Statement of the Problem
The Pygmalion effect (a variant of self-fulfilling prophecy) has long been recognized in the fields of economics, psychology, political science, medicine, administrative organization, and education.  The Pygmalion hypothesis is confirmed if subjects perform and achieve by what is expected of them, independent of their abilities.  Since the publication of the first experiment employing this design in the classroom (Rosenthal & Jacobson, 1968), there has been a great deal of research and controversy concerning the nature of Pygmalion effect, its magnitude, and possible methodological flaws in the study of expectations (Barber & Silver, 1968; Elashoff & Snow, 1971; Thorndike, 1968).  Subsequent studies and meta –analyses have now established the fact that raising teachers’ expectations influences pupil achievement (Jussim & Harber, 2005, Trouilloud et al, 2002, Babad, 1993; Dusek, 1975; Hall & Merkel, 1985; Harris & Rosenthal, 1985; Rosenthal, 1985; 1991).  Many authors (Benabou and Tirole, 2003; Kierein & Gold, 2000; McNatt, 2000, Barling & Beattie, 1983; Crawford, Thomas & Fink, 1980; Eden, 1984,1990,1995; Eden & Ravid, 1982; Eden & Shani, 1982; Eden & Zuk, 1995; King, 1971,1974; Livingston, 1988; Miller & Turnbull, 1986; Oz & Eden, 1994) have noted that the Pygmalion effect operates in a variety of military and industrial settings. There is enough experimental data to suggest that self-fulfilling prophecy is true (Garner, 2005).  Ferraro (2005) has established the fact that social science theories can become self-fulfilling by shaping institutional designs and management practices, as well as social norms and expectations about behaviour. Despite this lively treatment of the authority of expectations in the educational and social psychological literature, management and organizational behavioral specialists have barely dealt with it in the sales field, where motivation is of prime concern.  In addition since the preponderance of research has been experimental in nature, there have been few efforts at modelling casual sequences.  Although experiments have been the favorite method among psychologists, one of its limitations is the number of factors that can be controlled by the researcher.  Large-scale surveys, while foregoing researcher-manipulated variables, allow for sophisticated causal modelling using a large number of predictor variables. The study is designed to examine the impact of the Pygmalion effect in motivating a sales force using causal modelling.  The paper adds to the sales literature since the psychological effects of sales managers’ expectations of salespersons is an area, which has not yet been widely researched and validated. The development of the causal model also adds to the literature since there is not much work in this area.

Theoretical Rationale
The underlying theoretical assumption is that the higher a sales manager’s expectations of salespersons’ performance, the greater the likelihood that sales performance will be high. If sales managers’ expectations are low, salesforce productivity is likely to be poor. This is called a self-fulfilling prophecy, often referred to as the Pygmalion effect. A self-fulfilling prophecy is based on an assessment of previous experience and it is a projection or anticipation that effectuates its own accuracy.  For example, because of previous experience with the performance of a category of employees (e.g., women, older males), a supervisor may anticipate whether or not a specific employee who fits that category will be effective.  Sales managers who hold low expectations of certain groups of persons may subtly influence their workers to perform at low levels or may view worker performance levels as inadequate even though the workers may be actually performing better than others who are expected to perform higher, but are actually performing less effectively.  Thus, low expectations can influence subsequent perception, which may, in turn, influence evaluation.  At the same time, low expectations that are accepted by workers or performers can influence perception, which may, in turn influence behavior.

Background of the Problem
Thomas’s sociological dictum that “things perceived as real are real in their consequences” (Merton, 1957) is applicable to self-perception as well as to interpersonal perceptions.  In the area of sales performance, persons who perceive themselves as good salespersons are likely to be better salespersons than those who do not see themselves as good salespeople.  At the same time, as the social interactionists have noted, our perceptions of ourselves are a consequence of the ways in which others treat us (Stryker, 1980).  Thus, it is to be expected that self-perceptions be conceptualised as mediators between the perceptions and treatment of the individual by others and the individual’s behavior (see figure 1).
 

