Proportion Test

The Proportion Test computes a test statistic to determine if a specific category in a sample has a different proportion than the expected proportion, p0. The node calculates the sample proportion of the category p̂. With the given null-hypothesized proportion p0, a z-Score is calculated via (p̂ - p0)/se, where se is the standard error. If the corresponding p-Value is below the given significance level α, the null hypothesis can be rejected. Missing values will be ignored. See Wikipedia on test statistic, "One-proportion z-test" for further information.

Options

Category column
The distribution column, of StringType or BooleanType. Its total number of occurrences without missing values is nobs.
Category
The category to test for. The number of occurrences of this category in the category column is count.
Null hypothesis
The null hypothesized proportion 0 ≤ p0 ≤ 1.
Alternative hypothesis
The alternative hypothesis HA: whether the actual proportion =count/nobs is greater than, less than, or different from the null hypothesized proportion p0.
  • Greater than: Tests whether the actual proportion is larger than the null hypothesized proportion p0.
  • Two-sided: Tests whether the actual proportion is different from the null hypothesized proportion p0.
  • Less than: Tests whether the actual proportion is less than the null hypothesized proportion p0.
Significance level α
H0 is rejected if the p-value is below the significance level α.
Use sample proportion to compute standard error
Whether to calculate the standard error from the sample proportion p̂ rather than the default null hypothesis proportion p0.

Input Ports

Icon
The table from which to test samples.

Output Ports

Icon
Proportion test evaluation.

Views

This node has no views

Workflows

  • No workflows found

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.