Issue 32: T-Test uses incorrect degrees of freedom for one sample tes…

…ts (#33)

* gemspec: Upgrade dev dependencies.

* t-test: Use correct degrees of freedom when doing a one sample test.

fixes #32
This change fixes an issue where we use the size as the degrees of
freedom when perfoming a T test for one sample with comparative mean.

* travis: Run tests on Ruby 2.7.

* workflows: Run build with Ruby 2.7.

* workflows: Run tests on workflows.

.github/workflows/ruby.yml

@@ -18,3 +18,18 @@ jobs:
         gem install bundler
         bundle install --jobs 2 --retry 1
         bundle exec rake
+  build_2_7:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v1
+    - name: Set up Ruby 2.7
+      uses: actions/setup-ruby@v1
+      with:
+        ruby-version: 2.7.x
+    - name: Build and test with Rake
+      run: |
+        gem install bundler
+        bundle install --jobs 2 --retry 1
+        bundle exec rake

.travis.yml

@@ -5,4 +5,5 @@ rvm:
   - 2.6.0
   - 2.6.3
   - 2.6.5
+  - 2.7
 before_install: gem update --system && gem install bundler

README.md

@@ -9,6 +9,7 @@ Unit test runs under the following ruby versions:
 * Ruby 2.6.0.
 * Ruby 2.6.3.
 * Ruby 2.6.5.
+* Ruby 2.7.
 
 We got the inspiration from the folks at [JStat](https://github.com/jstat/jstat) and some interesting lectures about [Keystroke dynamics](http://www.biometric-solutions.com/keystroke-dynamics.html).
 

lib/statistics/statistical_test/t_test.rb

@@ -21,7 +21,7 @@ def self.perform(alpha, tails, *args)
                     raise ZeroStdError, ZeroStdError::STD_ERROR_MSG if data_std == 0
 
                     comparison_mean = args[0]
-                    degrees_of_freedom = args[1].size
+                    degrees_of_freedom = args[1].size - 1
 
                     (data_mean - comparison_mean)/(data_std / Math.sqrt(args[1].size).to_f).to_f
                   else

ruby-statistics.gemspec

@@ -28,7 +28,7 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
 
   spec.add_development_dependency "rake", '~> 12.0', '>= 12.0.0'
-  spec.add_development_dependency "rspec", '~> 3.6', '>= 3.6.0'
+  spec.add_development_dependency "rspec", '>= 3.6.0'
   spec.add_development_dependency "grb", '~> 0.4.1', '>= 0.4.1'
-  spec.add_development_dependency 'byebug', '~> 9.1.0', '>= 9.1.0'
+  spec.add_development_dependency 'byebug', '>= 9.1.0'
 end

spec/statistics/statistical_test/t_test_spec.rb

@@ -23,12 +23,24 @@
     # with a standard deviation of 0.452, whereas marks range from 1 (worst) to 6 (excellent).
     # The grade point average (GPA) of all fifth grade pupils of the last five years is 4.7.
     # Is the GPA of the 22 pupils different from the populations’ GPA?
-    it 'performs a t-test with one sample' do
+    it 'performs a t-test with one sample for one tail' do
       student_grades = [5, 5.5, 4.5, 5, 5, 6, 5, 5, 4.5, 5, 5, 4.5, 4.5, 5.5, 4, 5, 5, 5.5, 4.5, 5.5, 5, 5.5]
       alpha = 0.05
 
       result = described_class.perform(alpha, :one_tail, 4.7, student_grades)
 
+      expect(result[:p_value].round(6)).to eq 0.003114 # R 3.5.1. calculates the p_value as 0.003114
+      expect(result[:null]).to be false
+      expect(result[:alternative]).to be true
+    end
+
+    it 'performs a t-test with one sample for two tails' do
+      student_grades = [5, 5.5, 4.5, 5, 5, 6, 5, 5, 4.5, 5, 5, 4.5, 4.5, 5.5, 4, 5, 5, 5.5, 4.5, 5.5, 5, 5.5]
+      alpha = 0.05
+
+      result = described_class.perform(alpha, :two_tail, 4.7, student_grades)
+
+      expect(result[:p_value].round(6)).to eq 0.006229 # R 3.5.1. calculates the p_value as 0.006229
       expect(result[:null]).to be false
       expect(result[:alternative]).to be true
     end