The self-fulfilling model posits that sales managers’ expectations and treatment of salespersons will be influenced by background factors of the salespersons (e.g., age gender, ethnicity, experience).  These expectations and treatment by the sales manager will, in turn, influence the salespersons’ self-expectations, which will ultimately influence the performance of the salespersons. Direct evidence of the relationship between perceiver expectations and target performance comes from Brew and Hall (1964), Eden (1984), Rosenthal and Jacobson (1968), Sterdy and Kay (1962), and Sutton and Woodman (1989).  These studies and others within the literature on the Pygmalion effect support the fact that individuals who successfully meet performance expectations set for them are usually rewarded with salary increases, good grades, or promotions. If performance expectations are usually high or close to peoples’ own levels of aspiration, they will feel personal satisfaction at having achieved the goal.  Such positive outcomes lead to a higher level of aspiration and a more positive attitude toward the job. The existence of a relationship between perceiver expectations and perceiver behavior is supported by extensive research in the educational literature (Babad, 1990, 1993; Harris & Rosenthal, 1985; Rosenthal, 1974).  The “teachers’ pet” studies of Babad (1993) and Tal and Babad (1989,1990) suggest that the inequitable and favoristic division of a teacher’s affect (“playing favorites”) evokes strong reactions among the students, reducing students’ morale, satisfaction, and the desire on the part of non favorites to continue to study with the teacher who plays fovorites.

Statement of Hypotheses
The model of self-fulfilling prophecy casual chain in figure 1 is the basis for the hypothesis stated below:
H1: The higher the sales managers’ expectations of salespersons, the higher their sales persons’ performance
H2: The higher the sales managers’ expectations of salespersons, the higher their salespersons’ self-expectations.
H3: The higher the sales managers’ expectations of salespersons, the higher their salespersons’ evaluation of their own performance.
H4: The more positively salespersons perceive their treatment by sales managers, the higher their self-expectations.
H5: The more positively salespersons perceive their treatment by sales managers, the higher their performance.
H6: The higher the salespersons’ self-expectations, the higher their performance.

Methods

Sample
In order to test the hypotheses, survey data were collected on 193 sales employees from three retail sales companies located in New York City. These retail sales companies are involved in mass merchandising, upscale merchandising, and specialty retailing. The companies provided a sample population of 193 salaried sales employees working under 24 managers. Upon receipt of permission, the researcher made arrangements with supervisory personnel to distribute the questionnaires. Supervisor’s ratings were collected on 174 of the sales persons with 19 employees not evaluated by their managers. Full data on the Pygmalion variables were available for 174 salespersons.

The sales forces are predominantly female (60.6%) and young, with nearly two-third (64.2%) under 30 years of age. An additional one-quarter (26.9%) was between 31 and 40 years of age. Most (53.4%) were single; slightly less than one –third (31.1%) were married. The workforce was multiracial and multicultural. Nearly half (45.6%) the sales staff were of Hispanic or Latino background; slightly less than one-quarter (24.9%) were African-American; 14.5% were white; 10.4% were Asian. Nearly half (48.7%) of the sales staff had completed high school, 28.0% had attended some college, and 6.2% had college diplomas. Slightly more than one-tenth (11.9%) had not earned a high school diploma. The respondents had a mean of 3.23 (SD =4.07) years of sales experience. They had a mean of 2.70 (SD =2.82) years of experience with their present employer.

Instrumentation
The data were derived from the company personnel survey (CPS) comprising the employee perception of Supervisors’ Expectations Scale (SES), the Supervisors’ Authoritarian Behavior Scale (SABS), the Positive Achievement Motivation Behavior Scale (PAMBS), the Employee Self-expectation Scale (ESS), and the Employee Self-Evaluation Scale (ESES) (Kohli, 1985).  In addition to the self-perception and supervisor behavior scales, several demographic items were included in the CPS.  They were age bracket, sex, marital status, ethnicity, educational attainment, years in sales, and years with the present company.  The Supervisors’ Employee Evaluation Survey (SEES) contained two scales: the Supervisors’ Evaluation of Present Work Scale (SEPWS) and the Supervisors’ Expectancy scale (SES).

