Sie sind auf Seite 1von 20

PERSONNEL PSYCHOIJOOY

199a 43

A TEST OF THE SITUATIONAL LEADERSHIP THEORY


WARREN BLANK Tbe Leadership Group Maharishi International University JOHN R. WEITZEL Syracuse University STEPHEN G. GREEN Purdue University

Hersey and Blanchard's sittiational leadership theory (1982) is widely known and tised, but has limited, mixed empirical validation. This study examines the underlying assumptions regarding the theory's prescriptions that subordinate maturity moderates the relationship of leader task and relationship behaviors with indicants of leader eifectiveness. Results of this analysis do not support these asstmiptions. An examination of the more complex predictions of the theory also show little support for it. Findings are discussed in terms of future research and theory development.

In their situational leadership theory (SLT), Hersey and Blanchard (1969,1982) argue that a leader's task behavior and relationship behavior interact with subordinate maturity to significantly influence leader effectiveness. SLT is one of a class of situational approaches to leadership. For example, Fiedler (1964,1%7) suggests three situational factors (leader-member relations, position power, and task structure) moderate the relationship between leader traits and leader effectiveness. Path-goal theory (House, 1971) proposes that task and subordinate characteristics moderate the impact of four types of leader behavior (supportive, directive, participative, and achievement oriented) on subordinate effort and satisfaction. SLT focuses on only one situational variable (subordinate maturity) as a moderator of two leader behaviors (task and relationship) and leader effectiveness. SLT is intuitively appealing (Yukl, 1981) and popular with practicing managers. Various training publications prominently advertise SLT materials, and managers attend literally thousands of SLT programs each year. Situational leadership theory has also been cited in the academic literature as an important situational approach to leadership effectiveness (Yukl, 1981).
Correspondence and requests for reprints should be addressed to Warren Blank. Ph.D., The Leadership Group, 3463 State Street, Suite 157, Santa Barbara, CA 93105.

COPYRIGHT 1990 PERSONNEL PSYtWOLOCW. INC

579

580

PERSONNEL PSYCHOLOGY

Yet the theory proposes complex relationships between variables and contains conceptual ambiguities and contradictions (Graeff, 1983, provides a ccmplete discussion). In addition, the theory has received only limited empirical attention (Yukl, 1981). IWo published tests of SLT (Hambleton & Giunpert, 1982; Vecchio, 1987) have shown mixed support and have various methodological Umiutions. Thus, the utility of this well known theory that has widespread use b difficult to evaluate. l b better understand SLX the basic assumptions underlying the nKxiel require examination. Existing research has not done this. The more complex formulations of the model may then be examined. Limitations in previous research regarding important methodological issues (e.g., measurement of key variables) also need attention. This study empirically examines SLT in this marmer to extend our understanding of the theory and its potential usefulness. SitutitionaJ Leadership Theory SLT focuses on two primary typ)es of leader behavior: task and relationship behavior (Hersey & Blanchard, 1982). Hersey and BlaiKhard (1%9, 1982) suggest that these behaviors are very similar to consideration and initiation of structure which are well grounded in leadership literature (e.g., Cartwright & Zander, I960; Halpin, 1959; Katz, Maccoby & Morse, 1950). Moreover, Hersey and Blanchard (1982) suggest that the task and relationship behaviors be operationalized in a manner that closely parallels existing operationalizations of consideration and initiation of structure (see Leader Behavior Descriptive Questionnaire [LBDQ-XII], Stogdill & Coons, 1957). Vecchio (1987) used the LBDQXII to measure task and relationship behavior, arguing it is a more widely accepted index of leader behavior than the Hersey and Blanchard LEAD instrument. Therefore, the present study measures leader task and relationship behaviors in terms of consideration and initiation of structure. SLT also focuses on subordinate "maturity" as the key situational characteristic that is said to moderate the relationship between leader behavior (task and relationship) and leader effectiveness (Hersey & Blanchard, 1%9, 1982). Maturity is defined as the "ability and willingness of people to take responsibility for directing their own behavior" (Hershey & Blanchard, 1982, p. 151). Hersey and Blanchard (1982) argue that subordinate maturity consists of two dimensions: psychological maturity and job maturity. Psychological maturity is somewhat ambiguously defined in Hersey and Blanchard's (1982) recent work. It is diaracterized as a "willingness or motivation to do something" and as having "to do with confidence and commitment" (p. 157). Examples of its

WARREN BLANK ET AL.

