the nature of their question. The sources of assistance included: a 1-800 number for the CLAND research team, the DND hot-line, and the Ekos survey center. Although a few of individuals did request contact from a project representative, the number was very low (under 40). The original research design called for the first wave of data collection to be completed over a period of nine months. The difficulties encountered in generating an adequate sampling frame that were described earlier, however, caused that period to effectively double in length. The total period for the wave one data collection extended from June 26, 1995 to November 25, 1996. The extension of the initial data round beyond the March 1996 termination date for the CRP program also led to the inclusion of a small number of ERI/EDI cases (n=161) into the second cohort of the sample. The research methodology applied a staggered approach to data collection whereby individuals would be contacted at a point three months from the week in which they received their CRP letter of offer. Maintaining that rigorous practice in the face of a prolonged sampling period generated serious cost implications for the project. To prevent project costs from escalating dramatically, a modification was made to the scheduling of interviews. Instead of contacting individuals on a weekly basis, the survey administrator was authorized to collect data on a monthly basis. Although a degree of precision was sacrificed in this change, it was deemed an acceptable measure to achieve cost containment. Table 2 below illustrates the average duration between the date of offer and the initial interview for each cohort. The higher standard deviation noted for cohort two indicates the greater degree of variability resulting from the shift to a monthly "bundling" of interviews. Table 2: Mean duration: interview date from date of offer
Findings: Section 1 The second section of findings covers a broad range of items that are being tracked across successive waves of data collection. Since the distance from the adjustment event is viewed as critical feature of a longitudinal research design, our analysis examines recurring items at specific intervals. In this report, the information being tracked is presented for the contact point, three months from time of offer. As Cohort 1a participants were only contacted for the first time at nine months and thus do not possess three-month information, they are excluded from the second section of this report’s findings. The data that is reported has been weighted to ensure it is representative of the population of DND employees from which our sample has been drawn. As described earlier, weights were required for the occupation and region variables. Two separate formulas have been applied: one for the data-set used in producing section one findings (includes all cohorts) and another for the section two findings (cohorts 1b & 2 only). Prior DND
employment Job Satisfaction Statistically significant
associations were noted between overall job satisfaction and three key
demographic indicators: age (c 2=69.2, p=.000), education (c 2=29.5,
p=.000), and occupation (c 2=18.2, p=.02). A large proportion (24%)
of younger workers, those under 35 years of age, report high levels of
dissatisfaction while employees over 55 were least likely to report being
dissatisfied (10%) with their DND job. An inverse relationship between job
satisfaction and education (more educated workers report lower levels of
satisfaction) is also noted. These two factors are closely related as younger
workers in our sample tend to be the better-educated. Finally, employees in the
Administrative Support category stood out as the least satisfied respondents. Younger employees are more likely to have experienced recent career progression: 27% report receiving a promotion in the last three years. A greater degree of movement is expected during earlier stages of individuals’ career development. The process for gaining federal government employment, whereby individuals often accept entry-level positions below their qualifications in order to obtain a foothold, might explain this finding. Other factors that appear to be
associated with internal movement within the federal government include
education (those with low education are less likely to be promoted) and
occupation (the operational category reports the lowest levels of mobility, with
81% having experienced no change). Measures of the statistical significance of
the association between these variable are not calculable, however, because
internal movement is based upon a multi-response item. The CLAND study queried respondents about "moonlighting" to identify those participants who were working outside DND and the level of their commitment (measured in hours) to these endeavors. It was interesting to note that while the great majority reported no such activity, 7.3 per cent of the respondents reported alternative employment at the time they received their letter of offer. The average number of hours worked was 13.1 per week. A strong relationship was noted between age and supplementary employment (c 2=51.3, p=.000); the likelihood of working decreasing steadily as age increases. Males were also slightly more inclined towards supplementary employment (c 2=3.78, p=.000). CRP services received The first wave questionnaire
incorporated a set of one-time questions to capture individuals’ utilization
of these services so that their efficacy and impact on subsequent outcomes can
be analyzed. Participants were asked to comment on 1) the relative usefulness of
the information they received from the respective organizations in helping them
to reach a decision, and 2) the specific types of adjustment assistance and
services received from the Department of National Defence. Strong associations were noted
between individuals’ understanding of their CRP options and both the age and
occupation variables (see Table 3 below). The increased knowledge level among
older workers corresponds with their high degree of interest in the CRP offer as
presented in Table 4. Scientific/professional and administrative category
employees tend to be more knowledgeable about the CRP. This is likely due to
their higher educational background and their position in the organizational
hierarchy. University graduates demonstrate a much higher level of CRP
knowledge. It was interesting to note that CRP type was not associated with
