**Chapter THREE**

**Methodology**

3.1**Introduction**

In this chapter we dealt with the methods and techniques that was employed in the assemblage and analysis of informations. The chapter states the population and sample of accounting alumnuss in Nigeria that was studied, describe the information, informations types and beginnings. The chapter besides explains relevant variables, variable appraisals, the analytical theoretical accounts and the methods of informations analysis that was employed in this survey.

3.2**The Population and Sample**

All accounting alumnuss from Nigeria constitute the population. Nigeria is divided into six ( 6 ) geopolitical zones, North-East, North-West, North-Central, South-East, South-West and South-South with 30 six provinces ( 36 ) and a federal capital district. This survey adopts the purposive random trying method to obtain the sample size of accounting alumnuss. In order to accomplish this, the survey focused on accounting alumnuss working in the South-West and South-South geopolitical zones. Accounting graduates working both in public and private sectors including those that are freelance constitute the sample for this survey. Because accounting alumnuss by and large behave and ground likewise as a consequence of the nature of their preparation, the findings from the survey can be used to generalize for others.

3.3**Research Design**

In the class of this survey, we adopted the study method. The pick of the study method is on the land that the study method, which could be cross-sectional on longitudinal, was used because our survey is to happen out the extent of the relationship between the dependant and independent variables. By the study method, we used the questionnaire to arouse all the relevant information from our respondents for the intent of deriving understanding and to measure the relationship of the variables that were studied.

3.4**Sampling Technique**

In this survey, we adopt the simple random trying ( strontium ) technique to take members of our sample size. By this method, each alumnus ( accounting ) working in the 12 ( 12 ) provinces and metropoliss has equal opportunity of being selected to be portion of the sample size. The research worker or his helper visited the respondents during their on the job hours and administered questionnaire after obtaining permission from them.

3.5**Beginning of Data**

We obtained our informations from primary beginning to consequence the analysis of this research work. The primary information was elicited through the usage of questionnaire which was administered to the respondents. The figure of questionnaire administered in each province is one 100 ( 100 ) doing a sum of one 1000, two hundred ( 1200 ) questionnaire in all. The respondents were selected at random during their on the job hours. This enabled the research worker or his helpers to run into the respondents personally, and therefore supply the best response to the questionnaire.

3.6**The Research Instrument**

The research instrument used in this survey is the questionnaire. This is shown as Appendix in this work. The questionnaire was given to the respondents to make full after they have been indiscriminately sampled by the research worker. In all, there are 30 five ( 35 ) inquiries all drawn to prove the five hypotheses stated in the first chapter of this work. The questionnaire for this survey was adapted from the work of Bundy and Norris ( 1992 ) .

3.7**Questionnaire Administration**

The research instrument, which is the questionnaire, was administered through the usage of research helpers. Six ( 6 ) B.Sc. grade holders in Accounting and Business Administration were employed on impermanent footing. Equally shortly as the footings of employment were concluded, they were trained. The preparation was to familiarize them with how to run into people and elicit responses from them. Second, the questionnaire was read and explained to them and they were given the chance to inquire inquiries on issues that they themselves did non look to understand in the questionnaire. To guarantee that the questionnaire is non manipulated by any of the research helpers, unscheduled visits was made by the research worker to the provinces and metropoliss where the questionnaires were being administered. At the terminal of the exercising, one 1000, one hundred and thirty five ( 1,135 ) pieces of questionnaire were retrieved out of which one 1000, one hundred and 14 ( 1,114 ) samples were decently filled and used for our analysis. The balance 20 one ( 21 ) questionnaire sheets that were non decently filled were separated and discarded by the research worker.

3.8**Method of Data Analysis**

The survey employed a combination of statistical and econometric tools in the information appraisal and analyses process. The multidimensional nature of the dependant variable suggest that appraisal may be bias without using scientific techniques of dimension decrease before carry oning subsequent analysis. Consequently, the survey employed explorative factor analysis and so the multivariate arrested development utilizing the Ordinary Least Squares techniques. Both methods are discussed below.

