Statistical analysis consulting firms / company, as a third party to help researchers or students gaining a master’s degree or doctoral degree. It has multiple roles and tasks as well as restrictions of providing services. Why should be some restrictions? Obviously, there is task of consultant and task of student. main idea of the research is student task. The consultant can not replace its task. Otherwise,the main task of consultant is to assist technical tasks. The task as follows:

Data Analytics Consultant Give Consider Methods

Due to lake of statistical analysis knowledge, student need help of data analytics consulting. They can assist a researcher to provide appropriate methodology and research design, rightly. Whether quantitative or qualitative research methods are suitable the research. Prospective research design or retrospective. Is it necessary observational studies, surveys, or even experimental design? Furthermore, they can assist technical task to input the data into computer. In term of  research methodology, the consultant has comprehensive overview. He can give alternative view of some research methods and its advantages. Indeed, the researcher has to choose appropriate one.

Statistical Data Analysis Services

Researchs use consultant when they have to time to running the data. Commonly, they use statistical software such as: SPSS, EVIEW, Amos, PLS, Lisrel. Hence, they would need to technically assist in the work of running data. It should not be lost is that researchers stay understand the processes and procedures of data processing. The reason to choose a particular software and reason to use certain analytical methods. Technical matters related to the running of data can be done by the consultant.

Provide Consultation of Data Analysis

One of the most important in research is to interpret the data. Without interpretation, statistical results only numbers without meaning. Researchers will discuss with data analytics consulting to find any figures. They may explain the data meaning on the software output.

Actually, consultant expected able to help researchers writing better research. Better use appropriate research methodology. Researchers, actually can just use a random method of statistical analysis and statistical software like SPSS and the output would be come.

Secondly, statistical consulting firms / company can give time efficiency to the researchers. Researcher could save time one or two weeks or even months. To learn various methods of research is not an easy thing. Moreover, statistical analysis methods are many kinds. Researchers who do not have background in mathematics or statistics may be difficult to learn strange terminology.

Restrictions of Data Analytics Consulting Task

Eventually, there are some limitations in using the services of consultants.

• Researchers should not use the services of consultants to find main ideas of the research.
• Researchers should not ask for the help of consultants to do plagiarism.
• Researchers should not use consultants to replace the role of researchers.
• Finally, Researchers should not assume that the consultants’ work is theirs

Descriptive statistics definition and examples will be discussed in this article. The definition is a method of statistical analysis that only describe the condition of the data examples. Descriptive statistics do not draw conclusions larger than the sample. For example: Researcher conducts a research with 100 students as sample size. The conclusions drawn only for 100 students, no more. An example of a descriptive statistics in the health subject: A study measures the prevalence of TB in certain areas. It measures the effectiveness of a drugs in healing a disease. Necessarily, its conclusion only cover for that area. Other example of descriptive statistics in economic research: Average movement of stock prices on the stock exchange. The volatility of commodity price indexes on the futures exchange.

Commonly, Descriptive statistics describes a variable. Possibly, it also illustrate between two or more variables. For example, study on gender towards blood pressure in a hospital. The variable only blood pressure and gender as control variable. We may increase the number of variable such as age or life style. In political study, gender, age range, educational level to be important variable in choosing a candidate for leader in the election. Descriptive statistics does not merely involve only one variable. Two or more variables are possible. Essentially, the conclusions only cover the sample size.

Descriptive statistics is different with inferential statistics. In inferential statistics, the conclusion can be drawn beyond the sample size. For example, a research using 100 students as sample size, but the conclusion draws for whole college. Researcher want to examine the study habits of students in a college that has a total of 20 thousand students. Researchers certainly don’t need to interview 20 thousand students. It is enough to interview 100 to 300 students  to get the conclusion of all college. 

Descriptive VS Inferential Statistics

Inferential statistical examples in health and medicine subject:  we want to examine effectiveness of anti-hypertensive drugs A compared to drugs B. It takes 50 sample size for control and 50 sample for treatment. In inferential statistics, the conclusion was not only for 100 samples but can be greater. Another example in economics: the effect of fiscal policy toward economic growth in a country. The study uses time series data samples for 5 years. In inferential statistics, conclusions can be more than 5 years or more.

Key word of inferential research is testing the hypothesis. we temporarily test the allegations of researchers by using little data or samples of the real situation. hypothesis testing is what distinguishes descriptive statistics and inferential statistics. In descriptive statistics, there is no hypothesis testing.