The Employee Expectation Scales (EES) is a 5-item instrument with four items using a 5-point Likert-type response mode.  A typical item is “I expected my sales performance to be.”  The response modes are as follows: 1 = far below most of my peers.  A fifth item asked respondents to rate themselves on their performance relative to their peers on a scale of 0% to 100 %, with 0% indicating that they are better than none of their peers and 100% indicating that they are better than all of their peers.  Because the fifth item used a different measuring scale than the other four items, EES items, scores were transformed into standard scores with means of 0 and standard deviations of 1.  Sutton and Woodman (1989) reported a reliability coefficient of .89 for the measure.  The Employee Evaluation of Present Work (EEPW) and Employee Perception of Supervisor’s Expectation Scale (EPSES) used the same items as the EES; however, the respondents were asked to report their perceptions of actual quality of their present work in the former, and their perceptions of supervisor expectations for their work in the later.

Supervisors’ behavior was measured through the CPS by having employees report on their supervisors’ behavior.  Three scales refined by Kohli (1985) were used: the Supervisor Arbitrary and Punitive Behavior Scale (SAPBS), the Supervisor Contingent Approving Behavior Scale (SCABS), and the Supervisor Achievement –Oriented Behavior Scale (SAOBS).  All supervisor behavior scales were keyed to 5-point Likert-type response modes from “very false” (1) to “very true” (5).  All of the supervisor scales reported by Kohli (1985) began with generic pronoun ‘‘He.” This was changed to a stem of “My supervisor.”  This change brought the items into conformance with the mandates of the publication Manual of the American Psychological Association, 4th edition (1994), which advises the avoidance of “generic he” construct.  It also reflects the fact that women are in supervision, and hence, is more accurate.  This SAPBS consists of 8 items.  A typical item is  “ My supervisor speaks in a manner not to be questioned”.  The theoretical range of scores is from 8 to 40, with 8 indicating low levels of arbitrary and punitive behavior, and 40 indicating the high levels.  Kohli (1985) improved the scale and reported an alpha reliability coefficient of .83. Supervisor Contingent Approving Behavior Scale (SCABS) was measured by a 4-item scale originally developed by House (House & Mitchell, 1974).  A typical item is “ My supervisor gives clear recognition for outstanding work”.  The theoretical range of scores is from 4 to 20, with 4 indicating low levels of contingent approving behaviour and 20 indicating high levels.  The measure has been found to discriminate between contingent approving and contingent disapproving behavior (Fulk & Wendler, 1982).  Kohli (1985) improved the scale further and reported an alpha reliability coefficient of .95 for the scale.

The SOABS contained 4 items, a typical item is “ My supervisor shows approval of me when I put forth my best efforts”.  It has the same theoretical range as the SCABS; Kohli (1985) improved the scale and reported an alpha reliability coefficient of .70.
Supervisors’ behavior was scaled into two variables with eight items each.  One scale was termed “Authoritarianism,” since it dealt with the supervisors’ authoritative behavior, such as, “My supervisor rules with an iron hand.”  The other was named “Positive Achievement Orientation Behavior (PAMB) since the items that comprised it dealt with supervisors’ attempts to positively influence sales persons’ performance. A typical item was: “My supervisor gives me recognition for improvement in my performance”.  Both scales had alpha reliability coefficients higher than .90, indicating strong internal consistency. The two supervisor behavior scales were developed by taking the 16 supervisor items and conducting a factor analysis on them using principal components factor extraction and varimax rotation.

Data collection
The researcher received from each participating supervisor a list of the salespersons that he or she supervised.  These lists were then coded with four-digit code numbers that were keyed as follows: the thousand and hundreds column digits identified the supervisor and the tens and unit columns identified the individual salesperson.  For example ID No. 1307 indicated salesperson number 07 who was supervised by supervisor number 13.  That is, salesperson ID numbers were nested within supervisor IDs.  Therefore, for each participating salesperson, there were two completed surveys with identical numbers: the CPS completed by the salesperson and the SEES completed by the supervisor.  Only the researcher had the key that matched the respondent to the code numbers.  The Pygmalion variables all had alpha coefficients above .90, indicating high levels of internal consistency.