581

operationalization focus on willingness to take responsibility, achievement motivation, and commitment to an objective (p. 159). In earlier works, psychological maturity was defined in terms of ' ^ e relative independence, achievement motivation, and ability to take responsibility" of the subordinate (Hersey & Blanchard, 1969, p. 221). Hersey and Blanchard suggest the relative independence component is drawn from Argyris (19S7) and involves an individtial's self-sufiSciency. The achievement motivation component reflects work by McClelland and Atkinson (e.g., McQeUand, Atkinson, Clark & Lowell, 1953). Hersey and Blanchard provide no citable reference for ability to take responsibility. They do not explain its inclusion as an element of psychological maturity. Job maturity is defined in terms of the "ability to do something" and is seen as strongly related to educational and job experience (Hersey & Blanchard, 1982, p. 157). To test the theory's underlying assumptions, both psychological and job maturity need to be addressed. Since maturity is a key element of SLT, its measurement is central to testing the theory. Existing research has not attempted to utilize a measure with adeqtiate psychometric properties. Both Vecchio (1987) and Hambleton and Gumpert (1982) used the Hambleton, Blanchard, and Hersey (1977) measure which has several psychometric problems. For example, the measure contains only five items to measure both categories of maturity. Single items are used to represent complex constructs such as achievement motivation and commitment Each item has only polar anchor descriptors (using an eight-point scale). Single item measures using only polar anchors have questionable reliability and content validity (Nunnally, 1978). A new 11-item measure of psychological maturity, developed following scale construction procedures suggested by Nunnally (1978), has shown adequate internal consistency and both predictive and concurrent validity (see Blank, Weitzel, Blau, & Green, 1988). Use of this scale to meastire psychological maturity represents a methodological advance over the two existing studies. A second issue regarding maturity is that SLT suggests that job and psychological maturity be combined. Graeff (1983) details that doing so results in internal consistency problems with the model, l b overcome the problems raised by Graeff, an alternate approach would be to consider both maturity dimensions separately when testing the theory's basic assiunptions. To test the theory's underlying assumptions, the maturity measures are considered separately. SLT argues that leader effectiveness results from appropriate amounts of leader t a ^ and relationship behaviors being provided for subordinates at different levels of maturity (Hersey & Blanchard, 1982).

582

PERSONNEL PSYCHOLOGY

The theory recommends a linear relationship between subordinate mattirity (both psychological and job) and task behavior. When subordinate maturity is low, ieaders need to provide high amounts of task behavior. When subordinate maturity is high, leaders should provide low amounts of task behavior. The relationship between subordinate maturity and relationship behavior is somewhat more complex since it is proposed to be curvilinear. When subordinate maturity is high or low, ieaders need to provide low amounts of relationship behavior. When maturity is moderate, leaders need to provide high amounts of relationship behavior (see Figure 1). When levels of the leader behaviors are provided for subordinate maturity levels as indicated in Figure 1, SLT predicts greater leader effectiveness. SLTs prescription to apply a combination of task and relationship behavior also presents problems. Combining task and relationship behavior results in the model being unable to handle some situations in a logical manner (see Graeff, 1983). Existing studies (Hambleton & Gumpert, 1982; Vecchio, 1987) only report tests of the combined effects of task and relationship behaviors, and results indicate mixed support for SLT. For

o 3 c

Task Balwvior Rtationsiip B Ml Low M2 Modarat* Pollowar Maturity M3 M4 High

1: Amounts of Leader Belicviors Prescriiied in tfae Sitaational Lcadenhip Theory tor Different Levels of Subordinate Maturity. Adapted from Leadership in Organizations (p. 142) by G. Yukl, 1981, Englewood Cliffs, NJ: Prentice-Hall. Copyright 1981 by Prentice-Hall. Adapted by permission.

WARREN BLANK ET AL.

583

this study, ourfirststep was to examine the prescribed linear relationship between task behavior and maturity and the prescribed curvilinear relationship between relationship behavior and maturity. This might shed more light on the predictions of SLT and provide a comparison to existing, more complex, published assessments of the model. Our second step in examining SLT was to test the more complex formulation of SLT. Concerns with conducting this analysis are noted above; however, a complete understanding of SLT requires taking this step in addition to focusing on the underlying assumptions. Analyzing the more complex formulation of SLT requires combining task and relationship behavior to create four "leader styles" suggested by the model (see Figure 2). SLT argues that each style is most appropriate for one of

(LOW)-*

Figure 2: Relationships Betwera Leader Style and Level of Subordinate Maturity. From Management of Organizational Behavior (4th ed.,

p. 200) by P. Hersey and K. Blanchard, 1982. Englewood Qifb, NJ: Prentice-Hall. Copyright 1982, Prentice-Hall. Reprinted by permission.