The similarity of descriptive and inferential statistics is both methods are quantitative method. They use statistical measurements as conclusions. Mean, median, mode and standard deviation are the statistical measurements. Merely, In a descriptive statistics,  mean, median and mode and standard deviation only describes the sample size. whereas, in inferential statistics, there is a standard error which is the gap of mean median and mode in between sample and the population.

Furthermore, we discuss descriptive a statistics definition and me examples, we need to analysis the software. SPSS is a familiar tools widely used in various field of science. In the next article, it will be peeled out.

Definition of Data Analysis

Data analysis in quantitative research proposal is one part of the chapter that researchers need in the beginning of writing a research proposal. Whereas in the research, it is an activity after the data from all collected. Activities in data analysis are: grouping data based on variables and types of respondents, tabulating data based on variables from all respondents, presenting data for each variable studied, doing calculations to answer the problem formulation, and doing calculations to test the proposed hypothesis.

Quantitative Data Analysis Techniques

In a research proposal, it must be clear what method of analysis is capable of answering the research hypothesis. Hypothesis is a temporary answer to the research problem. Data analysis techniques in quantitative research commonly use statistics. There are two kinds of statistical data analysis in research. These are descriptive statistics and inferential statistics. Inferential statistics include parametric and non-parametric statistics.

Descriptive statistics

In preparing research proposals, researchers need to explain what is descriptive research. Descriptive statistic is a method to analyze data by describing data without intending to make generalizations. Descriptive statistics only describes the sample data and does not make conclusions that apply to the population. While, conclusion that applies to the population, then the data analysis technique is inferential statistics.
In addition descriptive statistics also function to present information in such a way that data generated from research can be utilized by others in need.

Inferential Statistics

When researchers want to generalize broader conclusions in the research proposal, it is necessary to write inferential statistics. Inferential statistics (often also commonly inductive statistics or probability statistics) are statistical techniques used to analyze sample data and the results are applied to populations. It requires a random sampling process.

Inferential research involves statistical probability. Using of probability theory is to approach sample to the population. A conclusion applying to the population has a chance of error and truth level. If the chance of error is 5%, then the truth level is 95%. While the chance of error is 1%, then the truth level is 99%. This opportunity for error and truth is the level of significance. Statistical tables are useful for carrying out tests of the significance of this error. For example, t-test will use table-t. in each table provides significance level of what percentage of the results. For example the correlation analysis found a correlation coefficient of 0.54 and for a significance of 5% it means that a variable relationship of 0.54 can apply to 95 out of 100 samples taken from a population.

Inferential statistics is a higher level then descriptive statistics. To that in the research proposal, the flow of conclusions becomes clear. Data Analysis is to make general conclusions (conclusions), make a prediction (prediction), or make an estimate (estimation).

Quantitative research method defines as a scientific approach that views a reality. It should be classified, concrete, observable and measurable the relationship of variables. Commonly, the data presents in the form of numbers and uses statistics method as a tool of analysis.

Quantitative research process begins with an activity of exploring the problems. Therefore, formulating research problems must be clearly. Even though, research problem in quantitative research is still temporary but it will develop after along the process. Relevant theories are to design research models and research parameters as well as the basis for making hypotheses. Theory will be the basis in the preparation of variables. In survey research, questionnaires/instruments are arranged based on research variables. The instrument need to test whether is valid and reliable. Data analysis which leads to examine the hypothesis.

Basically, quantitative analysis consists of several stages. Approximately, formulating the problem, constructing a model, obtaining data, finding solutions, testing the hypothesis, analyzing the data, and interpreting the results.

Method of analysis should refer to the purpose of study. Previously, researcher needs to identify whether the data meets the basic assumptions. While using questionnaire, data checking needs to support analysis. It need to clean the data, extreme values ​​(outliers), completeness the data.

Selection of Quantitative Methods

Format of quantitative research in social science depends on the problem and purpose of the research. There are two quantitative research formats: descriptive format and explanatory format.