Data Analysis
Data analysis follows a correlation design.  All hypotheses were tested using Pearson correlations and partial correlations, controlling for background variables for gender, marital status, ethnicity, length of employment, and years of sales experience.
Finally the model was tested using a series of multiple regression equations. The dependent variable was the score on the SEPWS.  In the model, the relationship between salespersons’ self-expectations and salespersons’ performance was assessed with a zero-order correlation. The indirect effects of sales managers’ treatment of salespersons’ performance was tested using a hierarchical multiple regression analysis, entering in the sales managers’ treatment of salespersons’ variables and salespersons’ sales-expectations prior to the entry of sales managers’ expectations. A similar regression was run using salespersons’ background variables as a predictor.  In addition, sales managers’ treatment of salespersons was regressed on salespersons’ background variables and sales managers’ expectations. These analyses were conducted recursively until path coefficients were assigned to all arrows in the model.  

Results

Hypothesis Testing
Six hypotheses of the study are tested. The first three hypotheses are concerned with the relationships between supervisors’ expectations, their ratings of the performance of sales employees, employee self-expectations, and employee performance self-ratings. Table 1 presents the Pearson and partial correlations for supervisors’ expectations and employee outcomes. The partial correlation coefficients were computed controlling for gender, marital status, ethnicity, sales experience, length of employment, and years of sales experience.

 

The Pearson correlation comparing supervisor expectations and supervisor ratings indicates a very strong positive relationship (r = .72, p <. 01. The partial correlation coefficient was similar (r=. 67, p<. 01). The data support the hypothesis that supervisor expectations and ratings are positively related. The Pearson correlation comparing supervisor expectations and employee self-expectations indicate a moderate positive relationship (r=. 45, P <. 01). The partial correlation coefficient was similar (r=. 49, p<. 01). The data support the hypothesis that supervisor expectations and employee self-ratings are positively related. The Pearson correlation comparing supervisor expectations and employee self-ratings indicate a moderate positive relationship (r=. 42, p<. 01). The partial correlation coefficient was similar (r= .45, p < .01). The data support the hypothesis that supervisor expectations and employee self-ratings are positively related.

           
Hypotheses 4 through 6 examine relationships between supervisor behaviors and supervisor’ rating of employees, employee self-expectations, and employee performance self-ratings, respectively. Table 2 presents the Pearson and partial correlations controlling for the effects of gender, marital status, ethnicity, sales experience, length of employment, and years of sales experience. There was no relationship between supervisor authoritarianism and their ratings of employee performance. However, the higher the ratings of supervisors on PAMB, the higher the ratings of employees’ performance (Pearson r =. 31, p<. 01; and partial r =. 26, p<. 01). The data supports hypothesis 4 partially.

           
Hypotheses 5, which specified relationship between employee self-expectation and supervisor behaviors was not supported for supervisor authoritarianism. Although, there was a significant relationship using a Pearson correlation, the correlation attenuated to nonsignificance when background factors were controlled. Hypothesis 5 was, however, supported for PAMB (r= .40, p <. 01; partial r = .43, p <. .01).
 

Hypothesis 6 was also not supported for authoritarianism. However, it was supported for PAMB (r =. 37, p < .01; partial r  = .40, p < .01).

Assessing the Model

The self-fulfilling prophecy model was tested using a series of multiple regressions. The dependent variable was the score on the Supervisors’ Evaluation of Present Work Scale (SEPWS).  The first stage of the model indicated sales managers’ expectations and treatment of employees being influenced by employee backgrounds.  Table 3 presents the results of the regressions analysis for supervisor expectations.  Coefficients used in the model are typed in boldfaced print.

Table 3 Summaries of Multiple Regressions of Sales Managers’ Expectations on Background Variables (N = 130)
 

Note, R =.  48, R2 = .23, p < .01.  Bold/italicised coefficients were used in the model in figure 2.  * P < .05.  **P < .01.

The regression analysis indicates that the background variables accounted for 23% of the variance in supervisors’ expectations, which was significant (R=. 48, p<. 01).  However, when examining the bivariate relationships, sales experience, was the only variable significantly related to expectations (b = .26, p< .01).  The greater the length of sales experience, the higher the supervisor’s expectations.

Table 4 contain the results of the regression analysis for sales managers’ authoritarianism.  The data in Table 2 indicate that employee background variables accounted for 39% of the variance in supervisors’ authoritarianism (R = .63, p < .01).  Bivariate analyses indicate that the greater the sales experience of the salespersons, the higher the perceived level of supervisors’ authoritarianism (b = .47, p< .01), and the greater the supervisors’ knowledge of the employee, the lower the authoritarianism (b = -.57, p< .01).
 