584

PERSONNEL PSYCHOLOGY

four levels of maturity. For this analysis, psychological and job maturity are combined since SLT suggests this. Again, we have noted above that there are limitations implied by this approach. However, in faithfulness to SLT, we conducted the analysis as SLT specifies. Hersey and Blanchard (1982) sugge^ a difference in "effectiyeness" when there is a leader style/level of maturity "fit" as opposed to when there is "nofit"between leader style/maturity. SLT suggests leader style SI fits maturity level Ml, S2 fits M2, S3 fits M3, and S4 fits M4. We examined thisfit/nofit relationship in relation to criteria of effectiveness. Figure 2 depicts the prescribed relationships between leader style and subordinate maturity. The curve represents the change of leader style required to achieve leader effectiveness for each level of maturity. Hersey and Blanchard (1982) define effectiveness in terms of goal accomplishment and the internal states or predispositions of the subordinate. An effective leadership "style" influences subordinates to "do a certain job" (p. 109), suggesting that subordinate performance should be affected. Effective leadership style also creates a climate where the subordinate "respects (the leader) and is willing to cooperate with (the leader)" and finds the job rewarding (p. 109), suggesting that subordinate affea also should be affected. Thus, tbis research focuses on subordinate performance and satisfaction with the supervisor and work as indicants of leader effectiveness for both steps in the analysis (i.e., examination of underlying assumptions and the complex formulation of the model). Use of multiple indicants of leader effectiveness provides a more complete test of SLT In summaiy, the present study first examines the basic assumptions of SLT. This is a reasonable step, not taken by previous research, to assess the strength of SLT. In an effort to more fully examine SLT, our second step was to analyze the relationship between combined task and relationship behavior (leader style) and subordinate maturity. For both analyses, particular attention is given to the use of measures with adequate psychometric properties since previous research on SLT, discussed above, uses measures that strain standards of validity and reliability.
Hypotheses Underlying SLT are assumptions about the contingent relationships between leader behavior and subordinate maturity (see Figure 1). These relationships suggest the following: Hypothesis 1: Subordinate performance and satisfaction (supervisor and work) will be related to an interaction of leader task behavior and subordinate maturity (job or psychological):

WARREN B I J ^ N K ETAL.

585

a. Under low subordinate maturity, task behavior will be positively related to performance; under h i ^ subordinate maturity, task behavior will be negatively or not related to performance. b. Under low subordinate maturity, task behavior will positively related to satisfaaion; under hig}i subordinate maturity, task behavior will be negatively or not related to satisfaction. Hypothesis 2: Subordinate performance and satisfaction (supervisor and work) will be related to an interaction of leader relationship behavior and subordinate maturity (job or psychological): a. Under moderate subordinate maturity, relationship behavior will be positively related to performance; under h i ^ and low subordinate maturity, relationship behavior will be negatively or not related to performance. b. Under moderate subordinate maturity, relationship behavior will be positively related to satisfaction; under high or low subordinate maturity, relationship behavior will be negatively or not related to satisfaction.

The fundamental hypothesis regarding the more complex formulation of SLT (see Figure 2) suggests the following:
Hypothesis 3: Subordinate performance and satisfaction will be higher when leadership style fits maturity level as prescribed by SLT than when leadership style does not lit maturity level. For example, when leader style 1 (high XasiL, low relationship) is used for subordinates in maturity level 1 (low), performance and satisfaction will be higher (see Figure 2).

Method Sample The. sample consisted of 27 hall directors (HD: Leaders) and 353 resident advisors (RA: Subordinates) from two large midwestern universities. Of the RAs, 54% were females, 46% were males, 82% were graduate and upperdass students, 17% were sophomores, and 1% were first year students. TWelve (45%) of the HDs were female. The two organizations are very similar in structure and procedures. The leaders (HDs) are full-time professionals supervising paraprofessional subordinates (RAs) who have undergone extetisive selection and training. HDs play an important role in guiding, directing, and supporting the work of the RA's. Weekly staff meetings and frequent formal and informal oneon-one HD and RA interactions occur. HDs conduct formal job evaluations which impact retention decisions. The HD role has significant impact on RA activities similar to any hierarchical organization. Maturity of the subordinates is considered a central concept in this context

