Impact of Performance Appraisal on Employee Motivation and Productivity in Turkish Forest Products Industry : A Structural Equation Modeling Analysis

The present study investigates the infl uence of performance appraisal on employee motivation and productivity in Turkish forest products industry using a structural equation model. The questionnaires prepared for this purpose were applied to 432 people that work in 14 forest products industry businesses operating throughout Turkey. A total of nine hypotheses were established to determine the relationship between performance appraisal and employee motivation and productivity. The results indicated that fi ve hypotheses were accepted, while four of them were rejected. In the light of the fi ndings of this study, it may be generally said that the effect of performance appraisal on employee motivation and productivity was high. In other words, performance appraisal was found to be a major factor on employee motivation and productivity. The present study is expected to help managers and decision makers in selecting suitable motivating factors in order to retain and satisfy their employees. Consequently, this study will fi ll an important gap in the area of performance appraisal in the forest products industry by structural equation modeling and presents the opportunity for further studies.


INTRODUCTION 1. UVOD
Employee performance plays an important role in achieving organizational goals.In other words, employee performance is directly linked to the success of any organization.Therefore, organizations try to improve their productivity by providing an improvement in employee performance throughout the world.In this regard, various strategies have been developed to measure employee performance; one of them is performance appraisal (PA) (Wholey, 1999;Gichuhi et al., 2013).
The PA can be defi ned as a strategic and important approach that requires a regular inspection of the performance of employees in the organizations for performing the assigned tasks and responsibilities (Salau et al., 2014).This strategy aims to measure, assess and improve the performance of employees to achieve various objectives (Singh et al., 2010).The majority of organizations also use the PA approach to determine the factors such as salary increases, promotions, the need for individual development and training of employees (Gürbüz and Dikmenli, 2007).PA is thus accepted to be an important tool in order to improve both organizational performance and individual performance of employees.Such advantages of PA make it an important tool to reach organizational goals in many sectors (Vasset et al., 2011).
It is also important to emphasize that human capital (employees) in the organizations is considered as one of the most critical factors in gaining sustainable competitive advantage in the competitive situation of today's world.The organizations that have skilled and motivated employees may be regarded as advantageous in reaching such goals.Thus, many organizations endeavor to improve the human resource management and its functions (Chang and Han, 2006;Fakhimi and Raisy, 2013).In this respect, PA may be regarded as one of major functions of human resources managers.
PA intends to evaluate the performance of employees as objectively as possible.Kumbhar (2011) reported that a well-designed PA system helps develop employee performance-related criteria, provides a feedback mechanism and enables a more equitable reward system.It was also stated that employees' productivity increases when a fair system is available for PA in any organizations.In this regard, PA undertakes a critical task for organizations in order to reach their strategic objectives (Kumbhar, 2011).On the other hand, some researchers have pointed out that a rewarding system, applied as a result of PA, contributes to an increase in employee mo-tivation through infl uencing directly the performance of employees (Prowse and Prowse, 2009;Ochoti et al., 2012).Ali and Ahmed (2009) mentioned that rewards are managerial tools that help to reach organizational goals by affecting individual or group behavior.Therefore, most organizations use various types of rewards such as pay, promotion, bonuses in order to motivate employees and increase employees performance (Ali and Ahmed, 2009).
Furthermore, PA is a continuing procedure that measures expertise and achievements of employees based on an acceptable accuracy and equality.The employees are subjected to an appraisal process periodically after they are hired.The appraisal activities serve as a guide for managers in making a decision on employees' current job performance, awards, career goals and other job-related actions based on performance (Gürbüz and Dikmenli, 2007).However, the factors such as differences in the level of education, hiring temporary workers, employment, work experience, time pressures, and shift work might affect the appraisal process as well as employee motivation and productivity (Vasset et al., 2011).Jelačić et al. (2008) found that the factors such as employment assurance, job organization and workplace activity affect employee motivation in forest products industry.In another study, no differences were found in terms of motivation factors among employees of Slovenian and Croatian wood industry companies.Besides, it was stated that managers should ensure the safety of employees and pay attention to their mutual relations (Kropivšek et al., 2011).Hitka and Štipalova (2011) compared the motivation of employees in the wood industry and other manufacturing enterprises for various categories throughout Slovakia.It was observed that the diversity of the importance of motivational factors is less signifi cant in the category of employees.In another study on the impact of the economic crisis period on the motivation of forest products industry employees, it concluded that employee motivation does not differ in the period before and after the crisis (Hitka et al., 2014).Lorincova (2016) mentioned that increasing professional satisfaction of employees in the furniture industry may lead to an increase in motivation and business productivity.
In addition to the above mentioned investigations, a comprehensive analysis that considers the infl uence of PA on employee motivation and productivity is needed.Structural equation modeling (SEM) is an appropriate approach for this purpose.SEM is a versatile multivariate statistical approach that enables re-searchers to effectively evaluate the relationships between observed and latent (unobserved) variables (Chan et al., 2007).The emergence of this approach has been regarded as a signifi cant development and it has been widely employed in various scientifi c disciplines (Xiong et al., 2015).SEM is a combination of regression analysis, correlation, factor analysis, and path analysis.In comparison to other multivariate techniques, SEM is capable of estimating multiple and interrelated relationships and representing unobserved concepts in these relationships.This method also considers measurement errors in estimation.Further, it defi nes a model explaining an entire set of relationships (Xiong et al., 2014).Because of these advantages, SEM has been increasingly employed in many disciplines, such as banking industry (Osibanjo et al., 2014), education sector (Phin, 2015), service sector (Suki, 2014), business safety (Hsu et al., 2012), tourism sector (Hallak et al., 2012), food industry (Booth et al., 2013), and health sector (Bazarganipour et al., 2013), for uncovering the relationships among different kinds of variables.