individuals’ knowledge of the departure program. Table 3: Knowledge of CRP Options by Age and Occupation Age Knowledge Level Under 35 35 to 44 45 to 54 55 and over % % % % Poorly informed 16.0 17.2 11.6 7.7 Somewhat informed 19.5 21.8 22.5 16.0 Well informed 64.5 61.0 66.0 76.3 Total n (235) (565) (703) (766) X2 = 49.35 (6,2269), p =.000 Occupational Category Knowledge Level Administrative Administrative Operational Scientific Technical % Support % & Prof. % Poorly informed 7.3 9.6 14.9 10.1 7.5 Somewhat informed 14.1 21.9 20.5 12.0 19.8 Well informed 78.6 68.5 64.5 77.9 72.6 Total n (142) (552) (1203) (147) (225) X2 = 33.48 (8,2269), p =.000 Figure 1: Ratings of Usefulness of Respective Information Sources
The Union of National Defence Employees should also be recognized for having reached a broad segment, 72%, of the workforce sample with its CRP information (n=1653). Revenue Canada (n=1392), and Canada Employment Centres (n=1170) also were cited by a significant number of respondents as sources of information during the decision period. Figure 2 reveals the extent to which employees offered the CRP received the different forms of adjustment services made available by the Department. The financial planning services were the most commonly used: 66 percent of respondents mention receiving individual financial counselling while 69 percent participated in group sessions. Retirement counselling was also popular with 59 percent of the sample taking advantage of the provided service. The percentage totals cited in Figure 2 exceed 100% due to their derivation from a multi-response variable in which an individual may identify more than one provider or source of information. Figure 2: Frequency of Adjustment Services Received 
Figure 3 charts respondents’ identification of the most useful adjustment services. The sampled employees demonstrate a strong appreciation for the financial counseling provided; together, the group and individual financial planning services were considered to be the most useful service by approximately half the respondents. The only other service meriting special mention was the retirement planning identified as the most useful service by 22 percent. Figure 3: Perceived Usefulness of Adjustment Services Received
Age and gender appear to influence respondents’ determination of which services are most valuable. Younger employees are more likely to identify job search and career counseling services as valuable while employees over 55 years of age account for most of the interest in retirement planning. On the receipt of financial planning services, women report a stronger preference for individual meetings over group sessions. Unfortunately, the small size of many of the cells in tables involving this variable limit our power to detect statistically significant associations. The decision whether or not to accept a buyout package such as the CRP is a rare and significant event in an individual’s working life. Despite the magnitude and potential consequences of the decision, the sample demonstrated a strong interest in the departure incentives, with 74.7 per cent opting to accept an offer (includes both CRP and ERI/EDI). Seven per cent declined directed offers while a further 18 per cent declined group offers or were able to secure a substitute (an innovative feature of the CRP was its substitution component, under which non-affected employees could volunteer for the departure incentive, replacing affected employees who wished to remain in National Defence). Previous Treasury Board research (1995) initiated shortly after the introduction of the CRP, investigated the factors influencing decision outcomes among the same DND population. The CLAND research presents another opportunity to build understanding of the dynamic process leading to these decisions. The questionnaire contained two single items aimed at identifying the motivation underlying employees’ decision to accept or reject the offer. The primary reasons cited by those accepting the buyout were to facilitate retirement (26%), seeing no personal future within National Defence (24%), and a favorable perception of the CRP’s terms (13%). Decliners meanwhile overwhelmingly identified the same CRP terms as their primary reason for turning it down (31%). Individuals rejecting the CRP were almost three times more likely to cite the financial disincentive than the next most common factors: liking their job (13%), suffering a pension penalty (10%), and having been refused (10%). The last factor cited, having been refused, represents, almost exclusively, group offer recipients whose application for the CRP was not approved by their section management. The CLAND data set allows for analysis of CRP type by a host of demographic and subjective variables to provide further, and perhaps more informative, insight into factors influencing employees’ decision-making behavior. Highly significant differences in the decision-making were apparent across cohorts with participants in Cohort 2 being much more likely to accept the departure incentives (see Table 4). Whereas approximately two-thirds of Cohort 1 participants elected to accept a package, fully 94 percent of Cohort 2 respondents received either the CRP or EDI/ERI. The large difference is attributable to the severe restriction on individuals’ alternatives imposed by two practices enacted during the sampling period of the second cohort: 1) the discontinuance of group offers, and 2) the removal of the employer’s obligation, under the Work Force Adjustment Directive, to provide a reasonable job offer to affected employees. The demographic item demonstrating the strongest association with CRP type was the respondent’s age. As Table 4 illustrates, older workers were strongly inclined towards accepting the departure incentive, with a take-up rate of 92 percent. Employees at the other end of the career continuum, those under 35 years of age, also show a relatively high propensity (72%) to opt for the CRP. As we move towards the mid-range age categories, however, the interest in leaving DND plummets dramatically to approximately two-thirds. This 20-year band represents the age cohort likely to be most dependent upon their employment; they are more likely to have family and financial commitments and are too far away from a pension to consider retirement. An interesting regional pattern also emerges from the data. The acceptance rate is highest in the Atlantic provinces (84%) and declines progressively from east to west with the Pacific region reporting a level almost 20 percentage points lower. Interpretation of this finding is somewhat perplexing. While the proportion of respondents from the Atlantic