Page |1

The Study of effects of employee Behavior on Predicting Customer Response

Page |2

Name: Reg# Day: Timing: Date of Submission: Submitted To:

Muhammad Zeeshan Motiwala 16824 Tuesday & Thursday 6 to 9 9/Aug/2011 Dr. Basheer

Page |3

Abstract The focus of this study is to describe the relationship between Employee Behavior and Customer Satisfaction. And to predict customer satisfaction at a given level of employee behavior and to test if there is a relationship between employee behavior and customer satisfaction. Hence the data of employee behavior of 30 OSS (One Stop Shop) Branches of PTCL is used. The objective of this data collection was to predict customer satisfaction / dissatisfaction using the data of employee behavior. First task was to evaluate the performance of employees based on certain categories like knowledge of employees, empathy, assurance etc. Likert rating scale was developed and employees were observed and rated on scale of 1-5. 5 rated the best performance where 1 indicates poor performance. After these ratings, customers were also asked to rate the employees to find out the relevancy or to link the customer satisfaction/dissatisfaction that is customer response with employee behavior. Information of employee behavior of 30 OSS branches of PTCL is contained in data file. Logistic regression model is used to identify the customer response of remaining 15 OSS Shops of PTCL.

Page |4

Table of Contents

Introduction Literature Review... Description of the data file. Problem Description... Application of technique Results Recommendations.. Conclusions

5 6 12 12 13 14 15 15

Page |5

Introduction

Logistic Regression Model Information on 30 OSS branches of PTCL is contained in data file. These file contains complete data of employee behavior of all OSS shops and customer satisfaction/dissatisfaction that is customer response data of 15 branches and based on the given data, logistic regression model is created to identify the customer response of remaining 15 OSS Shops of PTCL.

Logistic regression: Determines the impact of multiple independent variables presented simultaneously to predict membership of one or other of the two dependent variable categories.

Binary Logistic Regression Logistic Regression provides the following unique features: Hosmer-Lemeshow test of goodness of fit for the model Stepwise analyses Contrasts to define model parameterization Alternative cut points for classification Classification plots Model fitted on one set of cases to a held-out set of cases Saves predictions, residuals, and influence statistics Logistic regression is an increasingly popular statistical technique used to model the probability of discrete (i.e., binary or multinomial) outcomes. When properly applied, logistic regression analyses yield very powerful insights in to what attributes (i.e., variables) are more or less likely to predict event outcome in a population of interest. These models also show the extent to which changes in the values of the attributes may increase or decrease the predicted probability of event outcome. The dependent variable that is customer response is dichotomous. Independent variables are ordinal. So Logistic Regression technique is used and based on this technique incomplete data of customer response is completed thorugh prediction based on past data.

Page |6

Literature Review

Research 1 Factors influencing child welfare employee's turnover: Focusing on organizational culture and climate Many researches have been done on the issue of organizational culture and climate within the context of understanding the effectiveness of business .The aim of this research is to focus on the organizational factors related to child welfare area. This research paper focuses on a systematic analysis of organizational culture and climate and their effects on employee turnover in child welfare. These types of organizations required a stable workforce. However, high turnover rates in child welfare have been recognized as a major concerned. In this research Logistic Technique is applied Assumption 1 Employee's unwanted turnover negatively impacts remaining employees, clients, and the organization itself. This study assumes that employee's undesirable turnover might be reduced if organizational culture and climate changed positively. Assumption 2 Organizational culture is related to employee work attitudes and organizational performance. This study assumes that positive organizational culture, as defined by constructive behavioral expectations and norms, might be negatively associated with employee's intention to leave. Hypothesis There will be a negative association between organizational culture and employee's intention to leave in public child welfare agencies. Hypothesis 2 There will be a negative association between organizational climate and employee's intention to leave in public child welfare agencies.

This study uses the workforce retention study survey data that the New York State Social Work Education Consortium (SWEC) conducted and collected in two years (2002 and 2003).Among the 62counties in New York State, public child welfare agencies in 25 counties participated in the workforce retention study.

Dependent Variable=Intention to leave Independent Variable=Organizational Culture and Climate

Page |7

Logistic regression is employed to predict an employee's intention to leave using organizational culture and climate variables, after controlling for individual-level and countylevel characteristics. Regression analyses executed in this study have shown that organizational culture and climate are significant predictors of an employee's intention to leave. Specifically, public child welfare employees with higher values of organizational culture and climate have less intention to leave than those with lower values. Findings suggest that child welfare organizations can improve organizational culture and climate by supporting ER(Emphasis on Rewards) and reducing EE(Emotional Exhaustion).