Objectives and hypotheses 1.1. Ciljevi i hipoteze
Although there are many applications to examine the effect of PA on motivation and productivity in various disciplines, limited information is available to investigate the infl uence of PA on employee motivation and productivity in forest products industry by the SEM.This study, therefore, analyses the relationships among various factors that affect motivation and productivity of employees of Turkish forest products industry by means of SEM to create a model of factors that both directly and indirectly affect employee motivation and productivity.
Additional objectives of this study were: -to detect the factors that increase employee motivation and productivity; -to investigate the relationship between employee motivation and productivity; -to assess the effect of the PA on the motivation and productivity of employees.In this regard, the following hypotheses were established based on independent and dependent variables.
H 1 1, 2, 3, 4: There is a relationship between employee motivation and the purpose of PA (ppa), PA criteria (pac), PA practices (pap), Feedback in PA (fpa); H 1 5, 6, 7, 8: There is a relationship between employee productivity and the purpose of PA (ppa), PA criteria (pac), PA practices (pap), Feedback in PA (fpa); H 1 9: There is a relationship between employee motivation and employee productivity.
In this perspective, we proposed the model composed of all the hypotheses that describe the relationships between two variables, as depicted in Figure 1.
From the model in Fig. 1, it can be seen that the purpose of the PA (ppa), PA criteria (pac), feedback in PA (fpa), and PA practices (pap) are considered as independent variables.

Population and sample 2.1. Populacija i uzorak
The participants of the present investigation are composed of businesses that have ISO 9001:2008 Quality Management System (QMS) certifi cations and operate in forest products industry throughout Turkey.A previous study conducted by Serin (2004) has revealed that businesses that employ more than 100 employees give more signifi cance to total quality management applications because they have a more institutional structure.Fourteen companies that meet these criteria agreed to implement the study.The survey was conducted between March 2015 and December 2015.
A mong the 14 participating businesses, 7 deal with furniture, 6 with board products and 1 with paper.Surveys were administered to 432 people, including senior, middle and lower level employees.

Data collection instruments 2.2. Način prikupljanja podataka
A questionnaire prepared by compiling data from the surveys found in the literature was used in this study (Tarlığ, 2006;Yılmaz, 2006).The survey is comprised of two sections.The fi rst section contains 7 questions related to personnel information, while the other section includes Likert-type (5 scales) 51 questions related to the PA, employee motivation and employee productivity.After the surveys were completed, the data obtained were entered into SPSS (Statistical Software for Social Sciences).Analyses for reliability and validity were then performed by AMOS (Analysis of Moment Structures) package program.

