Choosing the Correct Statistical Test - Brianna Hazel G. So STEM 12.pdf.pdf

CHOOSING THE CORRECT STATISTICAL TEST Brianna Hazel G. So STEM 12 Lesson Outline Overview 3 Essential Questions Types of Data Types of Samples Purpose of the Analysis Types of Statistical Tools  Overview Choosing the right statistical t...

CHOOSING THE CORRECT STATISTICAL TEST Brianna Hazel G. So STEM 12 Lesson Outline Overview 3 Essential Questions Types of Data Types of Samples Purpose of the Analysis Types of Statistical Tools  Overview Choosing the right statistical test can be challenging. This presentation summarizes key concepts and tests involving means, proportions, and relationships. Three important questions guide the choice of statistical tests. 3 Essential Questions What level of measurement is used for the data? How many samples do we have? What is the purpose of our analysis?  Types of Data Understanding Levels of Measurement Nominal Data Ordinal Data Interval Data Ratio Data Also called qualitative or With a meaningful order Numerical data with Numerical data with categorical but unknown intervals equal intervals but no equal intervals and a between ranks true zero point true zero point Examples: color, defective parts, and Examples: survey ratings Examples: temperature in Examples: weight, height, chocolate preference (poor, fair, good, and Celsius, and IQ scores and sales figures excellent) Summary Values: Summary Values: means, Summary Values: means, frequencies, proportions, Summary Values: and standard deviations and standard deviations and percentages medians, and ranks Independent T-test, Independent T-test, Chi-Squared test Kruskal-Wallis test Dependent T-test, One- Dependent T-test, One- Way ANOVA test, and Way ANOVA test, and Pearson Correlation Pearson Correlation Coefficient test Coefficient test  Types of Samples Number or Amount Two Two Multiple One Sample One Sample Independent Dependent Independent with Two Samples Samples Samples Variables Test a single Compare two Compare Compare more Analyze sample against a different groups measurements from than two groups relationships hypothesized the same group at within one group value Example: Compare two different times Example: Compare with two scores of drivers or conditions stress levels across measurements Example: Test if who watched vs. four different Example: Compare the average weight did not watch “Top conditions Example: Correlate times taken for two of packets meets a Gear” different types of daily sales with standard chocolate bars by One-Way ANOVA temperature for Independent T- the same test, and Kruskal- the same days One Sample T-test test, and Chi- participants Wallis test Squared test Pearson Pearson Correlation Correlation Coefficient test, and Coefficient test Dependent T-test  Purpose of Analysis Importance or Significance Testing Comparing Finding Comparing Against a Two Correlation Multiple Hypothesized Statistics Between Two Groups Value Variables To determine if a sample Compare two groups or Assess the association or Determine if there are statistic differs from a conditions correlation between differences among more known value variables than two groups Independent T-test (for One-Sample T-test (for comparing means of Pearson Correlation One-Way ANOVA test single groups) two independent Coefficient test (for (for comparing means groups) linear relationships) across multiple groups) Dependent T-test (for Chi-Squared test (for Kruskal-Wallis test (for compares means of two independence) comparing medians related groups or across multiple groups) conditions)  Independent T-Test Data: Interval or Ratio (Continuous) Sample: Two independent samples Purpose: Comparing the means between two groups  Dependent T-Test Data: Interval or Ratio (Continuous) Sample: One sample measured twice (paired data) Purpose: Comparing the means within the same group at two different times or conditions  One-Way ANOVA Test Data: Interval or Ratio (Continuous) Sample: Multiple independent samples Purpose: Comparing the means among three or more groups Pearson Correlation Coefficient Test Data: Interval or Ratio (continuous) Sample: One sample with two variables Purpose: Measuring the strength and direction of the linear relationship between two variables  Kruskal-Wallis Test Data: Ordinal or Non-Normal Interval or Ratio Sample: Multiple independent samples Purpose: Comparing the ranks among three or more groups  Chi-Squared Test Data: Nominal (categorical) Sample: Two or more independent samples Purpose: Assessing the association between categorical variables  References Nicola Petty. (2014, August 3). Statistical Tests: Choosing which statistical test to use [Video]. YouTube. https://youtu.be/rulIUAN0U3w Thank You! Brianna Hazel G. So STEM 12