Table 4: Summary of Regression Analysis of Sales Managers’ Authoritarianism Background Variables (N = 130)
 

Note, R =.  .63, R2 = .39, p < .01.  Bold/italicised coefficients were used in the model in figure 2. * P < .05.  **P < .01.

Table 5 presents the results of the multiple regression of supervisors’ PAMB by the background variables.  The data indicate that the background variables accounted for 14% of the variance in supervisor PAMB (R = .38, p < .01).  The only variable that accounted for a significant proportion of the variance is the supervisors’ knowledge of the employee (b = .24, p< .05). 

Table 5: Summary of Multiple Regression of Sales Managers’ Positive Achievement Orientation Behavior on Background Variables (N = 130)
 

 
Note, R =.  .38, R2 = .14, p < .01.  Bold/italicised coefficients were used in the model in figure 2.   * P < .05.  **P < .01.

The next link in the causal chain was salespersons’ self-expectations.  They were conceived as being influenced by sales managers’ expectations, sales managers’ authoritarianism, and sales managers’ PAMB.  This link was tested using a hierarchical multiple regression analysis, first entering the background variables followed by the supervisor variables.  The results of the final solution are presented in Table 6. The data in table 6 indicate that the demographic background variables accounted for 35% of the total variance in employee self-expectations (R= .60, p< .01).  The longer the supervisor knew the employee, the lower the self-expectations (b = -.19, p< .05); and the greater the amount of sales experience, the higher the self-expectations (b = .32, p< .01).  The supervisory variables added 22% to the explained variance(R = .76, p < .01).  Supervisor expectations were positively related to self-expectations (b = .38, p< .01) and supervisors’ PAMB was positively related to self-expectations (b = .28, p < .01).

Table 6 Summaries of Multiple Regressions of Employees’ Self-expectation Background Variables and Supervisor Variables (N = 130)

           
















Note, Step 1:  R2 = .35, p < .01.  Step 2: R2 = .22, p <.  01.  Bold/italicised coefficients were used in the model in figure 2.  * P < .05.  **P < .01.

In order to analyze the direct effects of the supervisor variables on employee performance as rated by supervisors, while controlling for employee background variables, a hierarchical multiple regression analysis was performed by regressing supervisor ratings of sales employees.  The results are presented in Table 7.

Table 7:  Summary of Multiple Regressions of Employees’ Performance Ratings by        Supervisors on Background Variables and Supervisor Variables  (N = 130)
 

Note, Step 1:  R2 = .27, p < .01.  Step 2: R2 = .34, p <.  01.  Bold/italicised coefficients were used in the model in figure 2.  * P < .05.  **P < .01.

Although background variables accounted for 27% of the variance (r= .52, p < .01), no single background was significantly related to the employer ratings of sales employees.  The supervisor variables added 34% of explained variance to the equation (R = .78, p < .01).  Literally all of the variance was accounted for by supervisors’ expectations (b =. 61, p < .01).  The final multiple regression analysis is a three-step regression of employee ratings over background variables, supervisor variables, and employee self-expectations in hierarchical order.  The data are presented in Table 8.

Table 8 Summaries of Multiple Regressions of Employees’ Performance Ratings by Supervisor on Background, Supervisor Variables, and Self-expectation (N = 130).
 

 Note, Step 1:  R2 = .27, p < .01.  Step 2:R2 =.  34, p <.  01. Step 3:  R2 = .03, p < .01. Bold/italicised coefficients were used in the model in figure 2.  * P < .05.  **P < .01.

The background variables accounted for 27% of the variance in supervisors’ performance ratings of employees.  No single background variable accounted for a significant proportion of the variance in supervisors’ ratings of employees.  The supervisor variables added 34% to the explained variance in the equation.  Supervisors’ expectation was significantly related to performance ratings (b =. 51, p < .01).  Employee self-expectations added 3% to the total explained variance and was itself significantly related to supervisors’ rating (b - .27, p < .01).

On the basis of series of multiple regressions analyses, the model presented in this study can be reanalysed using path coefficients to indicate direct and indirect effects of the set of predictor variables.  The model is presented in figure 2.
 