The quantitative approach is deductive (from general to specific). It starts from a theory. This approach aims to test hypotheses and establish the facts or truths of a theory. The theories put forward as a standard to state whether or not a phenomenon occurs. Hypothesis is as a reinforcement of the assumption that quantitative research intends to see the relationship between variables. The final orientation is to make a conclusion that can be generalized more broadly. At the beginning, it has to be clear objections, and clearly design. For example, clarity of design, subject, variables, data, and analysis techniques. Hence, the quantitative approach is more stable. It does not allow re-designing the research procedure.

Quantitative research Characteristics

Certainly, Quantitative research has characteristics as follows:

Using deductive thinking patterns that try to understand a phenomenon by using general concepts to explain specific phenomena .
The logic of thought used is the logic of positivism and avoiding things that are subjective.
The aim of quantitative research is to compile nomothetic science, that is science that seeks to make laws from generalizations.
Data collection measures objective and use standard tools.
Contrarily, researchers position separately from the object of research. so the researcher does not emotionally involve with the research subject. Otherwise, qualitative research depends on the researcher.
In data analysis, researchers require to understand statistical techniques. It is able to analyze data after collected.
The quantitative research process follows a planned procedure involving numbers and data quantification.
The results of the study are generalizations and predictions. It is free of the context of time and situation.

A qualitative research methods is a process of research and understanding based on methodologies that investigate social phenomena and human problems. In this approach, the researcher makes a complex picture, examines words, detailed reports from the respondents’ views, and conducts studies in natural situations. Bogdan and Taylor (Moleong, 2007: 3) suggest that qualitative methodology is a research procedure that produces descriptive data in the form of written and oral words from people and observed behavior. Hence, this is the advantages of qualitative research methods. Researchers don’t need to use complicated statistical methods. it just use descriptive analysis.

Moreover, Qualitative research conducts natural conditions and discovery. In qualitative research, researchers are key instruments. Therefore, researchers must have the provision of theory and broad insights. So they can ask questions, analyze, and construct the object under study to be clearer. This research emphasizes the meaning and bound values. Qualitative research is powerful if problems are not yet clear. Even more, It finds out hidden meanings, understand social interactions, develop theories, ascertain data truths, and to examine developmental history.

Characteristics of Qualitative Research

Simply, Qualitative research is different from other studies. To find these differences, there are come characteristics of qualitative research:

1. In qualitative research, a researcher collects data in original or natural conditions.
2. Particularly, qualitative research concerns with the process than the results. It means in data collection often pay attention to the results and consequences of various variables influencing each other.
3. The background of behavior or deeds search for meaning. Thus, what is behind human behavior is central to qualitative research. Prioritizing direct data or “first hand”. Qualitative research demands as much as possible for researchers to conduct their own research activities.
4. Furthermore, the triangulation method is extensively both triangulation methods and triangulation of data sources.
5. Concern on contextual details. The researcher collects and records very detailed data about things in the study.
6. The subject under study is the same position as the researcher, so it is not an object or a lesser position.
7. Prioritizing the emic perspective, meaning that the respondent’s views are concerned, namely how he views and interprets the world and its standpoint.
8. Verification. The application of this method is through conflicting or negative cases.
9. Use the “Audit trail”. The method in question is to include the method of data collection and analysis.
10. Theory is basic. With data obtained from research in the field, formulate conclusions or theories.

Advantages of Qualitative Research

1. Purposive sampling. The qualitative method uses a small sample and is chosen according to the purpose of the study. Less data means less cost.
2. Since the beginning of the study. The data obtained are immediately analyzed until considered adequate results. it does not require a long process of data analysis.
3. The researcher as a research tool, means that the researcher as the main tool for collecting data is the data collection method based on observations and interviews. It has flexibility in data collection
4. In qualitative research, descriptive data collection write in the report. Data obtained from this research are in the form of words, images, and not numbers. it does not require complicated statistical methods.

Correlation vs Causation is a common term in scientific and health research. Among which several differences can occur. Identified Finding the true cause of a phenomenon is difficult to do as scientists tell you. Sometimes, the correlation and causation are closely related, but often they are not. Basically, tend to assume that if two events have correlation, they also have causation. But this does not happen in the majority of cases. This is a problem the difference between cause and correlation. It is a mistake to assume that just because two events have correlation, both tend to cause each other. This error or tendency refer to as causal non-causal pro-causa in Latin, or simply because of the cause.

Correlation

Correlation highlights that there is a relationship between two things; but does not predict causality. Indeed, correlation and causation is similar. One event definitely leads to another thing, it’s easy to build a causal relationship. If two events occur in a phenomenon, but one does not cause the other, they are only correlated and not causal.

