May 27, 2021 SAS
The hypothesis test uses statistics to determine the probability that a given hypothesis is true. T he usual process of hypothesis testing consists of four steps as shown below.
Develop a zero hypothesis H0 (usually, observation is a purely accidental result) and an alternative hypothesis H1 (usually, observation shows a component combination of real effects and changes in opportunity).
Identify test statistics that can be used to assess the authenticity of the zero hypothesis.
The P-value is calculated on the assumption that the test statistic is at least as valid as the observed statistic when the invalid assumption is true. T he smaller the P-value, the stronger the evidence relative to the zero assumption.
Compare the p-value with an acceptable α value, sometimes α values. I f the observed α is statistically significant, the zero hypothesis is excluded and the alternative hypothesis is valid.
The SAS programming language has the characteristics of performing various types of what-if tests, as shown below.
Test | Describe | SAS PROC |
---|---|---|
T-Test | Use the t-test to test whether the average value of a variable is significantly different from the assumed value. We also determine whether the averages of two separate groups are significantly different, and whether the averages for dependent or paired groups are significantly different. | PROC TTEST |
ANOVA | It is also used to compare the means when there is a separate classification variable. When testing, we want to use a one-factor variance analysis to see if the interval-by-variable mean varies depending on the independent classification variable. | PROC ANOVA |
Chi-Square | We use the card-side fit degree to assess whether the frequency of classification variables may occur by accident. If the proportion of the classification variable is a hypothetical value, you need to use a card square test. | PROC FREQ |
Linear Regression
|
Use simple linear regression when you want to test a variable to predict the validity of another variable. P olylinear regression allows you to test how multiple variables predict variables of interest. When using multiple linear regressions, we also assume that the predictors are independent. | PROC REG |