region over the age of 55 is relatively high (35%), it still trails Ontario in that category and thus the age variable cannot account for the regional variation. The poor labour market conditions in the East also would be thought to mitigate against voluntary job displacement. Other factors, possibly culturally-driven, may better explain their behavior in this instance. Another interesting dimension in the pattern of CRP acceptance is found in the relationship between CRP decision type and occupation. Acceptance of the departure incentive is dramatically lower among technical category employees. No single factor emerges to explain that finding, although the technical category does contain a high proportion of older workers (45%) and those reporting low levels of skill portability (23%). Table 4: Key Demographic Variables by CRP Type Demographic Variables CRP Type Admin. Admin Support Operational Scientific Technical % % % % % Accept CRP 77.7 78.6 75.4 71.6 61.1 Declined direct 7.9 7.1 6.9 14.8 6.0 Declined group 14.4 14.3 17.7 13.6 32.9 Total n (144) (559) (1212) (147) (227) X2 = 54.4(8,2288), p .000 Under 35 35 to 44 45 to 54 55 and over % % % % Accept CRP 72.8 62.0 67.2 91.6 Declined direct 7.9 13.9 8.4 1.5 Declined group 19.3 24.1 24.4 6.9 Total n (237) (571) (711) (769) X2 = 198.6(6,2288), p .000 Less than High School Some Post College University High School % secondary % % % % Accept CRP 79.4 74.8 73.2 65.4 72.5 Declined direct 5.2 7.1 8.1 8.5 11.9 Declined group 15.4 18.1 18.7 26.1 15.6 Total n (672) (807) (293) (244) (271) X2 = 29.8(8,2287), p .000 Atlantic Quebec Ontario Prairies Pacific % % % % % Accept CRP 83.8 75.7 73.6 72.3 65.4 Declined direct 5.4 9.9 8.5 2.5 8.8 Declined group 10.8 14.4 17.9 25.2 25.8 Total n (464) (411) (785) (298) (330) X2 = 63.6(8,1537), p .000 Cohort 1A Cohort 1B Cohort 2 Accept CRP 62.4 65.8 93.7 Declined direct 10.0 7.8 6.3 Declined group 27.6 26.4 n/a Total n (825) (634) (830) X2 = 269.8(4,2288), p .000
Substantial increases in workload are also reported. Roughly half of those questioned indicate that their workload has increased over the past year with the remaining respondents equally divided between those who reported no change and those who felt their burden had decreased. CRP recipients who have not yet left the organization account for most of those reporting a decreased burden. In addition to CRP type, workload was also significantly associated with education (c 2=19.9, p=010) and occupation (c 2=23.4, p=003): employees with higher education were more likely to report increased workloads; employees in the Operational and Admin/FS categories were also subject to increases. In today’s economy a great deal of emphasis is placed on the acquisition and maintenance of a portable skill set. The possession of transferable skills is significant in the context of the DND reduction for two reasons. First, it is expected to influence individuals’ decision whether to accept the CRP and secondly, the ability to rebound successfully from job displacement may depend, in part at least, on the possession of marketable skills. Two sets of skills questions were included in the questionnaire. A measure of the possession of portable skills was derived from two items related to the perceived application of employees’ skills outside the federal public service. The second set measured basic skills: literacy, numeracy, and the ability to work with computers. In terms of the perceived portability of skills, respondents were most likely to adopt a favorable view of their skill set. Frequencies for the combined measure indicate that 23% identified low portability, 26% medium, and 51% identified a high level of portability. As Table 7 illustrates, both occupation and education are associated with differences in how respondents view their skills. Higher status and better-educated employees, comprising the scientific/professional as well as the officer-level administrative group, report more favorable perceptions. A regional variation emerged as Quebec residents distinguished themselves in reporting more favorable perceptions of their skills. Table 7: Perception of Skill Portability by Education and Occupation Education Skill Portability Less than High school Some post College University High School % secondary % % % Low 27.1 25.4 20.7 17.7 15.8 Moderate 30.5 26.4 22.6 20.6 19.6 High 47.4 48.2 56.7 61.7 64.7 Total n (427) (510) (177) (159) (157) X2 =36.1 (8,1429), p =.000 Occupational Category Skill Portability Admin. Admin. Support Operational Scientific Technical % % % & Prof. % % Low 17.5 27.1 26.5 9.0 21.0 Moderate 15.4 23.5 29.1 21.8 22.0 High 67.1 54.4 44.4 69.2 57.1 Total n (86) (360) (776) (78) (130) X2 = 38.8 (8,1430), p =.000 The relationship between CRP type and perception of skill portability is also statistically significant but not in a fashion that supports a clear explanation. All three categories of CRP types reported identical proportions of respondents reporting high measures (50%) of skill portability. The variation occurs among the lower levels of portability: a large proportion of group offer decliners (31%) are found in the "low portability", whereas direct offer decliners have a higher concentration (35%) in the "moderate portability" classification. Observable age-related differences in the perception of skill portability (decreasing with age) did not achieve statistical significance. Nor were significant differences were found across cohort or gender. On this issue of basic skills, over 80 percent of the sample assessed their personal levels of both literacy and numeracy to be above-average. There were very few individuals placing themselves in the below-average segments of the seven-point scale (only 3 percent for numeracy and 2 percent in regards to literacy). The same degree of confidence was lacking in computer proficiency, however, as 41 percent reported below average skill levels in that area. Expected differences in levels of literacy among occupation groups (c 2=73.05, p=.000) were confirmed by the data: workers in the operational category were less likely to view themselves as possessing above-average levels of literacy. The CRP included a training/education allowance which provided for a maximum of up to $7000 over four years for retraining towards employment outside the federal public service. The training option is liberally interpreted to apply towards a broad variety of training activities ranging from correspondence courses to university enrollment to on-the-job training. Individuals have one year from the date they accept