It is easy to say that students who watch and play video games that are full of violence, blood and gore become aggressive in nature. But it is not certain because many still remain normal even after playing in many of these games. Here it would be more appropriate to say that violent games and violent behavior has correlation. But do not always have a cause and effect relationship. Hence, there is a causal relationship between watching violent video games and subsequent aggressive behavior, every child who watches and plays this game will become abusive. The parents have to ban this game.

Sometimes researchers do not just want to know a variable correlated with other variables or not. What he wants to know more is the cause of the occurrence or emergence of dependent variables, or what independent variables are causing the emergence of dependent variables. For example, what causes employees to stress (dependent variable). Is it causes sales volume to decrease (dependent variable)? what causes consumers to be dissatisfied (dependent variable). Is it true that inappropriate advertising ( independent variable) cause the decrease in sales volume (dependent variable) ? Is it true that one’s work motivation is low (dependent variable) due to low wages or salaries (independent variable)? He wants to know is what variables are causing all the negative things (symptoms) that arise or occur. With the aim of providing solutions to existing problems (job stress, decreased sales volume, unsatisfied consumers, etc.)

Causation

The causation highlights that there is a causal relationship between two variables. Simply, this highlights that A causes B. In scientific and health studies, the problem of confusion between correlation and causation often happen. In theory, it’s easy to distinguish, but the reality is not that simple. Many events can cause another event, such as lung cancer that is only occurs due to smoking. If an event causes another event, it has correlation among others. Two events that occur together do not mean the cause, for example smoking and alcoholism go hand in hand. But it is not one other cause.

Likewise in non-social sciences, causation research able to use the influence among variables. Someone who wants to know the effect of fertilizer application on plant growth will conduct various experiments in the laboratory and in the “field”. Plant growth is not only influenced by the presence of fertilizer, in order to truly know the effect of fertilizer, other variables probably have an effect on plant growth.

When a researcher sets the title “Effect of the amount of wages on the level of employee performance”, it explicitly describes what research methods should causation research. Why? Because the word “influence has meaning” effect “or” cause “. Researchers consciously or not, have put one status variable as an independent variable (the amount of wages), and one other variable as a dependent variable (performance level).

Structural equation modeling (SEM) is a statistical modeling technique that is very cross-sectional, linear and general. Included in this SEM is factor analysis (factor analysis), path analysis and regression (regression). Another definition mentions structural equation modeling (SEM) is a general and very useful multivariate analysis technique that includes a number of other analytical methods as special cases.

The next definition says that Structural equation modeling (SEM) is a statistical technique to build and to test statistical models that are usually in the form of causal models. SEM is actually a hybrid technique that includes the confirmatory aspects of factor analysis, path analysis and regression.

Advantages of SEM

Slightly different from previous definitions, Structural Equation Modeling (SEM) develops and has a function similar to multiple regression. Although it seems that SEM is becoming a more powerful analytical technique. It ables to consider interaction modeling, non-linearity, correlated independent variables, measurement errors, correlated error terms, multiple latent independent variables. It has many indicators, and one or two latent dependent variables that has also several indicators. Hence, SEM can be stronger than using multiple regression, path analysis, factor analysis, time series analysis, and covariance analysis

Its ability to decipher relationships between variables and to test theoretical credibility (or models). Which uses statistical techniques based on a number of very strict assumptions (Pedhazur, 1982). Three of the assumptions are that the variables the path analysis must be without errors, errors are uncorrelated, and the variables in the model are unidirectional (not including reciprocal models).

In many studies, almost all variables are unobservable variables. Variables such as educational aspirations, anxiety tests, student perceptions are latent behavioral concepts. The use of a single indicator to fully capture something complex in a path analysis is not practical. The nature of these variables in the path analysis requires the involvement of one or several indicators to construct each latent variable.

Path Analysis

Path analysis is an extension analysis technique of the regression model, which is used to test the dependence of a number of variables in a model. This model is generally depicted in a circle image and arrow direction that shows the relationship between the independent variable, mediator variable, and the dependent variable. The model is suitable for testing research hypotheses that show direct and indirect relationships. The causal model formulated by this researcher must be based on strong theoretical considerations.

This path analysis has advantages compared other analysis techniques. It is relatively simple and easy to do. In addition, we can analyze many variables with a relatively smaller number of samples.