*Factor Analysis*

In the behavioral surveies, factor analysis is often used to bring out the latent construction ( dimensions ) of a set of variables to measure whether instruments measure substantial concepts ( Cortina, 1993 ) . It is good known that there exist important statistical redundancies in most primary informations generated utilizing questionnaires. The intent is to cut down the dimensionality of a information set by happening a new set of variables, smaller than the original set of variables retain most of the sample’s information and eliminates redundancies. The indispensable intent of factor analysis is to depict the covariable relationship among many variables in footings of a few underlying, but unobservable, random measures, called factors. The common factor theoretical account proposes that each ascertained response or step is influenced partly by underlying common factors and besides by alone factors, neither of which can be found.

By and large, in this survey, factor analysis was used to cut down the dimensionality of the dependant variable ( occupation penchant ) which has 30 five ( 35 ) points, into smaller figure of factors than the original entire figure of variables. By finding factor categorizations through the factor analysis with measured informations, and by utilizing these factors as variables alternatively of utilizing the ascertained responses, we can cut down the figure of variables to a set that are more describable and simpler. The new variables that will emerge called factors are uncorrelated, and are ordered by the fraction of the entire information each retains.

Therefore factor analysis ( exploratory ) explores empirical informations in order to detect characteristic characteristics and fascinating relationships without enforcing a definite theoretical account on the informations unlike the confirmatory factor analysis ( CFA ) . The focal point of factor analysis in this survey is to cut down the redundancy that could originate from among the variables/items used to capture employment penchant by utilizing a smaller figure of factors.

In factor analysis, we represent ascertained variables ( i‚¶_{1}, i‚¶_{2}… , i‚¶p ) as additive combinations of a little set of random variables {*degree Fahrenheit*_{1}*, degree Fahrenheit*_{2}*… , degree Fahrenheit** _{P}*( m & lt ; P ) } called factors. The factors are underlying concepts or latent variables that “generate” the i‚¶’s. If the original variables ( i‚¶

_{1}, i‚¶

_{2}… , i‚¶p ) are at least reasonably correlated, the basic dimensionality of the system is less than p. The end of factor analysis is to cut down the redundancy among the variables by utilizing a smaller figure of factors.

As a consequence of the “explosive” nature of the dependant variable ( occupation penchant ) originating from the figure of variables underpinning the concept, there is demand for a penurious summarization of the variables and this will be done utilizing factor analysis tonss generated after carry oning an explorative factor analysis on the primary informations generated. After that, the factors tonss will be regressed on the explanatory variables to bring forth the necessary beta estimations. This is of import as dimension decrease is one of the major undertakings for multivariate analysis, it is particularly critical for multivariate arrested developments ( Maitra & A ; Yan, 2008 ) .

The chief constituent pull outing method with an oblique rotary motion that consists of a direct Oblimin with the Kaiser standardization will be used as the method to carry on the explorative factor analyses. However, before carry oning the explorative factor analysis, we shall besides set about certain nosologies cheques on the information. First, we shall analyze the Kaiser-Meyer-Olkin step of trying adequateness trials ( KMO index ) and the Bartlett’s trial for sphericalness to determine if there exist important intercorrelations between points and hence to analyze if the correlativity matrix is an individuality matrix.

*Ordinary Least Squares Regression*

The survey makes usage of ordinary least squares regression analysis as the information analysis method. Gujarati ( 2003 ) suggests four critical premises that must be met before using the OLS arrested development. The premises are Normality, Multicollinearity, Heteroscedasticity and Autocorrelation. However, given that the information is non time-series, the autocorrelation premise does non use. For Normality, the survey will use the Kolmogorov-Simirnov trial ensuing from the non-parametric nature of the information. Multicollinearity is one of the most of import jobs confronting the usage of multiple arrested development analysis because of the chance of collinearity between independent variables are Variance Inflation Factor ( VIF ) for each independent variable. In proving for heteroscedasticity which checks for stability of the mistake footings, the Breusch-pagan-Godfrey trial was performed on the remainders as a safeguard. Where the presence of heteroscedasticity is found in the remainders, one appropriate method to handle heteroscedasticity is to accommodate Robust Standard Errors that addresses the issue of mistakes that are non independent and identically distributed.