the CRP to submit their training plan to the Directorate of Civilian Training (DCTD) at National Defence Headquarters. Individuals could also apply a portion (up to 35%) of their training allowance towards the purchase of tools and equipment. Another contingency under the training option was to invest these funds in training provided as part of an employment arrangement with another organization. The EDI also carried a training allowance providing a maximum payment of $7000 for costs incurred in training to prepare for new employment outside the federal public service. The EDI allowance, however, was more restrictive in its terms and application. The allowance only covered expenses directly related to the training: the purchase of tools and equipment permitted under the CRP was highly limited in the EDI. Also, individuals receiving a pension benefit under the federal superannuation program or those availing themselves of the ERI were specifically excluded. Given the prominence of training in labour adjustment and labour force development and the accessibility of the specific DND allowance, it was expected that a large proportion of departing employees would avail themselves of the training option. Earlier focus group research conducted by Treasury Board (1994) suggested that younger employees in particular found the training option attractive. The CLAND questionnaire contained ten questions dedicated to the utilization of the CRP training allowance, focusing on training receipt, type, duration, and progress. The survey results confirm the popularity of the training allowance among CRP recipients. Of those employees who have accepted the CRP and not withdrawn from the labour force, 64 percent (n=516) indicate an intention to apply for funds under this provision. The popularity of the training allowance was consistent across both cohorts as well as gender and regional groupings. Table 8 presents information on the association between utilization of the training allowance and age, occupation, and education. Table 8: Application for CRP Training Allowance by Age, Occupation and Education* Age CRP training Under 35% 35 to 44% 45 to 54% 55 and over Yes, has applied 38.9 29.6 19.2 7.8 Yes, will apply 58.0 65.1 34.6 24.8 No 3.1 5.3 46.7 67.4 Total n (98) (222) (244) (242) X2 = 257.0(6,806), p =.000 Occupation Admin.&FS% Admin.Sup.% Oper.% Sci/Prof.% Tech.% Yes, has applied 21.3 25.8 17.7 36.6 16.2 Yes, will apply 22.2 46.2 45.8 47.3 27.8 No 56.5 28.1 36.5 16.1 56.0 Total n (61) (227) (423) (40) (56) X2 = 42.2(8,806), p =.000 Education < High High Some Post. College University School School Sec. % % Yes, has applied 14.6 21.9 21.3 27.8 26.5 Yes, will apply 40.1 40.8 49.2 43.8 48.3 No 45.3 37.3 29.6 28.4 25.2 Total n (211) (312) (103) (86) (93) X2 = 20.6(8,805), p =.000 *excludes individuals citing retirement The expected differences in training interest across age groupings is reinforced: over 95 percent of employees under 44 years of age intend to access these funds as opposed to 54 percent of employees between 45 and 54. What is also interesting is that 32 percent of employees over 55 years of age expressed an intention to pursue training. Older workers are not normally inclined to undergo training (Crompton,1994). Their access of the CRP training allowance may be more a reflection of the liberal terms associated with its receipt than any job-oriented behavior. Educational attainment demonstrated a positive association with training interest; Table 8 illustrates that the better-educated the respondent, the greater likelihood an expression in pursuing training is noted. The relationship between training interest and occupation was less orderly. The low levels of training interest displayed by the technical and Admin/FS categories are surprising, running counter to expectations given the educational complexion of those two occupational groupings. At the three-month point of contact, only 51 individuals had actually commenced their training. Computer training was the preferred form of training, cited by over 35 percent of respondents. Training pursued by the balance of respondents was distributed across a wide range of alternatives. While the type of training is well-patterned, the duration and intensity is highly variable. Respondents reported an average course length of 18.6 weeks (s.d.=15.3) at 16.6 hours per week (s.d.= 15.7) spent on training. The large standard deviations attached to the means for both duration and hours per week indicate a great deal of variation in the depth of these training events. It would seem, however, that most training is part-time; only 26 percent of those enrolled in training are in programs demanding more than 20 hours per week. Investment in the purchase of new tools and equipment was made by 40 percent of individuals who have already applied for the training allowance. A much smaller percentage (5.1) have been able to put a portion of their funds to use in acquiring training as part of a new employment arrangement. In future CLAND reports, as greater numbers of sample respondents leave DND, more in-depth analysis of training activity will be performed. Although 39 percent of the respondents who have left DND under the CRP program viewed their exit to signal retirement, most of the remaining employees were in a position of having to search for new employment. Job search behavior is one of many variables that the CLAND study will track in order to gain understanding of the factors contributing to successful labour market outcomes. A series of questions examined four key facets of respondents’ job search experience: duration, technique, intensity, and barriers. The Civilian Reduction Program provided affected employees with a substantial notice period as represented by the time span between receipt of offer and actual departure date. In many cases the departure schedule was the subject of negotiation between individual employees and their unit manager. At our initial point of contact, three months from receipt of letter-of-offer, fully 59.4 percent of CRP takers (n=1171) had made their exit from DND. The proportion was significantly higher in Cohort 2 (62.2%) than in Cohort 1B (54.2%). In order to build upon research that examined the efficacy of job search while employed (Blau & Robins,1990; Jones & Kuhn,1994), CRP recipients intending to remain in the labour force were asked about the length of their job search both prior to, and following, their departure from National Defence. The initial expectation of people departing the public service was that it would take them a median number of 9 weeks to find new employment. Although almost 22 percent (n=52) of these respondents indicated no pre-departure job search, the median reported length was six weeks. The median post-departure duration was calculated to be four weeks. The intensity of job search is measured by hours per week invested in related activity (for example, reviewing job ads, completing resumes and applications, contacting employers). The mean response of 11 hours (s.d.= 9.8) suggests a fair degree of variation among CRP recipients. Individuals may employ a multitude of different techniques in searching for work. Our findings presented in Figure 5, suggest that the DND employees pursued a range of activities, the most common method being responding to newspaper advertisements, cited by 59.2 percent of respondents. Figure 5: Frequency of Job Search Techniques Used 