The Weakness

As a comparison with SEM Full model, this path analysis only uses the total score of the variables only. While the full model on the SEM is a latent score. Since, it uses total score, that there is no measurement error in one variable. Therefore it is necessary to ensure that the measuring instrument has valid and reliable. Besides, the path analysis also can not see the accuracy index of the model. Therefore between one model with another model can not be compared.

Another weakness of the path analysis is that it does not allow the possibility of a link between residual error values ​​for each endogenous variable. Testing the model with the hypothesis of a joint effect (simultaneous) is rare. It is conceivable, academic experience will not only affect student performance. There is also the possibility of student performance affecting academic experience (eg learning, participating in study groups, accessing academic resources, engaging in class discussions). Thus, the use of path analysis to overcome problems like this is inappropriate.

In general, a researcher arranges hypotheses based on the formulation of problems and theoretical studies. For quantitative research, the hypothesis used is a statistical hypothesis, meaning that the hypothesis must be tested using statistical rules. Whereas for qualitative research does not need to use statistical rules. In a quantitative study, the formulated statistical hypothesis has two forms, the null hypothesis (Ho) and the alternative hypothesis (Ha). In general, hypotheses for quantitative research have three types: Descriptive Hypothesis, Comparative Hypothesis, and Associative Hypothesis.

Descriptive Hypothesis

Descriptive hypotheses are temporary conjectures about the value of a variable, not expressing relationships or comparisons. Remember, only about the value of a variable. Statistics used to test descriptive hypotheses are sample mean tests or standard deviation tests. A researcher formulates hypothesis based on the problem formulation and theoretical study. Following are some examples of problem formulations (PF), hypotheses (H).
PF: What is the percentage of junior high school mathematics mastery in the subject matter of the set?
H: Junior high school mathematics teacher mastery in the subject matter reaches 70%.

PF: How good is the grade XI mastery of class XI material?
H: mastery of class X material by class XI students reaches 75%.

Comparative Hypothesis

The comparative hypothesis is a temporary construct that compares the values ​​of two variables. That is, in the comparative hypothesis, we do not determine with certainty the value of the variables we examine, but compare. Means, there are two variables that are the same, but different samples. The statistics used to test this comparative hypothesis are (assuming normality is met) using a t-test. But before that, the normality and homogeneity must be tested first.
Following are some examples of problem formulations (PF), hypotheses (H).
PF: Is there a difference in the problem-solving abilities of students who got X learning better than students who got Y learning?
H: the problem solving ability of students who get learning X is better than students who get learning Y.

PF: Are there differences in the critical thinking skills of students who study during the day are better than students who study in the morning?
H: there is no difference in the critical thinking skills of students who study in the afternoon with students who study in the morning.

The two hypothetical examples above are slightly different. In the first hypothesis, we claim that the problem solving ability of students who get learning X is better than students who get learning Y. While in the second hypothesis, there is no one-sided claim that the critical thinking skills of students who learn during the day are better or worse. We only state that there are differences. Which problem is better, it does not concern this hypothesis. The first hypothesis is a one-party test hypothesis, while the second hypothesis is called a two-party test hypothesis.

Associative Quantitative Hypothesis

The associative hypothesis is a relationship between the relationship between two variables, the dependent variable and the independent variable. The statistics are used to test this comparative hypothesis are (assuming normality is met) using Product Moment Correlation, Double Correlation, or Partial Correlation.
The following are examples of problem formulations (PF), hypotheses (H).
PF: Is there a relationship between student achievement and the level of student anxiety?
H: there is a negative relationship between student achievement with the level of student anxiety.

PF: Is there a relationship between student learning outcomes and seating arrangements?
H: there is a positive relationship between student achievement with the level of student anxiety.

In the first hypothesis there are the words ‘negative relationship’. Negative relationship means inversely proportional. That is if the level of student anxiety is high, then student achievement is low. Whereas in the second hypothesis there are the words ‘positive relationship’. Positive relationship means directly proportional. It means if the seating arrangement is good, the student learning outcomes are high.

Ordinal scale definition is a scale showing sequential order and also serves to group. On an ordinal scale, the difference between values does not have any meaning, but only a sequence. Ordinal scale is impossible to make statistical calculations based on average calculations and standard deviations.