Demographic characteristics Demografska obilježja ispitanika
In this section, demographic information of the participants such as age, education, gender, marital status, wages, working experience and position in the company were determined.Results regarding demographic information for participant employees are given in Table 1.
Table 1 shows that 49.4 % of the participants had vocational school and higher education levels.The majority of participants (82 %) were male, while 70.4 % were married.In addition, 43.8 % of the participants were employees, 49.1 % of them had 6 years and over of working experience.

Vrijednost i pouzdanost
For the present study, Cronbach alpha coeffi cient was used for reliability analysis.According to the results of the analysis conducted, general Cronbach alpha coeffi cient was found to be 0.950.This coeffi cient varies within a range of 0 to 1, and a value over 0.90 means an excellent agreement (Kalaycı, 2009).
Moreover, factor analysis was carried out to measure structural validity.However, it is important to state that conformity of the data set for the factor analysis is tested with Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy prior to the analysis.KMO should be over 0.5 for validity (Sharma, 1996).This value was found to be 0.917 for this study (Bartlett's Test of Sphericity Sig.:0.001).Thus, it can be said that the study was suitable for performing factor analysis.In the application of the factor analysis, varimax orthogonal rotation method and basic components analysis were considered.In this respect, factors that include less than three variables and variables that load on more than one factor were discarded.Furthermore, eigenvalues statistics was used in order to determine the number of factors.Dunteman (1989) reported that factors with an eigenvalue higher than 1 are considered signifi cant.Table 2 presents the results of exploratory factor analysis related to the scale.
Upon application of factor analysis, from the scale consisting of 51 items, 18 items that do not match the structure of the scale or load on more than one factor were discarded.The remaining 33 items formed a structure that consists of 6 sub-factors with eigenvalues higher than 1.Total expressiveness of this six-factored structure was found to be 59.942 %.