586

PERSONNEL PSYCHOLOGY

(e.g., Hoelting, 1980; KaufCman, 1968; Nkkerson & Haning;ton, 1971). Hersey and Blanchard (1982) describe the relevance of tbeir theory to educational settings and Vecchio (1987) used high school teachers and principles as subjects for his study. Thus, the sample offers and appropriate context in which to examine SLT. Measures Leader behavior. The Leader Behavior E>escdptive Questionnaire (LBDQ-XU, StogdiU & Coons, 1957) was used to measure task (initiation of structure) and relationship behavior (showing consideration). These measures have been widely used despite certain limitations (e.g., Kerr & Schriesheim, 1974; Kormaa, 1966). However, given their strong and direct links to SLX and their use in previous research (Vecchio, 1987), they were felt to provide a reasonable and appropriate test. The measure was completed by RAs in terms of the HD. iVaditional scoring of the LBE>Q-XII was used to yield the two leader behaviors. Items for each scale were summed to create a score for task and relationship behavior. The measures showed adequate internal consistency and a moderate interconelation (see Tkbie I). Maturity. Measures of job and psychological maturity were developed for this study. Psychometric procedures suggested Nunnally (1978) were used to develop the psychological maturity measure (see Blank, Weitzel, Blau, & Green, 1988 for a complete description of the measurement development procedure). Briefly, a set of 30 preliminary items generated to measure independence, ability to take re^mnsibility, and achievement motivation (ten for each dimension) were administered to a sample of 350 upper-class undergraduate students. Factor analysis (VARIMAX rotation; minimum loadings > .40), suggested a scale of 12 items, four from each dimension ofthe psychological maturity construct (alpha > .70). The 12 items were administered to 84 managers enrolled as part-time MBA students in two universities (universities other than the RA sample). These respondents described the matiuity of a randomly selected subordinate whom they currently or recently supervised in a full-time job setting. Factor ana^is (VARIMAX rotation) yielded a single factor (Cronbach's alpha > .70), with one item loading below .40. This independent sample provided initial sui^mrt for the reliability and generalizability of the scale. In the present study, the 12 item psychological maturity scale was completed by each RA who rated all other RAs in their hall (range of

WARREN BLANK E T AL. TABLE 1 Zero Ordar Correlation Variables Mean SD Range 1 n/a .13* .08 .04 .11* .40" .06 2 3 4 5 6

587

l.Performaocc 380.8 61.7 170-425 2.Supervtsor satisfaction 44.2 8.01 12-54 3. Work satisfaction 34.9 7.61 9-51 4.Usk behavior 39.4 4.83 23-50 5.Rclationship behavior 39.6 5.77 19-50 6.Psychological maturity 63.4 4.97 43-77 7.Job maturity 37.9 26.3 16-167 Note: Cronbach alpha on diagonal p<.05; "p<.001

(.78) .19** (.67) . 4 1 " . 1 9 " (.78) .54' . 1 9 " . 5 0 " (.84) .00 .10 - . 0 8 -.01 (.95) .00 .06 .04 -.01 .05 n/a

RA staff size was 4-35 with a median size of 9). This "peer rating" of psychological maturity was used because it avoided the bias of selfreports. It also provided a measure of maturity that was independent of the leader behavior descriptions (RAs) and the performance judgments (HDs) thus minimizing method variance problems. The psychological maturity measure was also completed by each RA for him/herself (a self rating) and by each HD for all his/her RAs (a leader rating). These additional rating sources were used to assess the validity and reliability of the psychological maturity measure (see below). Data from each of these three rating sources (peer, self, and leader) were factor analyzed (VARIMAX rotation). The three factor solutions indicated that the same 11 items as in the manager sample had adequate (i.e., > .40) loadings. Therefore, the 11 retained items were summed to create a score for the self (Cronbach alpha = .95) and leader rating (Cronbach alpha = .93). To derive a peer maturity rating, a mean score for each item was computed for the ratings provided by all peers. These mean scores were summed for the peer rating (Cronbach alpha = .95). The 11 items are: Tb what extent does : (1) follow through on job tasks, (2) act conscientiously on the job, (3) know what to do on the job without being told, (4) work hard on the job, (5) try hard to improve his/her performance on the job, (6) strive to do his/her best on the job, (7) make job related decisions on his/her own, (8) do extra on the job, (9) take care to do the job right, (10) do a thorough job on any tasks undertaken, (11) set hisylier own job goals. The peer rating of psychological maturity, which is used for all hypothesis testing, demonstrated adequate reliability and validity. The measure yielded substantially the same factor structure over five independent samples and repeatedly demonstrated strong internal consistency, suggesting good reliability. Its validity is supported by its strong, repeatable internal consistency (Runkel & McGrath, 1972, argue that

588

PERSONNEL PSYCHOLOGY

this demonstrates validity through concordance of results). The peer rating correlations with the leader rating of maturity (r = .40, p < .001) and the self rating (r = .23, p < .01) suggest reasonable convergent validity. In addition, if one assumes more mature subordinates tend to perform better, the relationship between peer ratings and performance (r = .40, p < .001) can be seen as a demonstration of predictive validity. TTie low intercorrelations between psychological maturity and job maturity and satisfaction also suggest discriminant validity (see Ikble 1). Psychological maturity is distinguishable for general affective states and Job experience. Finally, the measure has face validity. Thus, the weight of the evidence seems to indicate that the psychological maturity rating is a reasonable representation of the construct As a side note, use of the self and leader rating sources was explored in all hypothesis tests and results were similar to those using the peer rating. Job maturity was measured in terms of "past job ejqierience." Hersey and Blanchard (1982) suggest this is an important component of job maturity. Both overall job experience (the number of months in residence hall related work) and experience in their present job (number of months) were assessed. These two were summed to create single measure of job maturity. Performance. Each school provided a midyear overall performance rating for each RA in the study. Ratings were made by the HD for each RA in his/her hall on 45 specific job categories (five for each of nine general objectives) using afive-pointscale (1 = poor to 5 = excellent). Each school's confidentiality requirements resulted in their only reporting to the research team each RA's mean score of ail 45 items. Satisfaction. Respondents completed the Job Descriptive Index (Smith, Kendall, & Hulin, 1969) supervisor and work satisfaction scales as a measure of their satisfaction. These scales were chosen to represent affect toward the leader and the work itself. The JDI was also used by Vecchio (1987) in his test of SLT. Finally, to avoid confounding between-hall differences, the task behavior, relationship behavior, performance, supervisor satisfaction, and work satisfaction measures were standardized within-hall with a mean of 50 and a standard deviation of 10. Data Collection All data except the performance ratings were collected by questionnaire. In school 1, data were collected on site by the research team during meetings with HDs and RAs. In school 2, dau were collected by a