The path model in figure 2 presents the direct and indirect influence of the predictor variables on salespersons’ performance ratings by the supervisor.  The data clearly show that supervisors’ behaviors and expectations are strongly affected by salespersons’ backgrounds, especially their prior experience, length of service, and supervisors’ prior knowledge of employee.  The direct effects of supervisors’ expectations on their ratings of employee performance are extremely strong (.61).  The indirect effects are substantial, since salespersons’ self-expectations are influenced by supervisors’ expectations (.38) and are directly related to salespersons’ performance (.27).  Thus, the data suggest that supervisors’ expectations strongly influence their ratings of employees, a direct effect.  Supervisors’ expectations also influence employee’s ratings indirectly by influencing salespersons’ self-expectations, which, in turn, influence their performance ratings.

A secondary causal chain can be viewed in terms of the links between salespersons’ backgrounds, supervisors’ PAMB, salespersons’ self-expectations, and salespersons’ performance.  Salespersons’ backgrounds influence supervisor PAMB (.38), which in turn, influence sales persons’ self-expectations (.28), which influence performance ratings (.27).  The direct effect of supervisor PAMB on performance ratings is weak and nonsignificant (.12). The third potential causal chain begins with the influence of salespersons’ backgrounds on supervisor authoritarianism (.67). However, supervisors’ authoritarianism has no significant direct (.08) or indirect (.05 with self-expectations) effects on salespersons’ performance.

In summary, path analysis indicates that the prime source of influence on salespersons’ performance ratings was supervisors’ expectations, which had strong direct effects and moderate indirect effects through their influence on salespersons’ self-expectations.  Supervisor PAMB had weak direct effects and weak to moderate indirect effects on salespersons’ performance through its influence on salespersons’ self-expectations.

Conclusion
The data from this study support the model of the self-fulfilling prophecy.  However, those effects need to be specified.   Supervisors’ behaviors are influenced by the background of their salespersons.  They tend to be more authoritarian with Hispanics, males, and single persons.  Their authoritarianism, however, does not influence employee self-expectations, nor does it influence their ratings of their salespersons.  Supervisor PAMB was influenced only by the supervisors’ knowledge of the employee.  That is, the greater the familiarity the supervisors had with particular employees, the greater the probability that they would use PAMB.  PAMB, however, was positively related to employee self-expectations.  Supervisors’ expectations increased with employees’ sales experience, length of employment, and supervisors’ knowledge of the employee.  Supervisors’ expectations influenced employee self – expectations.  The data indicate that to the extent that supervisors engaged in positive motivational behaviors and expected greater performance from their salespersons, employee self-expectations were increased, which influenced supervisors’ performance ratings independently of the influence of other factors, such as the direct and indirect effects of supervisor expectations and supervisor authoritarianism.  That is, the data suggest a significant Pygmalion effect that accounted for about 7% of the variance in employee working ratings.
 

Although the data are robust, a major limitation of this study is that it employed a survey rather than an experimental design.  This means that the researcher could not control expectations, even though statistical controls were employed.  Because of the lack of experimental control, variables outside the model unknown to the researcher may have influenced performance ratings and salespersons’ self-expectations.  Data indicate that knowledge of the employees and employees’ prior sales experience influenced supervisors’ behavior toward treatment of salespersons and the expectations of their performance.  Although such variables were employed as statistical controls, unmeasured variables, such as friendship may have influenced the relationships in the chain.

Even though causal modelling was used in the predictive model, it is based on a correlational design.  According to Maclver (1942), “where there is causation there is also correlation, but were there is correlation there may be no corresponding causation” (p.92).  That is, correlation is a necessary, but not sufficient, condition of causation.  The causal chains in the path model are actually linkages of correlations.  The imputation of causality is a leap of faith. A second limitation of this study was that the criterion of performance was supervisors’ ratings, which have already been noted, were highly influenced by expectations.  Even though direct effects of expectations on supervisors’ ratings were controlled, the validity of the findings would have been enhanced if an objective external criterion of performance could have been used.  It was not possible to employ such a criterion in this study since the salespersons came from three different organizations selling different product and services.