Measurement model 3.3. Model mjerenja
Measurement model structure obtained as a result of factor analysis was tested by confi rmatory factor analysis for investigating the structure validity.In order to test the conformity of the model, confi rmatory factor analysis was performed by the AMOS.Standard prediction values, t values and reliability level of the variables in the model are given in Table 3.
In analyzing the standard prediction values of the variables, it was shown that the values vary between 0.491 and 0.846.Besides, the t values of these predic- tions were signifi cant at 0.05 signifi cance level.Therefore, it can be concluded that the validity of the model was provided.Two types of reliability measures were used in the measurement model: the announced variance of the factors and the reliability coeffi cients of the factors.While the announced variance estimates of the factors show the total variance value explained by the observed variables of each factor, the reliability coeffi cients of the factors show the intrinsic reliabil-ity of the factors.As shown in Tables 2 and 3, the announced variance values of the factors are above the required lower limit (0.50 %) and the reliability coeffi cients of the factors are again above the lower limit (0.70).
The correlations among the purpose of performance appraisal, performance appraisal criteria, feedback in performance appraisal, performance appraisal practices, productivity and motivation are presented in Table 4 along with a discriminant validity test.Table 5 gives the fi t indexes belonging to the resulting fi ndings as a result of the testing with the structural equation modeling analysis of the developed model.
As the fi t indexes of the structural model were analyzed, it was seen that the model is within acceptable limits determined by the following literature (Anderson and Gerbing, 1984; Hancock and Mueller, 2006;Al-Refaie, 2015).AMOS program output of the model is shown in Figure 2.  Hypothesis was examined after the validity of the model was provided.The results of hypothesis tests are given in Table 6.
When the results of the model were analyzed, it was seen that fi ve of the nine hypothesis (p <0.05) were accepted, while four of them were rejected.
H 1 1 examines the relationship between the purpose of PA and employee motivation.According to the model, there is no signifi cant effect of the purpose of PA on employee motivation (estimate=0.156;p>0.05).Therefore, H 1 1 hypothesis is rejected.
H 1 2 examines the relationship between the criteria of PA and employee motivation.According to the model, there is a positive and signifi cant impact of the criteria of PA on employee motivation (estimate=0.276;p<0.05).Therefore, H 1 2 hypothesis is accepted.Accordingly, an increase of a standard unit in the criteria of PA constitutes an increase of 0.276 standard unit on employee motivation.
H 1 3 examines the relationship between the practices of PA and employee motivation.According to the model, there is a negative and signifi cant impact of the practices of PA on employee motivation (esti-mate=-0.157;p<0.01).Therefore, H 1 3 hypothesis is accepted.Accordingly, an increase of a standard unit in the practices of PA constitutes a decrease of 0.157 standard unit on employee motivation.
H 1 4 examines the relationship between the feedback of PA and employee motivation.According to the model, there is a positive and signifi cant impact of the feedback of PA on employee motivation (esti-mate=0.278;p<0.05).Therefore, H 1 4 hypothesis is accepted.Accordingly, an increase of a standard unit in the feedback of PA constitutes an increase of 0.276 standard unit on employee motivation.
H 1 5 examines the relationship between the purpose of PA and employee productivity.According to the model, there is a positive and signifi cant impact of the purpose of PA on employee productivity (esti-mate=0.187;p<0.01).Therefore, H 1 5 hypothesis is ac-cepted.Accordingly, an increase of a standard unit in the purpose of PA constitutes an increase of 0.187 standard unit on employee productivity.
H 1 6 examines the relationship between the criteria of PA and employee productivity.According to the model, there is no signifi cant effect of the criteria of PA on employee productivity (estimate=0.005;p<0.05).Therefore, H 1 6 hypothesis is rejected.H 1 7 examines the relationship between the practices of PA and employee productivity.According to the model, there is no signifi cant effect of the practices of PA on employee productivity (estimate=0.074;p<0.05).Therefore, H 1 7 hypothesis is rejected.
H 1 8 examines the relationship between the feedback of PA and employee productivity.According to the model, there is no signifi cant effect of the feedback of PA on employee productivity (estimate=-0.022;p<0.05).Therefore, H 1 8 hypothesis is rejected.
H 1 9 examines the relationship between employee motivation and employee productivity.According to the model, there is a positive and signifi cant impact of employee motivation on employee productivity (esti-mate=0.365;p<0.01).Therefore, H 1 9 hypothesis is accepted.Accordingly, an increase of a standard unit in employee motivation constitutes an increase of 0.365 standard unit on employee productivity.

ZAKLJUČAK
In this study, the effects of PA applications applied in businesses that operate in the forest products industry on employee motivation and productivity were examined by means of a structural equation model.Upon analyzing the obtained results, it could be seen that the PA had a positive impact on employee productivity, while it did not have a signifi cant impact on employee motivation.There was a positive effect of the criteria of PA on employee motivation, while there was no signifi cant impact on employee productivity.There was a negative effect of the practices of PA on employee motivation, while there was no signifi cant impact on employee productivity.The feedback studies in PA have been found to have a positive impact on employee motivation.However, it was shown that its impact on employee productivity was not signifi cant.Besides, it was determined that employee motivation had a positive effect on employee productivity.
In the light of these results, the following suggestions have been made for the forest products sector.
-The purpose of the employee PA should be explained to the employees very thoroughly and thus their hesitations should be reduced or avoided.It should take into consideration that an increase in the employee motivation can be provided by using PA.-PA criteria are required so that employees can reach achievable targets.Besides, it should be explained how to help the employees to provide the criteria.It should be noted that in this way an increase in employee productivity together with employee motivation can also be provided.-In order to achieve the purpose of the PA, in making any decision about the employees, appraisal results should be used as the data source for education, career, job rotation, pricing, etc. -By sharing the results of the PA studies with the employees, elimination of errors and defi ciencies can be provided.In-service training activities can also be organized to solve employee errors.

Figure 2
Figure 2 AMOS program output of the analyzed model Slika 2. Rezultat programa AMOS za analizirani model

Table 5
Fit indexes belonging to the results of the structural equation model Tablica 5. Indeksi koji pripadaju rezultatima modela strukturne jednadžbe