Displaced employees often encounter barriers in their efforts to acquire new employment. The perception of barriers are expected to vary according to their human capital, local labor market conditions, and personal circumstances. Far and away, the single greatest problem identified by respondents who have left National Defence was the dearth of available jobs, cited by over 33 percent of those respondents intending to remain in the labour force. The age factor trailed a distant second at 14 percent. To determine the labour market outcomes of employees departing National Defence, the CLAND study identifies and tracks detailed characteristics of individuals’ subsequent employment experience. Recognizing that displacement from stable employment in the 1990s often results in multiple jobs, sequential employers and, in general, increased economic insecurity (Seitchik & Zornitsky,1989; Picot & Pyper,1993; Beck,1992; Krahn,1991), the questionnaire is designed to capture information from multiple jobs. Due to the proximity of the first wave of data collection to the receipt of CRP offer, however, it was expected that little labour market activity would be present. For this reason, the research team chose to confine questions to respondents’ first job following departure as well as their current job at time of interview. These assumptions were proven accurate as only 10 participants indicated more than one employer since leaving National Defence. For both first and current job, employees were questioned about their date of hiring, source of information leading to hiring, type of work, industry, unionization, employment status, earnings, and job satisfaction. A series of questions were also assembled for individuals declaring retirement, primarily looking at the timing of their exit and the appearance of any subsequent labour force participation. The introduction of the CRP appears to have enticed a large percentage of civilian employees to voluntarily retire from National Defence. Of the respondents who have accepted the CRP and since left National Defence, 39.2 percent claim to have shifted into retirement. In response to the question whether the reduction program had encouraged them to retire earlier than planned, 79 percent of retirees replied in the affirmative. This response is reinforced by an examination of the age breakdown among retired sub-group of departed employees. Table 9: Retirement Age
As well, only 4.5 percent of retirees currently report any work performance. Successive waves of date collection will be able to discern whether retirees later re-enter the workforce in some capacity. The percentage of CRP recipients who have secured full-time employment upon leaving National Defence is low. Just 9.3 percent (n=66) of those who have left were working full-time at the time they were interviewed while a slightly smaller proportion (6.9%) classify themselves as self-employed. The small numbers of workers engaged in full-time employment limit the researchers’ analysis of their circumstances. Combining full-time workers with contract employees and part-timers does allow us to detect a few emergent patterns. In accordance with the popular literature on effective job search behavior (Bolles,1986; Mackay,1993), personal contacts were identified as the principal means by which job openings were discovered in 45 percent of the cases. The average weekly earnings of those workers ($558) were low in comparison to those reported by their colleagues who declined directed offers and chose to stay in DND ($771) although the average number of hours worked per week by the latter group is slightly higher (38.6 versus 34.6). Although interpretation of the apparent wage disparity is restricted by the small numbers available at this point in the study, it does provoke interesting questions around their labour force participation: are CRP recipients unavailable to secure equivalent or better-paying positions outside the public service, are they willing to accept lower wages in light of the financial buffer provided by their departure package. One concrete and dramatic impact, however, can be found in the loss of union affiliation characteristic of employees who have left. Whereas 95 percent of the sample was covered by a collective agreement in their public service positions, only 16 percent report being union members in their new job. Satisfaction levels with
non-salary benefits is lower in new jobs ( A small number of individuals (n=56) in the sample identify themselves as self-employed (non-contract). Three-quarters of these respondents are operating home-based business; one quarter have taken on paid employees. The distribution across industry type is broadly dispersed, with the exception of construction sector where 20 percent of self-employment is located. Among those who have accepted the CRP and since left the Department, 19 percent (n=136) describe themselves as unemployed and looking for work, whereas another 10 percent (n=70) are unemployed but not actively seeking employment. The total level of unemployment among departed employees would inflate dramatically (to almost 50%) if retirees were excluded from the distribution. Among those respondents classified as "unemployed and looking for work", the largest proportion (27.1%) cited as their main reason for not working, the absence of available work while another 23% identified their own decision to take a temporary break from their job search. Various studies of unemployment have shown that displacement generally results in income decline extending over a considerable period of time (Jacobson et al, 1993,1993a; Hamermesh, 1989; Ontario Ministry of