Despite the caveats issued in the previous section, the data strongly suggest that supervisors’ behaviors and expectations influenced salespersons’ self-expectations, which in turn, influenced their performance as rated by the supervisors.  The two variables that were most important were supervisors’ using positive methods of motivation for their salesforce and their high expectations of salespersons’ performance.  The implication is that high expectations plus creating a positive environment through encouragement and positive reinforcement of appropriate sales behaviors will improve employees’ performance. The data suggest that this is especially true for those employees who have little sales experience and who are new on the job.
Although expectation effects have been studied in education, the military, and in business, few studied were directly related to sales personnel.  The data from prior studies suggest that expectation effects can be generalized to almost any hierarchical social relationship.  W. I. Thomas’ dictum that whatever is perceived as real in its consequences has been verified empirically by hundreds of studies.

Therefore, researcher needs to study to a greater extent how supervisors’ expectations are translated into supervisors’ behaviors that influence employee self-expectations.  Perhaps, high supervisors’ expectations are necessary for increased employee performance.  However, what need to be investigated are the intervening variables between supervisor expectations and employee expectations.  In this study, the variable of positive achievement motivation behaviors was treated as one such intervening variable.  Other variables may include socioemotional support of the employee by the supervisor, task support, communication behaviours, and use of various training techniques. The model, while having empirical support, nevertheless tends to be somewhat elliptical. The linkages between supervisor expectations and employee self-expectations need to be further specified in subsequent studies.  We do not know how such effects occur.
 

References
 

Ajay, K. (1985). Some unexplored supervisory behaviors and their influence on salespeople’s role clarity, specific self-esteem, job satisfaction and motivation. Journal of Marketing Research, 22, 424-433.

American Psychological Association.  (1994).  Publication manual of the American Psychological Association.  Washington, DC.

Babad, E.  (1990).  Measuring and changing teachers’ differential behaviour as perceived by student and teachers.  Journal of educational Psychology, 82,683-690.

Babad, E, (1993).  Pygmalion C25 years after interpersonal expectations.  In P. D.  Blanck (Ed.), Interpersonal expectations (pp.125-153).  Cambridge, England:  Cambridge University press.

Barber, T., & Silver, M (1968).  Fact fiction and the experimenter bias effect.  Psychological Bulletin Monograph Supplement, 70 (6, part 2), 1 –29.

Barling, J., & Beattie, R.  (1983).  Self- efficiency beliefs and sales performance.  Journal of Organization Behaviour Management, 5, 42-51.

Benabou, R & Tirole, J (2003), Intrinsic and Extrinsic Motivation, Review of Economic Studies 70, 489-520.

Berlew, D., & Hall, T. (1964). The management of tensions in organizations: Some preliminary findings. Industrial management review, 6, 31-40.

Crawford, K. S., Thomas, E. D., & Fink, J.  J. (1980).  Pygmalion at sea: Improving the work effectiveness of low performs.  The journal of Applied Behaviour Science, 16, 482-503.

Dusek, J. (1975).  Do teachers bias children’s learning?  Review of Educational Research, 45, 661-684.

Eden, D.  (1984) Self-fulfilling prophecy as a management tool: Harnessing Pygmalion.  Academy of Management Review, 9, 64-73.

Eden, D.  (1990).  Pygmalion without interpersonal contrast effects: whole group gains from raising expectations.  Journal of Applied Psychology, 76, 81-98.

Eden, D.  (1995).  Self-fulfilling prophecy and gender:  Can women be Pygmalion and Galatea?  Journal of Applied Psychology, 80, 253-269.

Eden, D., & Ravid, G.  (1982).  Pygmalion vs. self-expectancy: effects of instructor and self-expectancy on trainee performance.  Organization Behaviour and Human prformance, 30, 351-364

Eden, D., & Shani, A. B.  (1982).  Pygmalion goes to boot camp:  Expectancy, leadership, and trainee performance.  Journal of Applied Psychology, 67, 195-199.

Eden, D., & Zuk, Y.  (1995). Seasickness as a self-fulfilling prophecy:  Raising self-efficiency to boost performance at sea.  Journal of Applied Psychology, 80, 628-635.

Elashoof, J., & Snow, R.  (1971).  Pygmalion reconsidered.  Worthington, OH:  Charles Jones.

Ferraro, F (2005), Economics Language and Assumptions: How theories can become self-fulfilling, Academy of Management Review, Volume 30, No 1, 8-24.