WARREN BLANK ETAL.

589

site liaison during staff meetings within each hall. The liaison was a topranking administrator within the residence hall system and was someone with whom the research team had close contact and a high level of confidence. The liaison followed the research team's carefully prepared instructions. Participants at both sites were informed that the instruments they were completing were to be used in a research stutfy regarding various behaviors of themselves and those with whom they worked. Anonymity of their responses was assured and it was explained that only aggregated results would be analyzed. Participants did not indicate any concerns to the members of the research team at site 1 or to the liaison at site 2 about their participation in the study. Follow-up data collection sessions were provided at both schools. An overall participation rate of 88% was achieved across both schools. Results Since the first two hypotheses address the relationship between leader behavior and leader effectiveness contingent on subordinate maturity, moderated regression was used to test for a significant interaction term for leader behavior and subordinate maturity in predicting subordinate performance and satisfaction. Separate tests of the hypotheses were conducted for psychological and job maturity. For the hypotheses related to task behavior (la and b) both maturity measures were dichotomized at the median. Low maturity was coded as one and high maturity was coded as zero. The same analyses were also conducted splitting both maturity measures into thirds and coding the upper and lower thirds as one and zero. Since the results were almost identical to the median split analyses, the median split findings are reported here. Results show little support for the hypotheses related to task behavior (see Tkbles 2A and 2B). Only in the case of psychological maturity and work satisfaction do we find a significant task behavior by maturity interaction term. The form of the interaction does support the hypothesis 2b (see Figure 3). Tksk behavior alone makes signi^nt, unique contributions to both types of satisfaction but not performance. Subordinate psycholc^cal maturity alone makes a significant, unique contribution to work satisfaction. For the hypotheses related to relationship behavior (2a and b), both maturity measures were divided into four quartiles. The middle two quartifes were considered as the middle range or moderate group (i.e., the group which needed more relationship behavior) and coded as a one. The lowest and highest quartiles were put into the extremes group (i.e., the group \i^h needed less relationship behavior) and coded as zero.

590

PERSONNEL PSYCHOLOGY TABLE 2A Regression Models for Task Behavior tuid Psychological MtUurity Tiak behavior p< .18 .01 .01 Psyo^logica] maturity P< .10 .45 .01 Ibsk X Psychological maturity p< J8 .42 .03

Dependent variable Performance Supervisor satisfaction Work satisfactioa

A-Square .09*" .17"* .06"*

TABLE2B Regression Modebfor Task Behavior and Job Maturity TSak 1 jebavior p< .01 .01 Job maturity p< .29 .98 Iksk X Job maturity p< .41 .98

Dependent variable Perfonnance Supervisor satisfaction Work satisfaction

A-Square .01 .17"' .03"

Note: p values are for the partial rvalue for the terms and indicate unique contribution to the model. *p<.01; *'p<.001

high maturity

Work Satisfaction

low maturity

Task Behavior Figtuie 3: Plot ofbsk x Psycfaologica] Maturity InteractitHi

WARREN BLA>fK ETAL. TABLE 3A Regression ModelsRelationship Behavior and Psychological Maturity Dependent vamMe Perfonnance Supervisor satisfoction Work satisfaction

591

A-Square .01 .04"'

1 Relationship Psychological 1 ^ X Psychological behavior maturity maturity


p< p< p<

.01 .01 TABLE3B

.54 .48

.62 .68

Repession ModelsRelationship Behavior and Job Maturity Dependent variable Perfonnance Supervisor satisfaction Work satisfaction Relationship behavior A-Square .01 29"' .04*
p<

Job maturity
p<

Relationship x Job maturity


p<

.01 .01

.86 .45

.70 .29

Note: p values are for the partial F value for the terms and indicate unique contribution to the model. p<oi; *"p<ooi

