Data Source and Method of Data Collection
Data based on its nature can be grouped into three types, namely dichotomy, discrete, and continuum.
Dichotomous data
is data that is disaggregated with each other, for example ethnicity, religion, gender, education, and so forth.
Discrete data
is data whose data collection process is carried out by counting or counting. Like, the number of children, the population, the number of deaths and so on.
Continuum data
is data collection data obtained by measuring premises measuring devices that use a certain scale. Like for example, Temperature, weight, talent, intelligence, and others.
Data Sources and Data Samples
Data based on the source, can be grouped into two types, namely internal data and external data.
Internal data
data collected from the institution or organization where the research was conducted. For example, Banyumas Middle School 1 teachers conduct research on teacher performance and aspects that influence it. If the data is taken from the performance of Banyumas 1 Junior High School teachers, then the data includes internal data.
External data
is data obtained from other institutions or organizations where the research was conducted. For example if the researcher will examine the business development of a particular product, and he is looking for data outside the company concerned, for example in the Central Bureau of Statistics, then the data is external data.
Ways of Data Collection and Examples
Based on the way data is collected, it can be grouped into two, namely primary data and secondary data.
Primary data
is data obtained from the first source, or it can be said that the collection was carried out by the researcher directly, as was the result of the interview and the results of the questionnaire (questionnaire). Soeratno and Arsyad (2003: 76) state that primary data is data collected and managed by the organization that uses or publishes the data. Example primary data, the researcher will examine the work procedures of a particular application, then interviews can be done about it.
Secondary data
is data obtained from the second source. According to Purwanto (2007), secondary data is data collected by other people or institutions. While according to Soeratno and Arsyad (2003; 76), secondary data is data that is used or published by organizations that are not processors. Thus secondary data has two meanings. First, the data has been further processed, for example in the form of diagrams or tables. Second, data collected by other institutions or people, or data not collected by researchers themselves. for example population income data collected by BPS, data collected by survey institutions and others.
Scale of Measurement
Based on the measurement scale, the data can be grouped into four types, namely: nominal, ordinal, interval, and ratio data.
Nominal Data
It is data that can only be distinguished, cannot be sorted and compared with one another. Nominal or nomi which means name, shows a sign or label that is only to distinguish between one another. Examples of nominal data are gender, religion, type of work and so on. The numbers in the nominal variable are used to calculate, i.e. many men, the number who are present and others. Then the numbers are expressed as frequency. Nominal data is obtained from nominal variables.
Nominal data have characteristics, calculation results and no fractions are found, the numbers listed are only signs or labels do not have a sequence (ranking), do not have a standard size, and do not have absolute zero. Statistical analysis is appropriate for nominal data, such as: Binomium Test, One Sample Quadrate Chi Test, Two Sample Chi Square Test, Mc Change Mark Test. Nemar, Fisher opportunity test, Chohran Q test, and Kotingensi coefficient test. Whereas the statistical tests used are non-parametric statistics (riduwan.2009: 6-7).
Ordinal Data
Is data that has a sequence (order), but does not have the same distance difference between the sequence sequence. In other words, data has levels, so that respondents can be ranked according to the characteristics that exist in themselves. In ordinal data can state that the data is more, the same, or less than other data. Ordinal data can be distinguished and sorted but do not have the same distance in the order or differences that exist.
For example: Ranking of learning achievements, rank 1 is 1000, 2 is 980, and 3 is 960. Which is the distance between ranks 1 and 2, different from ranks 2 and 3. Ordinal data not only categorize variables that show qualitative differences between the various categories, but also sort categories according to a certain way. Ordinal data are obtained from ordinal variables.
Statistical analysis that is appropriate for ordinal data includes, one-sample Kolmogorov Smirnov test, Sign test, Wilcoxon Sign pair test, etc. The statistical analysis used is non-parametric statistics (Riduwan. 2009.7-8).
Interval data
It is data that has the same difference, sequence, and distance difference between the sequence sequences, but does not have absolute or absolute zero. The distance on the interval scale is set to follow a certain size that is easy to understand its meaning in order to arrange an interpretation. For example learning achievement test data given numbers 4,5,6,7,8,9 and so on. The distance between 5 and 6 is the same as 6 and 7, and so on. However, this number does not have a comparative meaning, where the number 4 that students get does not mean the intelligence is half of the students who get 8. This is because the numbers in the interval data do not have absolute properties so they cannot be compared.
Interval data is obtained from interval variables. Ordinal data collected by scoring rules that follow a certain scale can be assumed to be interval data even though basically ordinal, for example data obtained from questionnaires that use scoring rules with a certain scale.
Statistical analysis that matches the interval data includes: t test, two sample t test, one-way ANOVA test, two-way ANOVA test, regression test and others. The statistical test used is the parametric statistical test (Riduwan.2009: 9).
Ratio Data
It is data that has the same difference, sequence, distance difference between sequence sequences, and has an absolute point or absolute zero, so that it can be compared with one another. The value of zero (0) as an absolute zero, indicates that a phenomenon with all the elements or factors in it is completely absent. With the zero value, the object with zero value means that the object has nothing in the variable. The distance between two adjacent points has the same value, the use of this data states a definite comparison. So it is more widely used in the exact sciences environment than social science.
Ratio data is obtained from ratio variables. Ratio data is the data that has the highest level in scaling the measurement of variables, because it can indicate differences, levels, distances, and can be compared. For example the liquid temperature data, a liquid has a temperature of 20 degrees Celsius, half of a liquid whose temperature is 40 degrees Celsius. Ratio data has the most variation, i.e. differences, sequence, levels, similarity in distance, and comparison. Statistical analysis and tests that match the ratio data are the same as those used for interval data.
Data source
Data needed in research can be collected or obtained from various data sources. Understanding the source of data in research is the subject of data mama can be obtained. If the research uses interviews or questionnaires in the collection of data, then the source of the data is called respondents, that is, those who respond or answer the researcher's questions. If the data collection is done on the population, the research respondents are the population, whereas if the data collection is done on the sample, the respondent is the sample. Data is collected by giving a score of responses given by respondents. Questions regarding data will be collected relating to variables.
If the researcher uses an observation technique, the source of the data can be in the form of objects, movements or processes. Research that observes student activities in learning, the source of the data is students, while the object of research is student activities in learning activities. If the researcher uses document analysis, the document or notes that are the source of the data, while the contents of the research subject notes become the research variables.
Data sources can be grouped based on two things, namely based on the subject to which the data is attached, and based on the area of the data source. Based on the subject where the data is attached to the data source can be classified into 4 abbreviations P (4p) of English, namely:
p = person, data source in the form of person. Is a data source that can provide data in the form of oral answers / interviews or written answers through a questionnaire. The data source is called respondent.
p = place, data source in the form of place. Is a data source that presents the display in the form of a state of rest, such as tools, objects, colors, room conditions and so on.
p = process, source of activity data or activities. Is a source of data that presents a display in the form of a moving state, such as learning activities, performance, movement and so on.
p = paper, data source in the form of symbols. Is a data source that presents signs in the form of letters, numbers, symbols, and other images.
Data collection conducted on the population produces more accurate data and conclusions because no errors occur. This is because all data objects are collected, and analyzed. However, data collection like this is not uncommon because it has many obstacles. Under these conditions data collection is usually only done from the sample.