Labour,1993; Grayson,1985; Ruhm,1992). While there is also evidence that some workers prosper following their involuntary departure (Ruhm,1991; Lauzon,1995), one of the hypotheses guiding the CLAND research is that individuals accepting the CRP would experience economic decline though not necessarily in the short-term due to the generous buyout terms. The questionnaire contained a series of questions designed to track changing economic circumstances. In addition to a three-item scale of economic deprivation, individuals were asked about personal savings levels, any requirement to cut back on important expenditures, and the burden of basic housing expenses. Over three-quarters of the respondents report an overall standard of living that has either remained the same or improved over the past year. This stability was also noted on the specific economic dimensions: 86% reported no increase in debt burden; 73% reported that their personal savings were the same or better compared to a year ago; and 67% reported no requirement to cut back on their spending patterns. Table 10: Economic Deprivation by CRP Type CRP Type Accept CRP% Declined Declined Group Offer% Group Offer% Economic Condition Better 44.6 24.0 18.9 Same 13.9 16.0 21.2 Worse 41.5 60.0 59.8 Total n (1163) (83) (172) X2 = 51.4 (4,1417), p .000 Expenditure Cutbacks Yes 31.0 39.5 40.3 No 69.0 60.5 59.7 Total n (1161) (82) (171) X2 = 7.7 (2,1414), p .021 Table 10 illustrates, however, the clear gulf in perceptions of economic deprivation between CRP recipients and those who declined the offer. The CRP takers are significantly more likely to identify short-term economic improvements. They are also about 10 percent less likely to report having cut back on important expenditures. The injection of capital associated with acceptance of the CRP buyout appears to have created an economic honeymoon for those departing National Defence. It provides a stark contrast to the regression cited among the "survivor" group choosing to remain Departmental employees. The latter group is almost three times as likely to report a decline in economic conditions than to identify an improvement in this area. That finding is entirely expected given the legislative wage restraints that federal employees have been subject to for the past six years. Figure 6 presented below, outlines the principal areas in which respondents identify making cutbacks to their personal spending. Discretionary expenditures such as entertainment, vacations, and new clothing figure prominently. Contributions to personal savings are also among the tops items cited for reductions by respondents. Figure 6: Expenditure Reduction Types 
Research on the health impacts of unemployment has consistently demonstrated a significant correlation between displacement and health. Aggregate level studies (Brenner: 1979,1983; Catalano & Dooley 1977) have identified increased rates of morbidity and ill-health in response to economic downturn. Studies comparing employed and unemployed segments of the labour force (Hepworth, 1980; D’arcy & Siddique:1985) have confirmed higher rates of health problems among those without work. Finally, a few longitudinal studies (Cobb & Kasl: 1979; Grayson:1985; Liem & Rayman:1982) have noted the health impacts on workers in the wake of plant closures. The CLAND study contributes to this body of knowledge by monitoring a number of health dimensions for both displaced and "survivor" respondents. Individuals are questioned in regards to how they rate their own health (as well as that of their spouse and dependents), frequency of visits to physicians and mental health workers, and use of prescription medication. Individuals’ subjective rating of their personal health was fairly positive, averaging 5.6 on a seven-point scale. Perceptions of spousal health were roughly equivalent (5.7) while respondents described dependents’ health somewhat more favorably (6.3). These ratings were consistent across CRP type, suggesting no immediate evidence of differences amongst the groups. Analyses of the respective health measures were conducted controlling for gender. The significance of gender in health behavior was apparent in all measures with the exception of subjective ratings of individual health where the means were comparable. Table 11 outlines the differences between the gender categories for prescription drug use and contact with health professionals. Table 11: Prescription Drug Use and Contact with Health Professionals by Gender Gender Male% Female% Prescription Drug Use Yes 36.2 43.5 No 63.8 56.5 Total n (920) (508) X2 = 7.4 (1,1428), p .006 Contact with Health Prof. None 23.9 19.5 One 27.4 24.9 Two 20.3 15.5 Three-Four 13.9 18.0 Five or more 14.5 22.2 Total n (914) (502) X2 = 22.1(4,1415), p .000 There was a clear, strong linear association between age and prescription drug use (c 2=111.7, p=.000); workers over 55 years of age represented almost half (45%) of all respondents reporting the use of prescribed medication. Insight into respondents’ psychological well-being is obtained by administering constructs measuring depression and anxiety, two conditions associated with unemployment and job loss (Hayes & Nutman,1981; Hoffman et al 1991; Jackson & Warr,1984; Ferman & Gordus, 1979). The General Health Questionnaire (GHQ), a twelve-item scale measuring depression has been augmented to include four additional questions. Anxiety is measured by a condensed, seven-item version of Spielberger’s 1979 State Trait Personality Inventory. Although most respondents report
low levels of depression ( In addition to the depression and anxiety items, the CLAND questionnaire incorporated several psychological constructs thought to influence departure decisions and adjustment outcomes. Scales were created to gauge the extent of respondent happiness, self-esteem, locus of control, achievement orientation, and aspiration. Inversely mirroring the results of the depression scale, the scores for a combined measure of happiness revealed the same variations; men (F=10.38,p=.001), married individuals (F=5.74,p=.000), and older workers (F=21.58,p=.000). Figure 7 below displays the means for each of the psychological items broken down by CRP type. One-way analysis of variance, used to test whether means differ across groups, reveals that statistically significant differences occur among CRP types for a number of measures. Individuals deciding to accept the CRP tend to report: a stronger sense of personal responsibility for their own fortunes (F=11.0,p=.001) and lower levels of depression (F=13.62,p=.001) and anxiety (F=23.3,p=.001). Figure 7: Psychological Dispositions by CRP Type
