Self-efficacy and organizational efficacy effect on employee attrition in IT sector

Abstract

Purpose

Information technology (IT) plays a major part in the economy growth of any country. Attrition has been an important issue that influences the performance of the company. The increasing levels of attrition are the most immediate concern that IT companies are facing. The present study will help in knowing the influence of self-efficacy (SE) and organizational efficacy (OE) on the turnover of an organization.

Design/methodology/approach

This study employed an online survey method for data collection using a sample size of 250. Reliability and validity of the measurement scales were ensured, and hypotheses developed were tested through PLS-SEM using SMART PLS.

Findings

A large number of employees are the youth of age group 20–40 years. Work satisfaction within the organization is the main cause of attrition. Lack of contentment along with the biasness in the organizations hampers the dissatisfaction with the nature of work between employees. Thus, effective, timely communication of organizational policies and goals helps to pound satisfaction with work within the organization, which could result in lower employee attrition.

Research limitations/implications

The research has been conducted as not much research works have been found that show the relationship between SE and OE on employee attrition in the IT sector. The present study will help in further theory development and finding new aspects.

Originality/value

The research is the first of its kind, to the best of the authors’ knowledge, that shows the relationship of SE and OE to employee attrition.

Citation

Bhardwaj, S. and Sharma, S.K. (2024), "Self-efficacy and organizational efficacy effect on employee attrition in IT sector", Business Analyst Journal, Vol. 45 No. 1, pp. 1-10. https://doi.org/10.1108/BAJ-07-2023-0057

1. Introduction

Information technology (IT) industry is a major sector for Indian economy, but due to the ongoing circumstances since the past few years, there is a recession in the industry (Paul et al., 2023) [1]; however, we can also say that with time, the attrition will be decreasing in the coming future (Balasundaram et al., 2023) [1]. Self-efficacy (SE) is the one of the important factor for attrition. It is an individual’s belief that an individual is capable of performing a particular assigned task successfully (Bandura, 1977). Think of SE as a kind of self-confidence (Kanter, 2006) or a task-specific version of self-esteem (Brockner & Glynn, 1988). SE has powerful effects on learning, motivation and performance because people try to learn and perform only those tasks that they believe they will be able to perform successfully. Wood and Bandura (1989) proposed a persuasive discussion concluding that high-level SE will raise the personal performance. SE grows with the passage of time; employees learn how to deal with the conflicts that occur in workplace.

Bandura (1997 and 2000) suggested that SE might not be the sole phenomenon at the individual level but that construct may apply to several levels in the organization. At the group or team level, the construct is termed as collective efficacy. Collective or group efficacy represents the collective efficacy belief that it is competent and can achieve the desired goals (Zepp, Ohlert, & Eys, 2023). The organizational efficacy has gained less attention than its individual counterparts SE and group efficacy. Recently, few empirical research studies have applied the construct at the organizational level (Bohn, 2002; Bohn, Douthitt, & Eggers, 2006). Gully, Incalcaterra, Joshi, and Beubien (2002) found that there is a significant relationship between collective efficacy and performance of a group or organization. In the light of the aforementioned empirical research studies, it can be said that the organizational efficacy (OE) is one of the most important construct for the organization to maintain its performance midst adverse environments (Ntontis et al., 2021). OE is a construct that helps an organization to uplift its performance by showing belief in its capability to perform well in a competitive environment (Bachmann, Maran, Furtner, Brem, & Welte, 2021).

2. Literature review and hypothesis

SE is one’s belief in self-abilities to implement and achieve a given task effectively (Bandura, 1986). Social cognitive theory states that human functioning is determined by personal, behavioral and environmental factors, which all interrelate with each other. SE impacts one’s conduct in several ways. Individual with high SE faith is more risk-taking and acknowledges difficult tasks; however, people with low SE tend to escape difficult tasks and take easy ones to perform according to their efficacy principles. High or low SE faith also affects the rate of the efforts taken to accomplish achievement. Bandura and Adams (1997) observed that there is an important correlation between SE and successive task operation. Many empirical research studies have been found which shows the significant relationship between SE and attrition, for example, Gupta, Bhattacharya, Pratima, and Coelho (2017) in their study on industrial and commercial training stated that there is a significant relationship between SE, locus of control and turnover intention (TI). (Chu, Zhang, & Li, 2021) in their study on nurses show that there is a significant effect of SE on attrition. The research of Yan, Mansor, Choo, and Abdullah (2021), on 406 employees of four- and five-star hotels in China and analyzed using structural equation modeling (SEM), stated the same thing, where SE plays a role in reducing the level of TI. Kim, Park, and Kim (2020) in their study on South Korea’s casino industry also stated the significant relationship between SE and TI. Thus, in light of the aforementioned arguments, we propose that SE has a significant impact on attrition of employees. Therefore, we posit the following hypothesis:

H01.

