Statistical framework. Don’t be confused by the word statistics. According to the large Indonesian dictionary, statistics is a noun (n), which means (1) a record of numbers; estimation; and (2) data in the form of numbers that are collected, tabulated, and classified so that they can provide meaningful information about a problem or symptom. You will understand and be able to do what is described in statistics if you understand what statistics are.
Statistics also have groupings; understand these statistical groupings, and it will make it easier for you to determine what you will do and where you are when you are dealing with data. Just like a map, you don’t have to learn all parts of statistics; you just need to identify the data you have and focus on that part by learning the right statistical methods to use.
Statistics look complicated if you don’t try to look at them in general terms. Let’s fly up to see what Java looks like, so that if you are in West Java, you will realize that the sea is closer to the west than walking to the east.
The general statistical form is illustrated in the picture below. Statistics itself is generally divided into two categories: descriptive statistics and inferential statistics.
Descriptive statistics is a statistical method used to describe a group of data with no intention to apply it generally to the data. For example, data on visitors to a mall You can use the data to describe when the mall has the most visitors or what percentage of visitors are children, women, middle-aged, teenagers, men, and so on. In the mall data, of course, you cannot say that the data represents all malls in Indonesia or all stores or supermarkets. You will focus on the data you have and explain it.
Usually, these descriptive statistics are used by researchers in the results and discussion section to explain the results of inferential statistical research. For example, the regression found that land area affects rice production. You certainly don’t stop there, saying that it turns out that land area greatly affects production. You can explain the results of the inferential analysis with descriptive analysis, namely an explanation of how much land area there is currently, the development of land area from year to year, and if it turns out to be reduced, what is the cause? Well, this is where this descriptive statistic comes in.
The tools used in descriptive statistics are usually tables, diagrams, and graphs that contain basic statistics, such as averages, frequencies, and others.
Inferential statistics is a statistical method that analyzes sample data to describe a population. The data we process will be used to describe population data. This is where the complexity comes from. You must determine the exact sample and sample method used so that the data you will have is truly representative of the existing population. Therefore, you learn about sampling techniques.
Inferential statistical methods also vary depending on the scale of the data you have. Statistical data can be divided into nominal, ordinal, interval, and ratio. Most people have the misconception that ratio data is the best data compared to other data. If you agree with this opinion, you are indirectly saying that quantitative research will be better than qualitative research (ratio data is usually used for quantitative research), even though qualitative research is more widely used for management and leads to policy changes. Both quantitative and qualitative aspects are equally important.
The correct argument is that ratio data has more opportunities to be analyzed because it has more diverse analysis methods than other data. If you have ratio data, it is likely that analytical tools such as goodness of fit, model interpretation, hypothesis testing, and data evaluation can all be used. Rather than having nominal data,
Nominal-type data distinguishes data in qualitative groups. The identity given to the group is only a distinction between groups and has no relationship or level. For example, 1 is male, and 2 is female. We can’t say 2 is higher than 1. The numbers 1 and 2 are just the identities of men or women.
If you have nominal data, you can use it to test the hypothesis of one sample, compare two or more samples, or establish an association or relationship between samples. For example, in one sample, it can be said that the number of men is significantly greater than the number of women. Or, for example, comparing samples in the city of Malang with samples in the city of Makassar to determine whether they are similar or significantly different.
Ordinal data is almost the same as nominal data, except that ordinal data has levels in each group. For example, in education, SD (1), SLTP (2), SMU (3), and PT (4) PT has a higher level than SLTP, SMU, and SD. But still, we cannot interpret that 4 or PT is twice as much as SLTP (2). Still, the number is the identity of the group that has a level from low to high.
Other widely used examples are strongly disagree (1), disagree (2), doubt (3), agree (4), and strongly agree (5).
Ordinal data and nominal data are usually grouped under nonparametric statistics. So, if you hear the term nonparametric statistics, it means that the data used in the study is ordinal or nominal data. Some of the analysis tools used are run test, Wilcoxon, Mann Whitney, Kolgomorov Smirnov, Friedman, two-way anova, Kruskal Wallis, and spearman rank. Of course, I will not explain each of these analysis tools in this article. But usually nonparametric statistics discuss one sample hypothesis (for example, whether this data group is significantly larger with a certain score), comparing groups (whether this group agrees more than other groups), or group relationships (what is the relationship between education level and certain opinions). Of course, the data used is on an ordinal or nominal scale.
Interval and Ratio Data
Interval and ratio data fall into the realm of quantitative data. The difference between interval and ratio data is that 0 is an absolute number in ratios.
To understand it, I give the following example: for example, data on the value of English lessons that are given a value of 0 to 100. The number 0 in the data is a symbol of the lowest value, not an absolute number. It can also be interpreted that a student who has a score of 60 does not have twice the intelligence of a student who has a score of 30. This means that this value data is interval data.
Ratio data, for example, balance data in a bank The value of 0 in the balance is the actual value. And someone who has a balance of $100 million can be interpreted as having twice the balance compared to someone who only has a balance of $50 million. Balance data is ratio data.
This interval and ratio data is grouped into parametric statistics. There are so many methods that can be chosen to process parametric statistical data because this is quantitative data. Starting from hypothesizing one sample group, comparing two or more groups, correlating two groups, to regression
The question is, What area of statistics does the time series fall into? Of course, it falls into parametric statistics. So, if the time series is about the number of women, how? Data on the number of women is also ratio data because the word number means an actual number, not just a symbolic number. Is that clear?
The conclusion of this article is that we must know where we are to be able to use the right statistical method according to the data we have.
Thank you for visiting.