Locus of control refers to the extent to which individuals internalize responsibility for various outcomes. High scores on the scale indicate that the individual believes he or she possesses a strong degree of control over events while a low score reflects attribution of responsibility to external sources. Comparing the means for locus of control across the categories of select variables reveals that married respondents reported a greater sense of internal control than did others (F=3.20, p=.01). Locus of control also increased with level of educational achievement (F=5.04,p=.000); university graduates demonstrated the highest average score. Workers under the age of 35 also reported significantly higher degrees of internal locus of control (F=4.94,p=.002). Respondents scored high on measures of self-esteem. On a seven-point scale, the mean score for the sample was 6.4 (s.d.=.66). One-way analysis of variance revealed no significant differences in self-esteem when testing for the effect of age, gender, education, marital status, or CRP type. A measure of respondents’
achievement orientation was created by combining three related items from the
questionnaire. The need to realize achievement in the workplace was determined
to vary slightly by education level, most notably between university graduates ( Aspiration was another measure in which most respondents reported high scores. The mean of 6.5 suggests that respondents consistently expressed a strong desire to receive personal fulfillment and challenge through their work activity. As may be expected, this need to derive fulfillment through work declines slightly with age (F=8.66,p=.000) as careers begin to draw to a close. Women also reported higher levels of aspiration than did men (F=6.23,p=.001). The findings of the first wave of data for the CLAND study provide necessary baseline information against which results from future series can be compared. The overall quality of the data and the high response rate (72%) are positive features of the initial data collection in this longitudinal research project. Assessment of the services provided by National Defence, the Union of National Defence Employees and other branches of government indicate that these sources were successful in delivering information to the workforce and contributed to an intermediate level of understanding among affected employees. The findings also highlight the importance attached by employees to specialized financial planning and retirement services to help them better understand the impacts of their decisions. Analysis of the factors associated with individuals’ decision whether to accept the departure incentive suggests that age is the most significant factor. Older workers were far more likely than other groups to take advantage of the CRP, reinforcing its role as an early retirement vehicle. Indeed, approximately 40 percent of respondents who have left National Defence identify themselves as retired. Respondents in the Cohort 1 sample were offered the CRP prior to the elimination of the job guarantees provided by the federal government’s WFA directive. That strong default option combined with the departmental practice of issuing group offers to create a degree of voluntarism that the Cohort 2 sample simply did not enjoy. The wave one findings dramatically reflect the impact that the discontinuance of those two features exercised on individual decision-making. Whereas roughly two-thirds of the first cohort availed themselves of the offer, fully 94% of the second cohort elected to take the buyout. Our preliminary round of findings also depicts a high level of economic comfort among CRP recipients as reflected by short-term improvements in their debt and savings positions as well as their overall standard of living. Future waves of the study will closely monitor economic and labour market patterns of departing employees to determine the durability of these initial improvements. Employees remaining in National Defence have in most cases chosen continued employment over the uncertainty that accompanies the short-term financial boost. At the same time, however, the public service work environment in which they remain is characterized by slumping morale and intensifying work loads. The scarcity of information on labour force participation presented in this report reflects both the proximity of the baseline contact to the CRP offer date and the high proportion of retirees within the first cohort. More information on labour market outcomes will be forthcoming in later reports as more individuals leave DND and begin to make their way outside the federal public service. Chi-square statistic (c 2): the chi-square is used to test the statistical independence of two variables in a contingency table. F-test statistic: an hypothesis testing procedure used in one-way analysis of variance to compare means between groups. Letter-of-offer: letter received by affected employees outlining the terms of their individualized CRP incentive package and a defined departure date. Employees were given 60 days in which to reach a decision on whether to accept the offer. Likert scale: an ordinal level scale that combines the scores of several individual items to provide a single measure of a given trait; the alpha level cited refers to the degree of internal reliability, the strength of the association among the various items. Mean ( Non-sampling error: also known as method error, refers to errors that occur in the data collection process, either in the administration of the interview, or in the coding of data. One-way analysis of variance: a statistical procedure used to test whether means of different groups are equal to one another. P value: a coefficient that indicates the level of statistical significance of an hypothesis test, the probability of rejecting a true null hypothesis. Response rate: refers to the level of participation in the study presenting the number of completed interviews as a proportion of total contacts made. Sample: refers to a subset of a population; in the CLAND study the sample is drawn from a population of DND employees receiving a CRP letter of offer. Standard deviation (s.d.): the most frequently used measure of dispersion of scores in a distribution, it represents the average deviation from the mean as expressed in standardized units. Survivors: in the literature on displacement (Noer,1993), employees who remain after downsizing events are referred to as "survivors". Beck, N. (1992) Shifting