By: Lannah Saldana
The class discussions last August 22 and 24 revolved around two major topics: descriptive and inferential statistics. Descriptive statistics show the characteristics and behavior of the dataset, including the standard deviation, variance, and skewness (Lee, 2020). On the contrary, inferential statistics tests the hypothesis using a population subset. This type of statistics aims to conclude for the larger group by analyzing sample data (Purdue OWL® - Purdue University, n.d)
The discourse on descriptive statistics starts with distribution, demonstrating the data frequency. Distribution measures include mean, the center in terms of value, median, the center in terms of order, and mode, which is the value with the highest frequency. Distribution also has types, namely symmetrical and asymmetrical. Symmetrical or normal distribution follows a bell curve since values are balanced on both sides of the peak. Asymmetrical or skewed distribution happens when one end or tail is longer than the other, signifying the presence of outliers. Mean is usually used to measure central tendency when the distribution is normal. At the same time, the median is ideal to be applied when the distribution is asymmetrical since outliers do not affect it. Another topic is measures of variation, which demonstrate how spread out the data is for a variable. These include range, the difference between the smallest and largest value; variance, which measures how far each point is from the mean; and standard deviation, which is the average distance of data points from the mean. Moreover, a common practice during exploratory data analysis is visualizing variation and distribution to understand the dataset better. Graphs such as histograms, box plots, and Q-Q plots are the most popular visualizations to present data distribution and variation.
Inferential statistics lies at the heart of the data analysis project. In this phase, statistical tests are performed to test the hypothesis stated at the project's start. Samples are drawn from the population to perform hypothesis testing. Samples may come from strata or subsets such as age or gender or from clusters such as geographical regions, schools, and households. Strata sampling is called stratified random sampling, which is used when the study needs a representative sample for each subgroup. On the other hand, sampling using clusters is called cluster random sampling, which is used when the study wants to investigate heterogeneity among groups. Moreover, statistical tests that are applied include linear regression, logistic regression, correlation, and chi-square test. Regression analysis is used for prediction problems since it employs the independent variables (X) to determine the value of the target variable (Y). Correlation measures the strength of the association of two quantitative variables, and the chi-square test analyzes categorical variables and investigates if there is a significant difference between observed and expected results.
In conclusion, descriptive statistics is conducted to understand the characteristics and behavior of the dataset. Moreover, inferential statistics is employed to identify relationships, distinguish patterns, and forecast values to test the propositions of the researchers. Interpreting the results in descriptive statistics is crucial since it determines the appropriate statistical tool for inferential statistics. Interpretation in the inferential statistics phase, too, is essential, for it must be ensured that the project gives valid and meaningful conclusions.
Source:
Basic inferential statistics: Theory and application. Basic Inferential Statistics - Purdue OWL® - Purdue University. (n.d.). https://owl.purdue.edu/owl/research_and_citation/using_research/writing_with_statistics/basic_inferential_statistics.html#:~:text=The%20goal%20of%20inferential%20statistics,generalize%20to%20the%20larger%20group.
Lee, J. (2020). Statistics, Descriptive. International Encyclopedia of Human Geography, 13–20. doi:10.1016/b978-0-08-102295-5.10428-7
Posted using Honouree