There is a significant relationship between SE and employee attrition.

Social cognitive theory extends the formation of human agency to collective groups. Collective Efficacy (CE) is a construct that targets a “a group’s common belief in its combined ability to arrange and perform the courses of action necessary to construct given levels of acquirement”. Bandura (1986) suggested group efficacy as an expansion of SE and recommended that group efficacy processes (Cao, Tao, Wang, Tarhini, & Xia , 2020) are more than just the sum of individual efficacy levels among the group. Many empirical research studies have been found which shows the significant relationship between OE and attrition in different sectors, for example, Yan et al. (2021) in their study on the role of top leaders also stated the significant relation between the two. Thus, based on the argument, we propose that OE has a significant impact on attrition of employees.

H02.

There is a significant relationship between OE and employee attrition.

3. Methodology

3.1 Sampling and data collection

The study employed an online survey using a structured questionnaire for data collection.

Minimum sample size requirements were ensured, and it was found to be 213 through power analysis (at 0.05 level of significance and 0.05 effect size) using G*Power software (Faul, Erdfelder, Lang, & Buchner, 2007). Kline (2015) suggests at least 200 observations for reliable SEM estimates. The snowball sampling method was used, and a sample of 250 was utilized for the present study, which is well above the minimum sample size requirement.

The respondents comprise mainly IT employees above 20 years of age. More than half of the respondents were male respondents (62.9%), whereas female respondents constitute 37.1% of the whole data; minority of the respondents lie in the age group of 20–30 years (70.1%).

3.2 Measurement

To judge the relationship between SE, OE and attrition intent, three structured research instruments have been used. The questionnaire contains the statements on different aspects of the research problem and is based on a five-point Likert-type scale (where 5 = strongly agree, 4 = agree, 3 = neutral, 2 = disagree and 1 = strongly disagree). The structured questionnaire used for the survey is divided into three sections. The first section consists of 10 statements that constitute SE scale that measures the SE insight of an individual about his capabilities of accomplishing the assigned duties and responsibilities. The SE scale is developed by Ralf Schwarzer and Matthias Jerusalem. The scale was shaped to assess a general sense of apparent SE with the plan in mind to forecast coping with daily hassles as well as alteration after experiencing all kinds of stressful life events. SE has been recognized to be one of the indicators of psychological well-being and provides dependable results across multicultural contexts (Rimm & Jerusalem, 1999). Every individual do not behave optimally even though they know what to do. The rationale of the SE questionnaire is that it measures an individual’s prospect of how that person is likely to perform in a wide variety of situations (Tipton & Worthington, 1984). This instrument is open to use for research purpose.

The second part of the questionnaire is framed to measure OE, with 17 statements. OE is the social cognition that people in an organization have about the beliefs or insights of an organization’s capabilities. The epicenter of the OE is that people within the organizations have a sense of the capabilities of an organization to face challenges and generate solutions and their collective perceptions from a perception of OE (Bohn, 2010). The scale that is used in the present research has been developed by James D. Bohn.

The third part of the questionnaire constitutes attrition intent and consists of seven dimensions comparing 34 statements, namely (1) satisfaction with pay, (2) nature of work, (3) satisfaction with supervisor, (4) organizational commitment, (5) justice environment, (6) turnover intention and (7) perceived alternative employment opportunities. These seven dimensions have been used by Naresh Khattri and Pawan Budhiyar. Attrition could be a by-product of all the above discussed factors if an employee is dissatisfied with any of them.

3.3 Data analysis

The reliability and validity assessment of the measurement scales were ensured through confirmatory factor analysis, and hypotheses developed were tested through PLS structural equation modeling using. SmartPLS.

Since the data were not following multivariate normality (Table 2), the study preferred variance-based SEM over covariance-based SEM (Hair, Risher, Sarsted, & Ringle, 2019). A full collinearity assessment was also done, and it was found that all the variance inflation factor (VIF) values for the constructs were well below the conservative threshold of 3.33 (Kock, 2015), which establishes that the present study is free from common method bias.

4. Results

Table 1 shows that minority of the respondents lie in the age group of 20–30 years (70.1%), i.e. sample led by young generation respondents. As far as department is concerned, absolute majority of the respondents are from finance department (56.8%), followed by respondents from sales department (9.9%).