This Ordinal scale is higher than the nominal scale, it is a rating scale. In the ordinal scale, the symbol of the measured numbers is to show the difference and order or level of the object according to certain characteristics. For example, level of satisfaction on the product. We can give a number with 5 = very satisfied, 4 = satisfied, 3 = less satisfied, 2 = not satisfied and 1 = very dissatisfied. Or for example in a race, the winner gets rank 1,2,3 etc. On an ordinal scale, unlike the nominal scale, to replace the numbers, we need to sort the sequence from large to small or from small to large. So, we shouldn’t make 1 = very satisfied, 2 = not satisfied, 3 = satisfied etc. Which is 1 = very satisfied, 2 = satisfied, 3 = less satisfied and so on.

Ordinal Scale More Example

In addition, although ordinal scale has a clear boundary but does not yet have a distance (difference). We do not know the distance from satisfaction from dissatisfied to less satisfied. Eventhough, very satisfied with the number 5 and very dissatisfied we give the number 1. But, we cannot say that satisfaction that is very satisfied is five times higher than that which is very dissatisfied. As on a nominal scale, it also cannot apply standard mathematical operations (arithmetic). Such as subtraction, addition, multiplication, and so on. Statistical parameters that is in accordance with ordinal scale is similar on the number and proportion. Such as mode, frequency distribution, Chi Square and several other non-parametric statistical equipment.

Ordinal Scale is a scale that is the second level of measurement, tiering into something that is ‘more’ or ‘less’ than others. This measure is to sort objects from lowest to highest and vice versa. Example: measuring sports championships, work performance, employee seniority. For example: Answer questions in the ranking. For example: strongly disagree, disagree, neutral, agree and strongly agree get symbol numbers 1, 2,3,4 and 5. These numbers are only ranking symbols, do not express numbers.

The symbol indicates the order or level of the object measured according to the characteristics. For example, we want to know the preferences of respondents towards a noodle brand. brand A, brand B, brand C, brand D and then respondents rank noodle brands by giving number 1 for the most preferred brand, number 2 for the second rank, etc. A summary of the results as follows: Brand A = 1, brand B = 2, Brand C= 3, Brand D = 4

In statistics we know the scale of data measurement : nominal data, ordinal data, interval, and ratio. In general it can be said that the purpose of an observation is to obtain about the condition of an object in various circumstances. Among the various measurements for objects, which are numbers, rank, length, volume, time, weight, and physical-chemical measurements.

There are four measurement scales in Statistics:

Nominal Data

Firstly nominal Scale is the simplest measurement scale. It groups objects into several groups, which have similar features will be in one group. Nominal scale measurement results not to sort but can distinguish. Common examples commonly used are gender variables. In this case the measurement results cannot be sorted (women are higher than men, or vice versa). Examples of nominal scale applications: trademarks, types of shops, sales territories.

Ordinal Data

Secondly ordinal Data describes the position or rank but do not measure the distance between ranks. Size on an ordinal scale does not give an absolute value to an object, but only a relative sequence. Furthermore, the distance between rank 1 and 2 does not have to be the same as the distance between rank 2 and 3. On an ordinal scale, the rank does not have a unit of measurement. For example: social status (high, low, medium), measurement results that classify people into high, low or medium social status. In this case, we can know the level, but the difference between social statuses (high-low, low-medium, high-medium etc.) is not necessarily the same. Example application: preference level, management position, career path.

Interval Data

Thirdly interval scale gives the numeric characteristic to objects that have nominal and ordinal scales. It has the same distance in the order of the object. Interval scale is the level of this scale above the ordinal and nominal scale. Therefore, an important feature of this scale: we can add, subtract, duplicate, and share without affecting the relative distance of the scores.

Furthermore, this scale does not have an absolute zero. We cannot interpret in full the value of a certain ratio. In interval measurement switches, the ratio between two arbitrary intervals does not depend on the value of zero and the unit of measurement, For example, measurement of temperature on a Celsius scale. If a water bath is full of 0 degrees C, 50 degrees C, and 100 degrees C, the difference between 0-50 and 50-100 degrees C is the same. We cannot say that water at 100 degrees C is twice as hot as water 50 degrees C. Example application: Employee performance appraisal (on a scale from 0-100)

Ratio Data

Further ratio scale has all the properties of the interval scale plus one trait to give information about the absolute value of the object. It aims to distinguish, sort, certain distance, and we can compare (the most complete, including all of the scales above). Example: If we want to compare the weight of two people. A weight 40 kg and B 80 kg. We can know that A is twice as heavy as A. Because the value of the numerical variable weight expresses the ratio with zero as its default. Other examples: Age, value for money, height, etc.