Fulk, J., & Wendler, E.R.  (1982). Dimensionality of leader-subordinate interactions:  A path-goal investigation.  Organizational Behaviour and Human Performance, 30, 241-64.

Garner, E (2005). The Pygmalion Effect. Enzine Articles. Retrieved April 23, 2006, from http:// w.w.w. Enzinearticles.com

Hall, V., & Merkel, S.  (1985).  Teacher expectancy effects and educational psychology.  In J. B. Dusek (Ed.), Teacher experiences (pp.  67-92). Hillsdale, NJ: Lawrence Erlbaum Associates.

Haris, M. J., & Rosenthal, R.  (1985). Mediation of interpersonal expectancy effects:  31 meta-analyses.  Psychological Bulletin, 97, 363-386.

House, R., & Mitchell, T. R.  (1974, autumn).  Path-goal theory of leadership.  Journal of Contemporary Business, 3, 81-97.

Jussim, L & Harber, D. (2005). Teacher Expectations and Self- fulfilling Prophecies: Knowns and unknowns resolved and unresolved controversies. Personality and Social Psychology Review, 9 (2), 131-155.

Kierein, M & Gold, M  (200), Pygmalion in Work Organizations: a meta-analysis, Journal of Organization Behavior 21, 913-928

King, A. S.  (1971).  Self-fulfilling prophecy in training the hard-core:  Supervisor expectations and the underprivileged workers’ performance.  Social Science Quarterly, 52, 369-378.

King, A. S.  (1974). Expectation effects in organization change.  Administrative Science Quarterly, 19, 221-230.

Livingston, J. S.  (1988).  Pygmalion in management:  HBR Classic and retrospective commentary.  Harvard Business Review, 66 (5), 121-130.

Maclver, R. M.  (1942).  Social causation.  New York Haper.

McNatt, B (2000). Ancient Pygmalion Joins Contemporary Management: A Meta-Analysis of the Result. Journal of Applied Psychology 85, 314-322.

Metron, R. K.  (1957).  Social theory and social structure.  New York Free Press.

Miller, D., & Turnull, W.  (1986). Expectancies and interpersonal process.  Annual Review of Psychology, 37, 233-256.

Oz, S., & Eden, D.  (1994). Restraining the golem:  Boosting performance by changing the interpretation of low scores.  Journal of Applied Psychology, 79,744-754.

Rosenthal, R (1974). On the social psychology of the self-fulfilling prophecy: Further Evidence of the Pygmalion effects and their Mediating Mechanisms. New York: MSS Modular.

Rosenthal, R.  (1985). From unconscious experimenter bias to teacher expectancy effects.  In j. Dusek (ED.), Teacher expectancies (pp. 37-65).  Hillsdale, NJ:  Lawrence Eflbaum & Associates.

Rosenthal, R.  (1991).  Teacher expectancy effects:  a brief update 25 years after the Pygmalion experiment.  Journal of Research in Education, 1, 3-12.

Rosenthal, R., & Jacobson, L.  (1968).  Pygmalion in the classroom:  teacher expectations and the pupil’s intellectual development.  New York: Holt, Rineheart & Winston.

Sterdy, A., & Kay, E.  (1962).  The effects of goal difficulty on performance.  Lynn, MA:  General Electric Company (mimeo).

Stryker, S.  (1980).  Symbolic interactionism.  Menlo Park, CA: Benjamin / Cummings.

Sutton, C. D., & Woodman, R. W.  (1989).  Pygmalion goes to work:  The effects of supervisor expectations in a retail setting.  Journal of Applied Psychology, 74, 943-950.

Tal, Z., & Babad, E.  (1989).  The teacher’s pet phenomenon as viewed by Israeli teachers and students.  Elementary School Journal, 90, 99-110.

Tal, Z., & Babad, E. (1990). The teacher’s pet phenomenon; Rate of occurrence, correlates, and psychological costs. Journal of Educational psychology, 82 (4), 637-645

Thorndike, R. L.  (1968). Review of Pygmalion in the classroom by R. Rosenthal & L. Jackobson.  American Educational Research Journal, 5,708-711.

Trouilloud, D., Sarrazin, P., Martinek, T., Guillet, E (2002). The influence of teacher expectations on student achievement in physical education class. European Journal of Social Psychology, volume 32, issue 5, pp 591-607.