4.1 PLS-SEM results

Measurement model assessment results. All the constructs were reflectively measured. Therefore, the first step is to check whether the item loadings of all the construct indicators are above the recommended threshold of 0.708 (Hair et al., 2019). The results show that item loadings for every construct’s indicators were above 0.708, except indicators (SE9: loading 0.68), (OE5: loading 0.68), (OE10: loading 0.66), (OE13: loading 0.62), (PS3: loading 0.69), (SS3: loading 0.69), (JE5: loading 0.68), (TI3: loading 0.69), (TI5: loading 0.66), (PAEO3: loading 0.67), (PAEO4: loading 0.65), (PAEO5: loading 0.64). Before eliminating any item with loadings lower than the threshold value, checking the average variance extracted (AVE) value for that construct is suggested. Hair et al. (2019) recommend that one must retain that indicator if AVE is above 0.50.

The results of the measurement model assessment (Table 2) show that all the measurement instruments were reliable. Other reliability measures like composite reliability, Cronbach’s alpha and rho A are also provided in Table 2. The convergent validity of measurement scales was ensured by assessing whether AVE values were above the threshold of 0.5. From the results, it was evident that all the AVE values were well above the threshold.

Furthermore, discriminant validity (Table 3) was evaluated by heterotrait-monotrait (HTMT) ratio methods. HTMT is defined as the mean value of the item correlations across constructs. It is suggested that the HTMT ratio gives better results in evaluating whether the constructs are distinct or not. The results show that all the HTMT values were below 0.85, which is well within the advisable thresholds (Henseler, Ringle, & Sarstedt, 2015).

Structural model assessment. The VIF values were below 3.33 (Hair et al., 2019), suggesting no multicollinearity among the constructs. The results of the structural model assessment (Table 4) exhibit that both the hypotheses were accepted.

In the standardized estimates (β) column, the results indicate that the SE is related to the attrition intent in a statistically significant way (standardized regression weights: β = 0.364, p = 0.432) and SE is a predictor of employee attrition intent. It can be said that attrition intent is 36% explained by SE. It is concluded that SE is a cause of attrition intent because there is a positive and significant relationship between SE and attrition intent (Table 4); hence, the alternative hypothesis that there is a significant relationship between SE and attrition intent is hereby accepted. Standard regression weights (β) are standardized coefficients estimates and are independent of the units in which all variables are measured. These standardized coefficients allow the researcher to directly compare the relative relationship between each independent variable and the dependent variable.

In the standardized regression weights (β) column, the results indicate that the OE is related to the attrition intent in a statistically significant way (standardized estimates: β = 0.320, p = 0.001) and OE is a strong predictor of employee attrition intent. It can be concluded that 32% variance in the attrition intent is explained by OE. It stated that there is a positive and significant relationship between OE and attrition intent (Table 4); hence, the alternative hypothesis that there is a significant relationship between OE and attrition intent is hereby accepted (see Figure 1).

Predictive relevance of the model. The coefficient of determination (R²) was 52% for attrition which shows that the model has moderate explanatory power. Henseler et al. (2015) suggest that R² is not enough to determine the model’s prediction power as it is in-sample prediction (see Table 5). Therefore, we should see Q², a measure that incorporates out-of-the-sample predictive relevance of the model as well. Q² for avoidance behavior was 0.269, suggesting moderate predictive relevance. The model was found to be good (=1− SRMR), with an Standadized Root Mean Square Residual (SRMR) of 0.026.

5. Discussion

It can be noted that the standardized regression weight (β) on the all factors are both significant and positively related to attrition intent, i.e. organizational commitment, justice environment, satisfaction with pay, satisfaction with nature of work, satisfaction with supervision, turnover intention and perceived alternative employment opportunity. It can be said that attrition intent can be minimized by enhancing satisfaction with the nature of work between the employees and organization. Satisfaction with the nature of work has emerged as an important factor which can help the organizations to control the high attrition intent among the employees. Biased and lack of contentment in the organizations hamper the satisfaction with the nature of work for employees and the organization. Thus, effective, precise and timely communication of organizational policies and goals helps to inculcate satisfaction with the nature of work between employees and organization, which could result in lower employee attrition. As the study revealed that OE is a strong predictor of attrition intent in the organization, research showed that satisfaction with the nature of work is one of the most important factors in OE. Organizational commitment could help organizations to curb higher attrition intent among the employees.

In the standardized estimates (β) column, the results indicated that the SE is related to the attrition intent in a statistically significant way and SE is a predictor of employee attrition intent (Table 4). This can be said that 36% of the contribution in attrition intent is explained by SE. Rest of the 64% of the variance is attributed to the other factors. These factors could be many. Few possible factors are better future prospects outside, centralized decision and mismatch with the job description. It is concluded that SE is a cause of attrition intent because there is a positive and significant relationship between SE and attrition intent; hence, the null hypothesis that there is no significant relationship between SE and attrition intent was rejected. In the standardized regression weights (β) column, the results indicate that the OE is related to the attrition intent in a statistically significant way and OE is a strong predictor of employee attrition intent (Table 4). It can be concluded that 32% of the variance in attrition intent is contributed by the OE: rest of the 68% of the variance is attributed to the other factors.