Research 2 Will They Stay or Will They Go? The Role of Job Embeddedness in Predicting Turnover in Individualistic and Collectivistic Cultures

Turnover has been always a concern for any organization. Many studies have been done on turnover across culture. This research paper investigated the cross-cultural generalizability of the job embeddedness model (Mitchell & Lee, 2001) by examining turnover in an individualistic country (United States) and a collectivistic country (India). As the culture of USa and India is very different so these two countries are considered for this study. Call Centre industry was considered for this study. Data were collected from call centers in the spring and summer of 2006. In the United States, data were collected from three organizations. Survey invitations were sent to 486 employees, and responses were obtained from 344 employees, for a response rate of 70.78%. In India, data were collected from three locations of the same organization. Survey invitations were sent to 629 employees, and responses were obtained from 482 employees, for a response rate of 76.63%. Control Variables include External Prestige, Job Search Behavior, job Satisfaction, Job Alternatives and Organization commitment.

Hypothesis 1 proposed that organization embeddedness would account for variance in voluntary turnover after controlling for job satisfaction, organizational commitment, perceived job alternatives,

Page |8

and job search in both the United States and India. Hypothesis 2 proposed that community embeddedness would account for variance in voluntary turnover after controlling for job satisfaction, organizational commitment, perceived job alternatives, and job search in both the United States and India. Hypothesis 3 proposed that personjob fit would interact with country such that the relationship would be stronger in the United States than in India. Hypothesis 4 predicted that country would moderate the relationship between community fit and turnover such that community fit would predict turnover more strongly in the United States than in India. Hypothesis 5 predicted that the relationship between organization links and turnover would be moderated by country such that the relationship would be stronger in India than in the United States.

Turnover is a dichotomous variable. Therefore logistic regression is used to test the hypotheses. The goal of this study was to test the additional variance in turnover explained by job embeddedness over and above commonly used variables. Chi-square tests of model fit provided information on whether a model with the addition of a job embeddedness variable differed significantly from a model without the variable. When discussing tests of simple slopes for each country, there is a use of the odds ratio (b) for which values below 1 indicate a negative effect, values at 1 indicate no effect, and values above 1 indicate a positive effect.

The findings is that although job embeddedness is important in both cultures, as demonstrated by the overall contribution of organization and family job embeddedness, this research suggests that interventions targeting different dimensions of job embeddedness will be valuable for retention in different cultures. The finding that organization fit was more important in India has implications for both recruitment and organizational socialization. Second, findings suggest that enhancing family embeddedness could be an effective retention tool. There are a number of ways in which organizations might increase family embeddedness. In terms of family links, encouraging social links between organizational members could lead to increased family interactions with the organization.

Page |9

Research 3 Job stress models, depressive disorders and work performance of engineers in microelectronics industry Objectives Microelectronic engineers are considered valued human capital contributing substantial contribution toward economic development, but they may meet stressful work conditions in the framework of a globalized industry. The study aims at identifying risk factors of depressive disorders primarily based on job stress models, the DemandControlSupport and EVortReward Imbalance models, and at evaluating whether depressive disorders impair work performance in microelectronics engineers in Taiwan. Methods The casecontrol study was conducted among 678 microelectronics engineers, 452 controls and 226 cases with depressive disorders. In descriptive statistics, bivariate analysis was used to compare all variables between the cases and controls. The chisquare test was applied for categorical variables, and t-test was used to examine continuous variables. A two-tailed value of p < 0.05 was considered statistically significant. Results:- By hierarchical logistic regression, risk factors of depressive disorders are high demands, low work social support, and low frequency of physical exercise. Combining the two job stress models may have better predictive power for depressive disorders than adopting either model alone. Three multivariate linear regressions provide similar results indicating that depressive disorders are associated with impaired work performance in terms of absence, role limitation and social functioning limitation.

Research 4 Predicting Committed Behavior: Exchange Ideology and Pre-entry Perceived Organizational Support

A longitudinal field study conducted in a military setting examined the effects of exchange ideology, pre-entry Perceived Organizational Support, and their interaction, on initial and long-term committed behavior.

P a g e | 10

Hypothesis 1: Exchange ideology will be negatively related to initial and long-term indicators of committed behavior, namely to choice of and perseverance in combat duty. Hypothesis 2: Pre-entry POS will be positively related to initial and long-term indicators of committed behavior, namely to choice of and perseverance in combat duty. Hypothesis 3: Exchange ideology will (positively) moderate the relationships between preentry POS and initial and long-term indicators of committed behavior. Tests for difference between dependent correlations showed a significant difference between the correlations of exchange ideology and pre-entry POS both with choice of combat service, t(1047) = 2.22, p < .05, and with combat service perseverance, t(1001) = 2.57, p < .01. Note that because perseverance is a binary variable, significance tests may be somewhat biased, but we report below a logistic regression analysis that supports the results of the present test. Furthermore, exchange ideology was as effective as the biodata structured interview in predicting the two dependent variables. Tests for difference between dependent correlations indicated no significant differences between the correlations of exchange ideology and interview score both with choice of combat service, t(1045) = 1.02, p > .3, and with combat service perseverance, t(1001) = 0.62, p > .6. This was the result of this research.

