Lundy and Berkowitz
gather more information than those low on the trait (Streufert & Nogami, 1989). These activities should lead to the development of more tracks. Lundy and Berkowitz (1957) found that more cognitively complex individ- uals possessed more functional constructs in memory.
Hypothesis 6: Cognitive complexity is positively related (0 the number of script tracks available in memory.
Employees with many script tracks can draw on a wide array of poten- tial responses for handling problems. Script tracks are a cognitive resource that is expected to increase promotability. Leong, Busch, and John (1989] studied the relationship between the effectiveness of insurance sales agents number of actions in their scripts. They found that agents with more actions were more effective than those with fewer actions per script. Evaluating promotability involves a rating of an employee’s potential for performing a higher-level position. Promotability ratings concern not only current effec- tiveness, but also employees’ adaptability and likely application of cognitive resources to the position.
Hypothesis 7: The number of script tracks an individual has for handling job problems is positively reJated to his or her promotability.
Goodman (1968) found that cognitive complexity is directly correlated with advancement and job level. However, Goodwin (1991) and Streufert and Nogami (1989) concluded that in complex task settings, the relationship of cognitive complexity with effectiveness and advancement to higher-level positions is stronger than it is in structured or simple task settings.
Hypothesis 8: Joh complexity moderates the relationship between an individual’s cognitive compJexity and pro- motabiiity.
METHODS
Subjects and Setting
The directors of personnel at two companies requested the participation of employees across all departments and levels, from hourly workers to vice presidents. The overall participation rate was 81 percent, which provided a total of 74 subjects. One company was a manufacturer of specialized aircraft parts. Subjects in this company included 39 men and 8 women, 47 percent of whom were in the managerial ranks. Their mean age was 49 years, their mean tenure in their current positions was 9.6 years, and 47 percent had education beyond a high school level. The organization had a mechanistic structure with a high degrees of formalization, documentation, and govern- ment surveillance. The other organization was a wholesale and retail office supply company. Subjects in this company were 8 men and 19 women, 15 percent of whom were in the managerial ranks. Their mean age was 33.5,
1994 Wofford 185
their mean tenure was 4.1 years, and 37 percent had education heyond the high school level. This company was organic and had a fast rate of growth and change.
Procedures and Measures
Subjects completed a structured interview, a post-interview question- naire, a demographic survey, and the Role Construct Repertory Test (Brook, 1981), an instrument that measures cognitive complexity. Two researchers conducted the interviews and administered questionnaires.
Interviews with employees. Interviews were structured to elicit verbal protocols relating to job problems (1) that occurred frequently or very rarely (within subjects) and (2) that involved a minor goal-performance or a major goal-performance discrepancy (between subjects). The order in which the levels of job problem frequency were presented was determined randomly. Each respondent named as many approaches as he or she could think of for handling a problem that occurred frequently and one that occurred rarely (only once every two or three years). These protocols, which were tape recorded, allowed the assessment of the use of script tracks and strategy structures by the employees. The approach employees stated first was con- sidered to be their most accessible script track. Next, subjects identified the approach they were most likely to use. Half the subjects (randomly chosen) were asked what they would do if this approach were an “utter failure” (major goal-performance discrepancy), and half were asked what they would do if this approach “was not to perfection” (minor goal-performance dis- crepancy).
Coding verhal protocols. After instruction in the procedures for coding verbal protocols, two naive coders marked on the transcribed protocol sheets the numher of script track and strategy structures for each employee.^ When the coders obtained different values for a protocol, I selected the value to be used. I assessed interrater reliability by correlating the code values of the two coders. The low interrater reliability for cognitive processing for frequently occurring problems (.51, shown in Table 1) reduces confidence in the ro- bustness of results involving this construct. Several studies have shown the use of simultaneous verbal protocols for assessing cognitive processes to be valid (Ericsson & Simon, 1980; Russo. Johnson. & Stephens, 1989). Effect coding (Kirk. 1982: 186) was used to identify the structures that the employ- ees would access following the goal-performance discrepancy prompt: a previously accessed script track was coded – 1 , a new script track 0, and a strategy track, 1. A new script track was operationally defined as a track for handling a goal-performance discrepancy not included in an employee’s original list.
Survey and test. Immediately following the interview, employees com- pleted a demographic survey and the Role Construct Repertory Test. The
The definitions and examples used for coding can be obtained from the author.
186 Academy of Management Journal February
demographic survey assessed each employee’s age, job level, and tenure in the present job.
Supervisor questionnaires. At the end of the session, employees were informed of the purpose of the study and the questionnaires. The immediate supervisor completed and returned a questionnaire describing the complex- ity of the job one level ahove the employee’s current job and evaluated her or his promotability to that job. An 85 percent return rate was obtained. A third-party coordinator assured confidentiality.