A link between the SE perceptions of the employees and job descriptions should be made. Implementing the measures to find out the SE of the employees during the recruitment and selection would help the organizations to establish the linkage between SE belief of the prospective employee and the job for which he is to be selected. OE beliefs of the employees create group cohesiveness which leads to organizational commitment among the employees. Thus, a positive causal relationship has been found between SE, OE and attrition intent of the respondents of the Indian IT industry.

Research is an unending voyage with unfolded mysteries beneath. There is always a scope of improvization. Research studies always have some constraints and impediments, and this study also suffered from limitations. In majority of social science research due to time and money constraints, a small sample has been used, which is not sufficient to represent the universe. Generalization and implementation of the results of the study cannot be done on other organizations or geographical area. The study is restricted only to Pune, Bangalore and Delhi (including NCR) due to time and resource constraints. The possibility of respondents’ biasness in responding to the questionnaire cannot be ruled out. As this study is confined only to find out the implications of SE and OE for employee attrition, literature review suggested that there are numerous issues that might have an impact on attrition.

Figures

Figure 1

Path model

Table 1

Demographic profile of the respondents

Demographic features	Percentage
Gender
Male	62.9
Female	37.1
Total	100.0
Age of respondents (in years)
20–40	93.5
40–50 and above	6.5
Total	100.0
Experience of respondents (in years)
0–4	59.3
4–8 and above	40.7
Total	100.0
Departments
Technical/Technology	3.6
Marketing	8.1
Sales	9.9
Product management	8.1
Customer service	8.1
Human resources	5.4
Finance	56.8
Total	100.0
Designation
Senior level	20.7
Middle level	42.2
Junior level	37.1
Total	100.0

Source(s): Primary source

Table 2

Item loadings, reliability and convergent validity

Constructs/Indicator	Loadings	Cronbach’s alpha	Rho_A	Composite reliability	Average variance extracted (AVE)
SE1	0.85	0.87	0.88	0.86	0.87
SE3	0.82
SE6	0.80
SE2	0.76
SE5	0.70
SE9	0.68
OE1	0.80	0.86	0.87	0.87	0.88
OE3	0.78
OE4	0.77
OE6	0.76
OE9	0.71
OE5	0.68
OE10	0.66
OE13	0.62
PS1	0.70	0.82	0.88	0.84	0.84
PS3	0.69
WS1	0.86	0.80	0.76	0.81	0.82
WS2	0.84
WS3	0.81
SS1	0.70	0.77	0.70	0.77	0.78
SS3	0.69
OC1	0.86	0.84	0.86	0.86	0.77
OC3	0.83
OC5	0.80
OC8	0.76
JE1	0.71	0.69	0.78	0.74	0.73
JE2	0.76
JE3	0.70
JE5	0.68
TI1	0.70	0.68	0.77	0.75	0.69
TI3	0.69
TI5	0.66
PAEO1	0.70	0.66	0.74	0.74	0.67
PAEO3	0.67
PAEO4	0.65
PAEO5	0.64

Source(s): Authors’ contribution

Table 3

Discriminant validity

	SF	OE	PS	WS	SS	OC	JE	TI	PAEO
SE	0.663
OE	0.594	0.751
PS	0.625	0.767	0.727
WS	0.721	0.738	0.771	0.712
SS	0.632	0.52	0.662	0.532	0.550
OC	0.640	0.712	0.773	0.584	0.707	0.707
JE	0.708	0.537	0.714	0.547	0.611	0.678	0.627
TI	0.648	0.588	0.669	0.646	0.593	0.615	0.658	0.639
PAEO	0.561	0.727	0.632	0.668	0.632	0.636	0.536	0.661	0.681

Source(s): Primary data (PLS output)

Table 4

Path coefficient of the structural model and signification testing result

Hypo	Β	Ci-Min	Ci-Max	T	P	F2	Q2	Decision
H01	0.364	0.337	0.842	3.511	0.042	0.085		SIG
H02	0.320	0.065	0.362	4.652	0.001	0.062	0.541	SIG

Source(s): Authors’ contribution

Table 5

Predictive relevance

Model	R2	Adjusted R2	Q2 (=1- SSE/SSO)
Attrition	0.526	0.516	0.269

Source(s): Authors’ contribution

Note

1. Websites: https://www.livemint.com/companies/news/it-firms-may-see-a-sharp-decline-in-attrition-rate-11682270286789.html