Last week, the discussion was lengthy. I’d say, the topics were grouped according to their roles and uses in statistics.
The first part was about contingency tables and the chi-square test. A contingency table is a statistical tool used to display and analyze the relationship between two categorical variables. It organizes data into rows and columns, showing the frequency distribution of the variables' combinations. It is used heavily in survey research, business intelligence, engineering, and scientific research. These tables are great since they allow you to organize your data and provide answers to a variety of inquiries. With contingency tables, we can learn how to read, visualize, and evaluate data and to comprehend the connection between category variables. This is the tallying you do after you conduct a survey. Now, to interpret the data, for example if you want to know if the result means something or it happened purely by chance, then the Chi-Square test is the way to go. Chi-Square test is a statistical method applied to contingency tables to determine if there's a significant association between the variables or if any observed differences are due to chance. It compares observed and expected frequencies to assess the level of independence. A measure of variation, on the other hand, quantifies the dispersion or spread of data values. It provides insights into how much individual data points differ from the mean, giving an indication of data's overall variability and distribution.
Next part of the discussion was about data presentation. Here we discussed types of distribution particularly normal, kurtosis and asymmetrical distributions. Grasping these concepts aids researchers, analysts, and decision-makers in effectively describing data patterns, detecting potential outliers, selecting appropriate statistical methods, and making accurate inferences. Each concept serves as a valuable tool for ensuring data-driven insights are well-founded and representative of the underlying phenomenon being studied. Probability sampling was also discussed. Here, every member of the population has a possibility of getting chosen. Mostly quantitative research uses it. Probability sampling techniques are the best option if you wish to generate findings that are inclusive of the entire population.There were 4 types of this sampling namely simple random, systematic, stratified, and cluster.
The third part was about the statistics on understanding relationships between variables. And here, bivariate data and correlation are the fundamental tools. These help researchers, analysts, and decision-makers identify, measure, and make use of relationships between variables. They are crucial for data-driven decision-making, predictive modeling, research, and comprehending complicated interactions in the real world because of their wide range of applications.
Lastly discussed were, logical regression, hypothesis testing, and statistical errors. Logistic regression is a statistical technique that models a binary dependent variable and one or more independent variables to predict the likelihood of the dependent variable falling into a specific group or class. A real world example would be in medical research, predicting the likelihood of a patient having a specific medical condition based on various health factors. For hypothesis testing, it is an organized method that is applied in a variety of industries to assist people make defensible decisions. It uses sample data to assess the truthfulness of assertions or hypotheses about a population.It offers a framework for evaluating the reliability of hypotheses and enables researchers to reach insightful conclusions. Statistical errors might come consequent to hypothesis testing. These are inaccurate results that might happen while using statistical procedures to derive conclusions from data. Two main types of errors are Type I Error (False Positive) and Type II Error (False Negative). Type I Error occurs when a null hypothesis that is actually true is incorrectly rejected while Type II Error occurs when a null hypothesis that is actually false is not rejected.
Posted using Honouree