Research 5 Do Paid Vacation and Sick Leave Promote Job Retention?

This study examines whether employee access to paid sick or vacation days is associated with the probability of job retention. Paid sick and vacation days are understudied relative to family leave, health insurance, and other employee benefits; yet, it is theoretically plausible that workers with paid leave are better able to address both routine and urgent family issues without having to quit a job or risk being fired. In fact, this is one of the compelling, but unsubstantiated, arguments in favor of several current state and federal initiatives aimed at mandating employers to provide paid leave. Hypothesis 1. The availability of paid sick leave should have a stronger relationship with job retention than the availability of paid vacation. While vacation days can sometimes be used by employees to address illness or other family responsibilities, it is paid sick leave that is designed specifically for this purpose. Hypothesis 2. The relationship between paid sick days (or both types of leave) and job

P a g e | 11

retention should be stronger for workers with greater family demands or more constrained family resources. For instance, women continue to carry a greater burden of child care and home making responsibilities, which may make paid leave days more relevant to their job retention than they are for men. In addition, being unmarried, having young children, and having low income could all affect a workers flexibility in taking an unpaid day off and increase the value of paid leave. logistic regression models has been used that helps in predicting the probability of both shortand long-term job retention as a function of access to paid leave, controlling for a large set of worker and job characteristics. Results suggest that having access to both types of leave, but particularly paid sick days, increases the probability of short- and long-term retention.

P a g e | 12

Description of the data file Variables Variables OSS(Branch) Employee Attitude Employee Knowledge Initiatives by Employees Good Listener Employees Patience Empathy Tone of Voice Assurance Responsiveness Customer ResponseSatisfaction/Dissatisfaction Measurement Scale Nominal Ordinal Ordinal Ordinal Ordinal Ordinal Ordinal Ordinal Ordinal Ordinal Nominal Independent/Dependent Independent Independent Independent Independent Independent Independent Independent Independent Independent Independent Dependent

There are total 12 variables. Out of which 1 variable is dependent that is Customer Response and this dependent variable has two level that is 0=Dissatisfaction and 1=Satisfaction.

Problem Description Problem Statement To study the effects of employee behavior on predicting customer response

P a g e | 13

Application of technique

Hosmer and Lemeshow Test Step 1 Chi-square 1.585 df 3 Sig. .663

This shows the significance level above 0.05 which means the fit is good as it is greater than 0.05.

Variables in the Equation B Step 1a Tone of Voice Constant 2.116 -7.264 S.E. 1.052 3.919 Wald 4.047 3.435 df 1 1 Sig. .044 .064 Exp(B) 8.294 .001

P a g e | 14

Variables in the Equation B Step 1a Tone of Voice Constant 2.116 -7.264 S.E. 1.052 3.919 Wald 4.047 3.435 df 1 1 Sig. .044 .064 Exp(B) 8.294 .001

Variables not in the Equation Score Step 1 Variables Employees Attitude Employee Knowledge Initiatives By Employees Good Listener Employees Patience Empathy Assurance Responsiveness Overall Statistics .004 .178 .241 .253 5.963 1 1 1 1 8 .948 .673 .624 .615 .651 1.319 1 .251 1.053 1.931 .576 df 1 1 1 Sig. .305 .165 .448

Results The Hosmer and Lemeshow table shows the significance level above 0.05 which means the fit is good as it is greater then 0.05.

The table Variables on the equation shows that tone of voice is the most important variable in predicting customer response and it is below 0.05.

The table not in the equation shows variables that are not included in the equation and their significance level are high that is above 0.05 and these variables are not playing any role in predicting customer response. Ln(/(1-)=-7.2764+2.116*Tone of Voice

P a g e | 15

Graph showing probability of satisfaction predicted. From this graph it is interpreted that it a prediction that almost 96% customer will be satisfied with Employee behavior of PECHS and Orangi Branch. While Only there is a 4% chance that customer will be satisfied with Azizabad branch.

Recommendations It is better to use Discriminant function and Multiple Regression Analysis to identify whether the independent variables like variables related to employee behavior has any correlation or dependency on customer response. More Employee Traits can be observed and then this technique will give appropriate results

Conclusion The conclusion is that tone of voice play a large role in predicting customer response and through the use of logistic regression technique Customer response of remaining branches is predicted using employee behavior and past data.