Again, alternate splits were examined (i.e., upper and lower sixths coded as extremes group, and third and fourth sixths coded as moderate group). Almost identical results were c^tained. The findings using quartile splits are reported here. Results (see Tables 3A and 3B) indicate no support for the hypwtheses. In no case was there a significant contribution by an interaction term. In addition, neither maturity measure alone made significant unique contributions for any of the criteria. The leader's use of relationship behavior, however, did make significant, unique contributions to both types of subordinate satisfaction. Tb examine the complex formulation of the model, four "leader styles" and four "maturity levels" were created. Leader styles were created by dividing both task and relationship behavior at the median into a high and low levels. Each respondent was then assigned one of four leader style scores: SI = high task, low relationship; S2 = high task, high relationship; S3 = low task, high relationship; and S4 = low task, low relationship. Maturity levels were created by dividing both psychological and job maturity into four qtiartiles and then combining each respondent's score to o-eate his/her maturity level. As noted above, there are ambiguities inherent in this approach. The quartile splits of the job and psycholc^cal maturity scores might not place a respondent in the same maturi^ level. For example, the respondent might be in the second quartile for job

592

PERSONNEL PSYCHOLOGY TABLE4 MANOVAFit Between Leader Style and Subordinate Maturity

Subordinate maturity Multivariate Performaoce Work satisfaction Supervisor satisfaction

Fit (n = 128) 3.83 36.51 44.81

Nojl (n =m) _ 3.80 33.95 43.80

4.96 .23 9.46 .28

3345 1347 1,347 1347

.02 .63 .002 .26

maturity and in the third quartiie for psychological maturity. Hersey and Blanchard (1982) provide no guidelines on what to do with respondents in such "ambiguous" maturity group. In the absence of clear guidelines, we established a procedure for hypothesis testing that provides the most liberal test of SLT. To test the fit/not fit hypothesis, a fit grouping variable was created. If the leader style exactly matched the prescribed style for both the subordinate's job and psychological maturity level, the fit grouping variable was set to indicate "fit" (e.g., if leader style was S2 and both job and psychological maturity placed the respondent in level two, fit was indicated). For respondents in an ambiguous maturity level, the lit grouping variable was set to indicate "fit" when leader style showed a prescribed match with either of the maturity variable levels (e.g., for SI leader style, afitwas indicated if eu/ier job or psychological maturity were in maturity level one). Otherwise the grouping variable was set to indicate "no fit." This provided a liberal test of SLT. One-way MANOVA was used with performance, supervisor and work satisfaction as the dependent variable andfit/notfitas the grouping variable. To test whether or not fit added to variance contributed by task behavior, relationship behavior, and subordinate maturity, analysis of covariance was run with fit as the grouping variable and task behavior, relationship behavior, and subordinate maturity as covariates. All combinations of fit were tested against all combinations of no fit. Results indicate a multivariate result for fit between leader style and subordinate maturity (see Ikble 4). An examination of the one-way analysis of variance results for each criterion variable indicates the significant MANOVA is primarily due to work satisfaction (F = 9.46, 1^47 df, p = .002). Although all means are in the expected direction, this result provides not much more si^^port for SLT. As a further test, the same analyses were conducted using the self- and leader-ratings of maturity. Hiere were no significant relationships using the self-rating. A significant multivariate result was found using the leader-rating with none of the one-way analyses significant.

WARREN BLANK ETAL.

593

EHscussion
Previous research (i.e., Hambleton & Gumpert, 1982; Vecchio, 1987) showed mixed support for SLT. These studies attempted to test the complex and, as described by Graeff (1983), potentially ambiguous and contradictory prescriptions of the theory. The present study tested the model's basic assumptions that underlie its prescriptions. This approach was viewed as a reasonable way to help us understand SLT and perhaps shed more light on the mixed findings of previous studies. These results reveal a lack of support for the basic assumptions that underlie SLT. In only one case, p^chological maturity and task behav* ior, did an interaction of leader behavior and subordinate maturity predict subordinate outcomes, i.e., work satisfaction. Given the rather extensive analyses, 12 regression models repeated for two different partitions of the data, these findings do not bolster our confidence in the assumptions that underlie the predictions of SLT. This is disappointing because of the intuitive appeal of the theory. The results also raise questions about the results reported in the Hambleton and Gumpert (1982) and Vecchio (1987) work. Although both efforts show methodological care, the reports of these works do not allow us to evaluate if their data would support SLTs underlying assumptions. Perhaps perfonnance and satisfaction are only affected when task and relationship behavior take on certain values related to maturity. However, more complex tests (similar to those used by Vecchio) were conducted to assess this for the data used in the present study. These findings showed only mixed support for the complex matches between maturity and leader behavior. Although a multivariate effect was found, it was due primarily to work satisfaction which is similar to the findings regarding the underlying assumptions. Perhaps the lack of support for SLT found in this research can be explained by the measurement procedures used. The Hambleton, Blanchard, and Hersey measure of maturity may be tapping some dimensions that interact with leader behaviors that the present study's maturity measures do not tap. A question is what is the Hambleton, Blanchard, and Hersey measure addressing? How would this measure converge with the measure used in this study? In addition, in Vecchio's work, performance and maturity were measured by the leader. In the present study performance data was provided by leaders and maturity data provided by peers. Perhaps tius explains the difference in results. In that case, SLT may have a narrower range of utiMty (i.e., its prescriptioiK regarding interactions of leader behavior