*Model Specification*

The Factor Analysis theoretical account expresses each variable as a additive combination of implicit in common factors, {*degree Fahrenheit*_{1}*, degree Fahrenheit*_{2}*… , degree Fahrenheit** _{P}*} with an attach toing residuary term to account for that portion of the variable that is alone. For in any observation vector Y, the theoretical account is as follows:

i‚¶_{1}-i?_{1}= i?¬_{11}*degree Fahrenheit** _{1}*+ i?¬

_{12}

*degree Fahrenheit*

*+…… . + i?¬*

_{2}_{1m}

*degree Fahrenheit*

_{m}+ vitamin E

_{1}

i‚¶_{2}-i?_{2}= i?¬_{12}*degree Fahrenheit** _{1}*+ i?¬

_{22}

*degree Fahrenheit*

*+…… . + i?¬*

_{2}_{2m}

*degree Fahrenheit*

_{m}+ vitamin E

_{2}

i‚¶_{P}-i?_{P}= i?¬_{P}*degree Fahrenheit** _{1}*+ i?¬

_{p2}

*degree Fahrenheit*

*+…… . + i?¬*

_{2}_{1m}

*degree Fahrenheit*

_{m}+ vitamin E

_{P}

Where*m*is the figure of factors which should be well smaller than P, otherwise, we don’t accomplish a penurious description of the variables as maps of a few implicit in factors. i? represents average vectors associated with the variables. The coefficient i?¬ is the weights normally called the factor burden, so that i?¬_{ij}is the burden of the ith variable on the jth factor.*fi*represents the jth factor. with appropriate premises,_{ij}i?¬ indicates the importance of the factor (*fi*) to the variable ( i‚¶_{1}) and can be used in the reading of*degree Fahrenheit*_{I}. The vitamin E_{I}variable describes the residuary fluctuation particular to the ith variable. The factors (*fi*) are frequently called the common factors while the residuary variables ( vitamin E_{I}) are frequently called the specific factors.

Following the calculation of tonss for the figure of factors generated from the explorative factor analysis conducted on the dependant variable ( employment penchant ) , the factor tonss will so be regressed on the explanatory variables. The arrested development theoretical accounts are therefore specified therefore:

*Empr-f** _{1}*= i??

_{0}+ i??

_{1}Gender + i??

_{2}Indus + i??

_{3}Socvs + i??

_{4}Age + i??

_{5}Orgsize + vitamin E

_{1}

*Empr-f** _{2}*= i??

_{0}+ i??

_{2}Gender + i??

_{3}Indus + i??

_{4}Socvs + i??

_{5}Age + i??

_{6}Orgsize + vitamin E

_{2}

*Empr-f** _{3}*= i??

_{0}+ i??

_{7}Gender + i??

_{8}Indus + i??

_{9}Socvs + i??

_{10}Age + i??

_{11}Orgsize + vitamin E

_{3}

*Empr-f** _{N}*= i??

_{0}+ i??

_{N}Gender + i??

_{N}Indus + i??

_{N}Socvs + i??

_{N}Age + i??

_{N}Orgsize + vitamin E

_{N}

Where:*Empr-f** _{1}*= Employment penchant factor score 1

*Empr-f** _{2}*= Employment penchant factor score 2

*Empr-f** _{3}*= Employment penchant factor score 3

*Empr-f** _{N}*= Employment penchant factor mark

Gender = Student gender

Indus – Industry

Socvs = Societal values

Age = Student age

Orgsize = Organizational Size

vitamin E_{I}– vitamin E_{N}= mistake term

i??_{1}– i??_{N}= slope coefficients

Apriori mark: i??_{1}… i??_{N}= & gt ; 0