Secondary Data Data Collection Technique

42 A measurement of each variable in this research using Likert scale to measure attitudes, opinion and perceptions of individuals or groups social phenomena Sugiyono, 2013:94. By using a Likert scale, the measurement variable is an indicator variable that will be outlined. Using a Likert scale of five 5 levels to express the attitude or the respondent‘s answer as follows: Table 3.1 Likert Scale Strongly agree always very positive very satisfactory 5 Agree often positive satisfying 4 Neutral hesitant sometimes 3 Disagree never negative unsatisfactory 2 Strongly disagree strongly never very dissatisfy 1 Source: Sugiyono, 2013:133

2. Secondary Data

A secondary source is a source that does not directly provide data to data collectors, for example through others or through documents. Secondary data were generally obtained by the founders to provide additional information and images for further processing. Secondary data used in this study was obtained from books, journals, literature or other writings that are considered related to the problems studied by using written reports or 43 documentation of previous studies and other information that can be retrieved through the system online internet. In the process of secondary data collection, the researchers collected data related to and associated with the research. So it can support materials to support this research. D. Methods of Data Analysis 1. Validity Test Validity test used to measure the validity of the data in the study. According Ghozali, 2011:52 validity test is used to measure the legitimacy of a questionnaire. Validity test used to measure the validity of a questionnaire. Testing was conducted using Pearson correlation, guidelines for a model is said to be valid if the significance level below 0,05 then the questions can be said to be valid. In addition, the criteria used in determining whether or not valid questions or statements used in this study is the 95 confidence level α = 5, the number of respondents as many as 30 respondents to pre-test, and compared with the value of r table = 0,361 in the can of degree of freedom df = n - 2, in this case n is the number of pre-test sample of 30 respondents. A questionnaire is said to be valid when the value of r count larger than r table. Test will test the validity of each variable used in this study. Here are the results to test the validity of the variable service quality, 44 sales promotion and customer satisfaction with 30 samples of respondents.

2. Reliability Test