Research methodology on relationship between employee social support and performance
of NGOS in Kenya
In this study, the research designs adopted were descriptive survey and correlational designs. The choices of the
designs allowed the study to determine the strength and direction of a relationship so that later studies could narrow
the findings down and, if possible, determine causation experimentally. Descriptive design is the method of research
which concerns itself with the present phenomena in terms of conditions, practices, beliefs, processes, relationships
or trends (Salaria, 2012). The study used survey data (based on responses in a questionnaire) to record the data.
Surveys are often used with a correlational design. However, anytime data is used to determine if two or more
factors are related or correlated, correlational design is normally used, even if a survey or questionnaire is used to
record the data. This study required both the causal relationship and the extent to which each independent variable
influenced the dependent variable. Therefore, both descriptive design and correlational design was deemed
appropriate for this study.In this study, the target population included all employees working in 7,004 registered
NGOs in Kenya.
However, in this study, the accessible population was a list of all employees working in all the 17 sub- counties
authorized to operate in Kenya by the NGO coordination Council of Kenya. According to the records available at
FKE (2018) there are a total of 14,283 employees in different categories of employment such as : those in top
management, employees in middle management and employees in lower management.The sampling frame for this
study was drawn from the 14,283 employees working in NGOs in Nairobi County in Kenya. The list of the various
categories employees of NGOs was obtained from the Federation of Kenya Employees (2018).Since the current
study constituted heterogeneous population, stratified random sampling technic was the most appropriate for the
study. Stratified random sampling was then employed to ensure representation of employees working in the different
levels in the organization. The levels were stratified as: junior management, middle management and top
management.
Table 3.1: Proposed sample size and sampling frame
Target population Population size Proposed sample size
Top management 38 33
Middle management 159 97
Junior management 14126 242
Total 14283 372
Source: FKE (2018).
In addition to that, considering the number of sub counties which was used in the study, the expression given below
was used to arrive at the number of employees selected for the study. This was done to ensure representation of
various staff in the organizations and for the researcher to be able to draw inferences from the various sub-categories
of the employees (Garg & Kothari, 2014).Data was collected using semi-structured questionnaires and interview
questions generated by the researcher. The questionnaires had both open-ended and closed-ended questions
(Schwab, 2005). The secondary data was obtained from historical documents, annual reports of NGOs and
commentaries from World Bank, UNAID, USAID, UNICEF and WHO.A pilot study was conducted randomly. The
main purpose of the pilot study was to ascertain the validity and the reliability of the questionnaire (Garg & Kothari,
2014). The study generated both quantitative and qualitative data. Quantitative data in the form of descriptive and
inferential statistics was analyzed with the aid of the Statistical Package for Social Sciences (SPSS) version 23.
Hypothesis testing involved running an OLS regression model for each independent variable against the
organization performance as done. Several assumption tests were done to verify whether the researcher would
proceed and fit a regression model to the study or not. Various assumptions of regression models were tested and
confirmed whether they existed or not. The study employed skewness and kurtosis test, Q-Q plot, Kolmogorov-
Smirnov test and Shappiro-Wilk to test for normality of the dependent variable. Multicollinearity was tested using
Variance Inflation Factor (VIF) which is reciprocal to tolerance as a rule of thumb. Breusch-Pagan test was used to
test the null hypothesis that is if heteroscedasticity exists. To find out if there was a linear relationship between
dependent and independent variable; correlation analysis between the dependent variable was conducted. In the
autocorrelation test , the study used Durbin Watson test. In this study, an outlier test was conducted using Q-Q plot
test. In this study, confirmatory factor analysis was used because it allowed the researcher to test the hypothesis that
a relationship beten the observed variable and their underlying latent construct exists. Corellation analysis was also
done. The sample size adequacy test was measured using Kaiser-Meyer-Olkin Measure (KMO). This study used
linear multiple regression models to measure retention of employees in universities in Kenya. Thus the linear
multiple regressions used was as shown below.
Y= β0 + β1X1+
Where: Y = Organizational performance= Constant term while the Coefficient = 1…4 was used to measure the
sensitivity of the dependent variable (y) to the unit change in the predictor variables X1
X1 = Employee Social Support