594

PERSONNEL PSYCHOLOGY

and maturity apply only when the leader's performance and maturity assessments of subordinates are used). Obviously more research is needed on SLT to help us understand it. At this point, given the results of the small line of research to which this study adds, it is difficult to be very optimistic about SLT. The model's underlying assumptions were not supported and the complex formulation of the model received mixed support. Of course, one must always be concerned that a reasonable examination was provided. In the present case, we had reason to believe the sample and the setting employed would be pertinent to the processes described by SLT. Moreover, the SLT authors invoked similar settings in presenting their theory, and Vecchio used an educational setting for his work. Future research could only benefit by examining SLT across different samples (Vecchio, 1987), but we have no reason to believe that the present sample was inappropriate. Similarly, the measures employed attempted to faithfully represent the constructs described by SLT. The psychological maturity measure was developed with care and appears to demonstrate reasonable levels of reliability and validity. The measure of job maturity can be criticized for being narrow in its scope, for example, focusing on job experience alone; but, this aspect of job maturity is central to the concept as defined by Hersey and Blanchard (1982) and thus provided at least a partial test of this aspect of SLT. Use of the Hambleton, Blanchard, and Hersey (1977) measure cannot be viewed as a more rigorous approach since that measure has yet to demonstrate validity and reliability comparable to the measures used here. The measures of leader behavior and leader effectiveness also appear to be fair extrapolations of the theory as it is stated and were the same as those used by Vecchio (1987). In addition, different respondents provided the leader behavior, maturity, satisfaction, and performance data which minimized method variance problems. Finally, the analyses used to test SLT attempted to give the hypotheses every chance to be confirmed. Therefore, this study is considered a reasonable, though certainly not definitive, test of the underlying assumptions of SLT. In a larger sense, the conduct of this study revealed a number of things about SLT that may be useful to future work in the area. The idea of leader behavior and subordinate maturity interacting to predict leader effectiveness cannot be abandoned prematurely. It was supported in one instance here. This study also suggests that both managers and peers could recognize psychological maturity in subordinates, and Hersey and Blanchard (1982) present persuasive arguments for the potential importance of this concept Given the few main effects for subordinate maturity found in this study, it still may well be the case that the role of subordinate maturity is best understood in terms of interactions with leader

WARREN BLANK ETAL.

595

actions. A number of issues concerning both maturity and leader behavior need to be resolved, however, before this understanding can be advanced. For example, there is a high correlation between task and relationship behavior (r = .50) although this is not atypical for the LBDQ (in the Vecchio, 1987, study these variables are correlated .52). These behaviors were developed to determine the smallest number of dimensions to describe leader behavior (Korman, 1966). Their significant relationship may reveal that the Hersey and Blanchard model oversimplifies leadership by using these general dimensions. More complex formulations of leader actions and their interactions with maturity deserve attention. When considering how to measure psychological maturity, its dimensionality is not clearly defined in SLT. Discussions of the concept suggest a multidimensional construct composed of aspects of achievement motivation, commitment to work goals, and willingness to take responsibility. The measure used in this work did not support multidimensionality. Questions also arise as to who should rate subordinate's maturity. Hersey and Blanchard (1982) suggest manager and self ratings of maturity. In this work, peer ratings were used to avoid response-response bias in tests of the hypotheses. Given the convergence of the manager and self ratings with the peer ratings and the poor convergence of the manager and self ratinp (r = .06), peer ratings might be a better measure for future research. Vecchio (1987) suggests the peer measure might simply be a popularity index. However, in this study the peer measure correlated strongly with managerial assessments of performance and with the manager's perception of maturity. This raises another issue about what the maturity concept really means. Is it simply another way to characterize performance or perceived willingness to perform? Future research on psychological maturity should address this issue. SLT describes job maturity in terms of experience and knowledge. This study did not find the experience component very useful, indicating one might do well to concentrate on the knowledge component of job maturity. In that the maturity constructs are central to SLT, these issues need to be resoh'ed. Before SLT can be adequately tested, measures of psychological and job maturity need to be examined in a variety of contexts. The psychological maturity measure used in this study shows preliminary strength as a measure that might be of use for additional research. Furthermore, Hersey and Blanchard (1982) argue that subordinate maturity is task specific. A subordinate may bie mature on some tasks and immature on others. The task focus in this research (i.e., resident advisor duties) may not have been siifficiently focused. Future work will