The measurement scale above ranks from the low level (1 nominal scale) to the highest level (4 ratio scale). Measurement Scale with a higher level of measurement can change to a lower level, but the opposite.
Knowing this measurement scale will benefit a study. In the analysis, it needs measurement and what analysis tools are fit to answer the research objectives.

Comparative analysis aims to see the difference in the average of the dependent variable between two or more groups. The comparison test (dependent variable) for the two sample groups is the T test. The test for more than two groups of samples is the F test (ANOVA). Furthermore, comparative analysis requires normal distribution. Besides the variance must be homogeneous.

T test consists of one sample sample T test and two sample t test. One sample a test to examine a group of sample against a benchmark/standard value. for example, the researcher wants to test whether one group of students has a TOEFL test score above 500 or below 500. While in the two sample T tests, the researcher compares the population mean in the two sample groups. for example, researchers will compare the average TOEFL score in class A compared with class B. Does class A have an average TOEFL score that is the same as class B?

Furthermore, 2 sample t tests consist of paired samples and independent samples. We use paired samples when we compare two groups of samples which are basically one group. For instance, we compare the TOEFL score of class A before and after training.

It’s clear that the research subjects are the same but only at different times. In this paired sample t test, the number of samples must be the same. The subject will be tested before and after.

In independent samples, the subjects is different. For example we compare class A and class B at one time. The number of samples also allows different. Since the population of class A could be different with the population of class B.

When comparing more than 2 groups, we need ANOVA test. For instance, we examine whether group A, group B and group C have the same TOEFL score. In this test, it is possible to compare the same or different subjects. For example we compare the score of TOEFL in group A group B and group C. This test is only extension of the independent sample test. Likewise comparing group A before training 1, after training 1 and after training 2. The subject is only group A.

ANOVA test is technically different in calculations with T. ANOVA test, the basis of ANOVA is not average but variance. While the basic T test is average. Essentially the purpose of both calculations is the same.

Population is a generalization area consisting of: objects / subjects that have certain qualities and characteristics determined by researchers to get conclusions.

Finite Population

A finite population is a collection of objects or individuals that are objects of research that occupy a certain area. It clear boundaries that distinguish these population groups from other populations.

For example: the population of ducks in one cage, the number of A class students, the male population in an environment, and so on.

So the population is not only people, but also objects and other natural objects. Population is also not just the amount that exists on the object / subject studied. It includes all the characteristics / properties possessed by the subject or object.

For example, doing research in college Y, then college Y is the population. College Y has a number of people / subjects and other objects. This relates to the population in terms of quantity / quantity. But college Y also has the characteristics of its people. For example: its work motivation; work discipline; leadership; organizational climate and others. Furthermore: it also has other object characteristics: such as policies, work procedures, classroom layout, graduates produced and others.

One person can be as a population. Since, one person has various characteristics: style of speech, personal discipline, hobbies, ways to get along, his leadership and others. It will conduct research on President Y’s leadership then that leadership is a sample of all the characteristics possessed by President Y.

In the medicine, one person often acts as a population but not representing the unit of analysis. The blood is unit of analysis itself. If you want to observe the blood. Certainly, the population is person due to the blood inside the person.

Infinite Population

Infinite population is a collection of objects or individuals that are no boundaries or we can not measure about the total number of individuals in the occupied territories. For example, the population of stars in space, the number of red blood cells in a person’s body, and so on.

A large population will make it difficult for researchers to conduct a research. Requiring a lot of time and money to observe all populations. For the reason that a study of only needs some elements to represent the population. That is sampling.

Basically we can change infinite populations to finite population. We only need to impose limits on the population that we will observe. For instance, infinite population of fish in the ocean. By changing the fish population space, for example we put it in a pond or aquarium. It will become a finite/limited population. Even more, limited population is able to count, rather unlimited population has a number beyond our ability to count.

Regarding population, the smallest unit in a population is an element. Element is our observation focus. In these elements we will observe the characteristics of each element. Researchers call it a variable. As a result variable is source of information of the population.