Gears: Thriving in the New Economy, Harper Collins. Blau, D.M. and Robins, P.K. (1990) "Job Search Outcomes for the Employed and Unemployed", Journal of Political Economy, 98 ,June: 637-55. Bolles, R.N. (1986) What Color is Your Parachute?: A Practical Manual for Job Hunters and Career Changers, Berkeley CA: Ten Speed Press. Brenner, M.H. (1979) ''Health and the National Economy'' in Ferman and Gordus, op.cit. Brenner, M.H. (1983) ''Mortality and Economic Instability: Detailed Analyses for Britain'', International Journal of Health Services, Vol. 13 No. 4. Brockner J. et al (1985) "Layoffs, Self-Esteem and Survivor Guilt: Motivational, Effective, and Attitudinal Consequences", Organizational Behavior and Human Decision Processes, (36), 229-244. Catalano, R. and Dooley, D. (1977), "Economic Predictors of Depressed Mood and Stressful life-events in a Metropolitan Community", Journal of Health and Social Behaviour, 18: 292-307. Cobb, S. and Kasl, S. (1979) ''Some Mental Consequences of Plant Closings and Job Loss'', in Ferman and Gordus, op.cit. Crompton, S. (1994) "A Note on the Self-Initiated Training of Job Losers", Perspectives on Labour and Income, Statistics Canada, Spring, catalogue 75-001E. D’arcy, C. and Siddique, C.M. (1985) "Unemployment and health: An Analysis of Canada Health Service Data", International Journal of Health Services, 15,4,609-635. Ekos Research Associates (1997) Civilian Labour Adjustment in National Defence: Wave One Field Report Ferman, L. and Gordus, J. eds (1979) Mental Health and the Economy, Kalamazoo: WE Upjohn Institute. Grayson, P. (1985) ''The Closure of a Factory and Its Impact on Health'', International Journal of Health Services Vol. 15, No. 1. Grayson J.P. (1985) Corporate Strategy & Plant Closures: the SKF Experience, Toronto: Our Times Publishing. Hamermesh, D.S. (1989) "What Do We Know About Displacement in the U.S.?", Industrial Relations, 28 (Winter): 51-59. Hayes, J. and Nutman, P. (1981) Understanding the Unemployed, New York: Tavistock Publications Hepworth, S.J. (1980), Moderating factors of the psychological impact of unemployment", Journal of Occupational Psychology, 53:139-145. Hoffman, W.S., V. Lee Hamilton, C.L. Broman and D. Rauma (1991) ''Unemployment Depression and Self-Blame Among Autoworkers'', paper presented at the American Sociological Association Annual Meeting, Cincinnati, 1991. Jackson , P.R, and Warr, P.B. (1984), "Unemployment and Psychological Ill-health: the moderating role of duration and age", Psychological Medicine, 14, 605-614. Jacobson, L.S., LaLonde, R.J., & Sullivan, D.G. (1993) "Earnings Losses of Displaced Workers", American Economic Review, 83 (September): 685-709. Jacobson, L.S., LaLonde, R.J., & Sullivan, D.G. (1993) The Costs of Worker Dislocation, Kalamazoo,MI: WE Upjohn. Jones, S.R.G and Kuhn, P. (1994) "Is Employed Job Search Really More Effective" McMaster University. Krahn, H. (1991) ''Non-standard Work Arrangements'' Perspectives on Labour and Income, Winter, Statistics Canada, catalogue 75-001. Lauzon, D. (1995) "Worker Displacement: Trends, Characteristics, and Policy Responses" HRDC Applied Research Branch. Liem, G.R. and Rayman,P. (1982), "Health and Social Costs of Unemployment", American Psychologist, 37,10:1116-1123. Mackay, H (1993) Sharkproof: Getting the Job You Want, New York: Harper Collins. Marks, M.L. and Mirvis, P. (1992) "Rebuilding After the Merger: Dealing With Survivor Sickness", Organizational Dynamics, (21)2, 18-32. Ontario Ministry of Labour (1993) The Displaced Workers of Ontario: How Do They Fare? . Noer, D.M. (1993) Healing the Wounds: Overcoming the Trauma of Layoffs and Revitalizing Downsized Organizations, San Francisco, Josey-Bass Publishers. Picot, G. and W. Pyper, (1993)Permanent Layoffs and Displaced Workers: Cyclical Sensitivity, Concentration, and Experience Following the Layoff, Statistics Canada Analytical Studies Branch, Research Paper Series. Ruhm, C. (1991) "The Time Profile of Displacement-Induced Changes in Unemployment and Earnings" in Addison, J.T. (ed.) Job Displacement: Consequences and Implications for Policy, Detroit: Wayne State University Press. Ruhm, C. (1992) "Are Workers Permanently Scarred by Job Displacements?", American Economic Review, March, 319-324. Seitchik, A. and Zornitsky, J. (1989) From One Job to the Next: Worker Adjustment in a Changing Labour Market, W.E. Upjohn Institute for Employment Research. Treasury Board Secretariat (1995) "The Civilian Reduction Program at the Department of National Defence: Factors Affecting Early Departure Decisions", Ottawa. Treasury Board Secretariat (1994) "Consultations Among Civilian Employees Regarding DND Civilian Reduction Program", Ottawa, (prepared by Optima).
Each item in the questionnaire was examined for potential non-response bias. Those items for which the level of missing data exceeded five per cent of the expected number of cases were flagged and appear in the following table. Table 12: Item Non-Response
An examination of the variables for which missing cases exceeded the 5% threshold indicates the following patterns to account for the non-response.
Scale construction and validation procedures In a number of instances, items were combined to establish indices of particular concepts. These indices were of two types: Likert scales and combined measures. In both instances, questions have response categories along a continuum ranging from 1 to 7, where low values generally mean strong negative responses and high values denote strong positive responses. In cases where the scale was reversed, item codes were transposed prior to the calculation of the final summary measure. Thus, for all the indices cited in this report, low values are strongly negative while high values are strongly positive responses to a given topic. To retain the original 7-point scale and accommodate missing data, the summary measures were calculated by averaging the individual item values. The Likert scales require a minimum of four variables to generate meaningful alpha coefficients used to measure inter-item reliability. In situations where the condition for Likert-type scales could not be met, but two or three similar variables were moderately correlated, the items were combined to create a stronger overall measure. The following is a more detailed explanation of each summary measure constructed for this report. In all, seven Likert scales were created:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
=5.4, s.d.=1.56)
but express considerably lower levels of satisfaction with the adjustment
process (