Data Analysis Methods RESEARCH METHODOLOGY

37 1 Library Study, done for collecting the data with information through reference books, journals, and other information which suitable according to this study. 2 Notes of Al-Azhar Cooperative such as history company, organization structure, total of employee, total member, and product.

D. Data Analysis Methods

1. Descriptive Analysis According to Istijanto 2009 descriptive study was conducted where the research is done to provide wider exposure which has a purpose to change the raw data become easy to comprehended in the form of short time. The descriptive analysis used to explain general description about the respondent. This way is done by distributing the 60 questioners to respondent and then processed with statistical methods using SPPS 17 and Microsoft Excel. 2. Validity Test According to Malhotra 2004:269 validity of a scale may be defined as the extent to which differences in observed scale scores reflect true differences among objects on the characteristic being measured, rather than systematic or random error. According to Ghozali 2005:45 the validity of this research is used to measure the legality of a questionnaire. Test validity used to measure the legal valid or un-valid of a questionnaire. In order to determine whether an items that are to be used or not, usually done on the significance 38 correlation coefficients test on minimum correlation 0.30, means that an item is considered valid if the total score is greater that 0.30 Priyatno, 2010:90 3. Reliability According to Malhotra 2004:267 reliability refers to extent which has a scale produces consistent results if repeated measurements are made on the characteristic. According to Sugiyono 2009:456 reliability is often defined as the consistency and stability of data findings. From a positivistic perspective, reliability typically is considered to be synonymous with consistency of data produced by observation made by different researcher at different times. Reliability test conducted by researcher is to measure the consistency of the questions that include in the questionnaire on the variables. According to Malhotra 2004:268 a questionnaire is considered reliable when cronbachs alpha test reached 0.6 or more. Cronbach’s alpha is the average of all possible split half coefficients resulting from different ways of splitting the scale items Malhotra, 2004:268. 4. Path Analysis According to Riduwan and Engkos 2008:2 path analysis model was used to analyze the patterns of relationships between variables with the aim to find out directly or indirectly influence the set of independent variables exogenous against the dependent variable endogenous. There are some 39 benefits using path analysis based on Riduwan and Engkos 2008:2 are as follows : a. Description of the phenomenon studied or researched the problem. b. The prediction variable Y based on the value of the variable X and predictions with path analysis is qualitative. c. Determination of the determinant factor variable X where the dominant influence of the related variables Y, can be used to search for the mechanism line the influence of the free variable X against variables bound Y. d. Use model theory test trimming good test reliability and test new concepts of development. Application of method of trimming is used to correct a structural path analysis model by way of removing the exogenous variables the coefficient of the coaster is not significant. Figure 3.1 Path Analysis ρX1Y2 ɛ 1 ɛ 2 ρX1Y1 RX1X2 ρY1Y2 ρX2Y1 ρX2Y2 Source: Primary data Processed 2014 Service Quality X1 Customer Loyalty Y2 Customer Satisfaction Y1 Brand Image X2 40 Y1 = ρx 1 y 1 X 1 + ρx 2 y 1 X 2 + ρy 1 ɛ 1 y 1 Y2 = ρx 1 y 2 X 1 + ρx 2 y 2 X 2 + ρy 1 y 2 Y 1 + ρɛ1ɛ2 Description : a X1 = Service Quality b X2 = Brand Image c Y1 = Customer Satisfaction d Y2 = Customer Loyalty e ɛ1 = Error 1 f ɛ2 = Error 2 5. The Coefficient of Correlation According to Sarwono 2007:22 correlation of coefficient is used to see how strong relationship and direction between one or more variable. Coefficient of correlation can be shown by the number of pearson correlation. Pearson correlation ranged from zero until one. If pearson correlation is close to the number one, it means the relationship is getting strong. Otherwise, if pearson correlation is close to zero then the relationship is getting weak. Table 3.2 The Level of Coefficient Correlation Sarwono,2007:108 Internal Coefficient Level of Relationship 0.0 - 0.25 Very Weak 0.25 – 0.5 Weak 0.5 – 0.75 Strong 0.75 -1 Very Strong 41 6. The Coefficient of Determination According to Ghozali2006:202 the coefficient of determination essentially measure how far the ability of models to explain variation in the dependent variable. The value determination of coefficient is between zero and one. The small means that the ability of independent variables in explaining variations in the dependent variable is very limited. Basic weaknesses use the coefficient of determination is based on the number of independent variables entered into the model. Each additional one independent variable, then R 2 would increase, no matter whether these variables affect the dependent variable is not. Therefore, in this research used is the R square that have been adapted or adjusted R 2 as adjusted for the variables used in this research. Adjusted R 2 value can rise or fail if an independent variable added into the model. 7. Simultaneous Test F test According to Sarwono 2007:29 F test is used to see the influence exogenous variable to endogenous variable simultaneously. According to Sarwono 2007:30 the criteria for testing the significant level is 5 or 0.05. Step to examine the hypothesis with F test are as follow Sarwono, 2007:17 : a. Calculate F test by SPSS b. Calculate the F table with the criteria significant level is 0.05 c. Determine the criteria of hypothesis test as follows : 1 If F test F table , H0 is rejected and H1 is accepted, it means 42 exogenous variable has significant influence to endogenous variable simultaneously 2 If F test F table , H0 is accepted and H1 is rejected, it means exogenous variable doesn’t have significant influence to endogenous variable simultaneously. 8. Partial testT test According to Sarwono 2007:31 T test is used to see how big influence the exogenous variable to endogenous variable partially. Step to examine the hypothesis with T test are as follow Sarwono, 2007:32 : a. Determine the hypotheses : H0 : Exogenous variable doesn’t have influence endogenous variable H1 : Exogenous variable has influence endogenous variable b. Calculate the number of T test by SPSS c. Calculate the number of T table with the criteria significant level 0.05 and degree of freedom df = n – k number of respondent – exogenous variable d. Determine the criteria of hypothesis test as follows : 1 If T test T table , H0 is rejected and H1 is accepted, it means exogenous variable has significant influence to endogenous variable partially. 2 If T test T table , H0 is accepted and H1 is rejected, it means exogenous variable doesn’t have significant influence to endogenous variable partially. 43

E. Operational Variable