5%

PERSONNEL PSYCHCttXKJY

want to examine SLT in terms of more specific subtasks. Hiis is a selflimiting strategy, however. If SLT is only predictive at a veiy specific task level (e.g., bookkeeping versus rqx>rt writing), its generalizability may be too limited for it to be useful. F^iture research also needs to consider a longitudinal analysis of the interaction between leader behavior and subordinate maturity. Longitudinal data is typically viewed as a richer sourx% of analysis in behavioral science researdi. In the case of SLT it may be important since SLT can be interpreted as a within-group model that needs to be tested over time. More research is needed to clarify the degree of contributifm SLT makes to our understanding of leadership. It does seem clear that subordinate maturity is an important situational variable to be considered in leadership research. This study provides a measure to examine to facilitate that process. On the other hand, the singular focus of SLT on subordinate maturity may oversimplify the situational aspect of leadership. Hersey and Blanchard (1982) acknowledge that they only focus on one situational element to the exclusion of others. Yet they maintain that subordinate maturity is a central factor. Future leadership research needs to consider subordinate maturify along with other situational variables (e.g., position power, leader-member relations, etc.). In conclusion, the present work raises questions about SLT. Fimdamental measurement and design issues still need to be explored in order to ascertain the validity of SLT. The present study is an additional step in this process. A number of additional directions are suggested by this work. The widespread acceptance and use of situational leadership theory indicate it deserves more empirical attention. REFERENCES
Argyris C. (1957). PersonaWy and organization. New York: Harper and Row. Blank W, Weit2el J, Blau G, Green SG. (1988). A measure of psydiological maturity. Group and Organization Studies. 13,225-238. Cartwrigfat D, Zander A (Eds.). (1960). Group dynamics: Research and theory (2nd ed.). Evanston, IL: Harper & Row. Fiedler F. (1964). A contingency model of leadership effectiveness. In Berkowitz L (Ed.), Advances in experimental social psychology. New York: Academic Press. Fiedler F. (1967). A theory of leadership effectiveness. New York: McGTaw-HiU. GTaeff C. (1983). The situational leadership theory: A critical review. Academy of Managemeni Review, 8,285-291. Halpin A. (1959). The leadership behavioral school superintendents. Chicago: Midwest Administratira) Center. Hambleton R, Kanchard K, Hersey P. (1977). Maturity scalese^rating form. San Diego: Learning Resources Corporation. Hamtrfeton RK, Gumpen R. (1982). The validity of Hency and Blandutfd's theory of teader effectivenest. Group and Organization Studies, 7,225-242.

WARREN BLANK ET AL.

597

Heney P, Blanchard K. (1969). Life cycie theory of leadership. Trainir^and Devetopmeru Journal, 2,6-34. Hersey P, Blanchard K. (1982). Mantigemeru of organaational behavior (4th ed.). Englewood Cli&, NJ: Prcntice-HalL Hoelting F. (1980). Resident student devetopmenl. Normal, IL: National Entertainment and Campus Activities Association, Illinois State University. House RJ. (1971). A path-goal theory of leader effectiveness. Administrahve Science Quarterfy, 16,321-339. Katz D, Maccoby N, Morse N. (1950). Produaivity supervision, and morale in an office situation. Ann Arbor: University of Michigan, Institute for Social Research. Kauffinan J (Ed.). (1968). The student in hi^ier education. New Haven, CT: The Hazen Foundation. Kerr S, Schriesheim C (1974). Consideration, initiating stnicture, and organizational criteria: An update of Korman's 1966 review, PERSONNEL PSYCHOUXJY, 27, 555568. Korman A (1966). Consideration, initiating structure, and organizational criteria: A review, PERSONNEL PSYCHOIXWY, 66, 394-361. McQelland D, Atkinson J, Dark R, Lowell E. (1953). The achievement motive. New York: Appleton-Century-Crofts. NickersonD, Harrington J. (1971). The college student as counselor. Norvic, NY: Chronicle Guidance. Nunnally J. (1978). Psychometric theory (2nd ed.). New York: McGraw Hill. Runkel P, McGrath J. (1972). Research on human behavior A systematic guide to method. New York: Holt, Rinehart & Winston, Inc. Smith P, Kendall L, Huiin C. (1969). The measurement of satisfaction in yvork and retirement. Chicago: Rand McNally. StogdiU R, Coons A (Eds.). (1957). Leader behavior Its description and measurement (Research Monograph No. 88). Columbus, OH: Ohio State University, Bureau of Business Research. Vecchio R. (1987). Situational leadership theory: An examination of a prescriptive theory. Journal of Applied Psychohff, 72,444-451. Yukl G. (1981). Leadership in organizations. Englewood Qifis, NJ: Prentice-HalL

Das könnte Ihnen auch gefallen