The data were collected from rural and urban schools and health service organisations such as dispensaries, hospitals and health centers. Of 682 health workers including doctors, medical assistants, nurses, and midwives, 357 worked in urban areas and 325 in the countryside. Of the 396 teachers who took part in the study, 226 were in urban while 170 were in rural schools. A selected group of other personnel in the ministries and local administration of health and education who were considered to have a wider knowledge of their respective services was interviewed. These included, for instance, the head of the teaching services commission, professional hospital administrators, medical superintendents, chief nursing officers, head teachers, district medical and education officers and a group interview with the Uganda Medical Association. Not all information collected will be used in this study.
The data were collected by means of a structured interview and detailed questionnaire. The questionnaire sought four categories of information:
1) objective information about the respondent and his/her organisation such as education level, sex, magnitude of kinship responsibility, the length of service with a particular organisation and the size of organisation.
2) income-related information including monthly take-home pay, distance between work-place and home, other income generating activities. Under this category we also gathered information on absence incentives and barriers.
3) respondents' psychological assessment and evaluation of their work experience conceptualised in this study as the psychological contract. It included the hypothesized consequences of the evaluation namely organisational commitment and job satisfaction.
4) supervisory rating of the respondents discretionary cooperation operationalised as organisational citizenship behaviour.
A number of open-ended questions were asked concerning the major difficulties faced by health or education. These included a question seeking to have a general description of the state of the service. They included a description of the foremost management problem and priority. They also included questions tapping the external and internal environments of the institutions and of the specific organisation visited.
Another set of questions tried to tap what the respondent wanted to see done in order to deal or respond to what had been described earlier.
1) Psychological contract: It was measured by a battery of items comprising of
a) Fairness (Justice): A five item measure of distributive justice by Price (1988) was adopted for the purpose of this research. The measure is based on social exchange and especially equity theories.
b) Expectations (Need): A seven item measure adapted from Cook and Wall (1980) of need non-fulfilment was used.
c) Trust (Trust): A six item measure of organisational trust adapted from Cook and wall (1980) was used. Items for all the three were on 7-point Likert scales. The responses of the first two components ranged from strongly agree to strongly disagree. Responses on trust differed from item to item. For instance the first item responses ranged from extremely sincere to extremely insincere. The second item responses ranged from extremely good to extremely poor.
The items were pooled and subjected to a reliability analysis. A coefficient of reliability (Cronbach Alpha) of 0.85 was obtained.
2. Organisational Commitment (Affect): An eight-item measure of affective commitment by Allen and Meyer (1990) was adopted for use. The coefficient of reliability was 0.82.
3. Job satisfaction (Jobsat): Ten items from Cook and Wall were employed in the study. Cronbach Alpha was 0.85.
4. Organisational Citizenship Behaviour: Sixteen items were compiled from Munene (1979) and Organ and Konovsky (1989). Two components were measured:
a) Conscientiousness (Consc). Nine items measured general behaviour such as seeking responsibility and always being ready to accept more work. The coefficient of reliability was 0.89.
b) Altruism (Altruism): Seven items were used to measure this component. The coefficient of reliability was 0.6.
5. Intention to exit (Intent): Three items were used to measure exit intentions.
6. Personal and role-related variables of education, tenure, salary, gender, age, and kinship responsibility were measured. The kinship measure required the individual to state the number of adults and children he or she is responsible for and whether the respondent was married or not (Blegen, Mueller and Price, 1988). Each affirmative answer was awarded 1 score. The total kinship index load was 13.
7. Income-generating activities: An open ended question required the respondent to indicate what he or she does to make up for the difference between salary income and his or her expenditure.
