L'auteur présente la démarche suivie pour valider l'Inventaire de Satisfaction au Travail (IST). Le développement de l'Inventaire y est relaté ainsi qu'une description de la version retenue pour fin de validation. Administré à deux groupes de travailleurs (infirmières et contremaîtres), l'auteur explique comment l'instrument a été utilisé ainsi que la signification et la façon de calculer les scores sur les différentes échelles de l'Inventaire de Satisfaction. Il analyse également les résultats obtenus à la suite des tests de stabilité, validité et fidélité.
The Job Satisfaction Inventory : A Test of Validity
This study was undertaken in order to test the validity of the Job Satisfaction Inventory (JSI). To do so, the author goes through the following steps.
First, a preliminary version of the JSI was developed. This first version, the Work Attitude Questionnaire (WAQ), was administered in a pre-test context, to seven hundred (700) employed individuals. The WAQ consists of 102 items. Each item refers to a reinforcer in the work environment. The respondent indicates how satisfied he is with the reinforcer on his present job. Five responses alternatives are presented for each item « Very Dissatisfied ; Dissatisfied ; Satisfied ; Very Satisfied ; Extremely Satisfied ». Thus responses are scored 1 through 5 proceeding from left to right in the answer spaces. Scale scores are determined by summing the weights for the responses chosen for the item in each scale.
Each WAQ scale consists of three items. The items appear in block of 34 with items constituting a given scale appearing at 34 item intervals.
Data on the reliability and validity of the WAQ although adequate, showed that some of the scales had to be redefined or set aside because of a lack of relevancy.
Following this pre-test, it was decided to redevelop a « new » instrument, the Job Satisfaction Inventory (JSI).
The JSI was based on a new set of 20-scale Likert format questions.
The new instrument was constructed to sample more appropriate (intrinsic and extrinsic) reinforcement dimensions. Items stems were shortered and scales were extentioned to four items each. An attempt was made, through item wording, to make scale content more homegeneous. At the same time, items were worded to maximise readability (5th grade level).
The twenty scales of the JSI are as follows :
3— Work attractionA — Autonomy
7— Communication I
8— Communication II
9— Working conditions
14— Company policies and practices
17— Supervision — human relations
18— Supervision — technical
19— Ability utilization
The new instrument was designed to measure actual satisfaction with job rein-forcers. The JSI was administered to two heterogenous groups : 692 French Canadian registered nurses and 298 French Canadian foremen.
Evidence for the concurrent validity of the JSI is derived from the study of group differences in satisfaction, especially occupational differences in satisfaction. A large body of research accumulated over the last thirty years indicates that there are occupational differences in job satisfaction, in both level and variability.
To determine whether the JSI reflected these differences, data for 2 occupational groups were analysed by one way analysis of variance (to test differences in level of expressed satisfaction) and by Bartlett's test of homogeneity of variance (to test differences in group variabilities).
Group differences (among the 2 occupational groups) were statistically significant at the .05 level for means on 19- JSI scales. Group differences were also statistically significant at the .05 level for variances on 13- JSI scales. These data indicate that the JSI can differentiate among occupational groups (although this test will be carried out, in a near future, on many other groups).
To determine whether this differentiation was meaningful, means and variances for each of the 21 scales were examined to see which occupational groups had the highest and lowest means on the largest and smallest variance. (The research literature on job satisfaction has reported consistently that professional groups were the most satisfied and the non professional and the unskilled groups the least satisfied).
The data yielded by the JSI are comparable with those reported in the research literature.
Data on the internal consistency reliability of the JSI as estimated by the split half method show that for both groups the Pearson correlation coefficients (for the JSI scales) range from a high of .93 (working conditions) to a low of .68 on work attraction.
These data suggest that, in general, the JSI scales have adequate internal consistency reliabilities.
Data on the stability of the scores on the 21-JSI scales were obtained for one time internal — six months.
Test-retest correlation coefficients for the 21- JSI scales range, for the group of nurses, from .41 (Authority) to .74 (Working conditions). For the group of foremen the correlation coefficients range from .20 (Altruism) to .71 (Company policies and practices).
These data suggest that the JSI scales have adequate stabilities in time.
Data on the item homogeneity were also tested by intercorrelating the four items of each scale. Results show that the means of the correlation coefficients range from .67 (Q, Qs) to .81 (Q. QJ. These data suggest that the JSI scales have a strong item homogeneity.
These psychometric characteristics show that the JSI is a reliable and valid questionnaire in measuring job satisfaction among different occupational groups in both, level and variability.