Statistical tests Flashcards
(21 cards)
When to use Chi Squared test
- when values are discrete and when comparing expected values with observed values
When to use Correlation coefficient test
- when investigating relationship between 2 variables and the strength of a correlation
When to use T test
- when comparing 2 means
Null hypothesis for Chi squared test
- There is no significant difference between expected value and observed value
Null hypothesis for T test
- There is no significant difference between the 2 means
Null hypothesis for Correlation coefficient test
- There is no significant correlation between x & y
What to do when p≤0.05 for Chi squared test
- There is less than/equal to 5% chance that the probability is due to chance
- Reject null hypothesis
- There is a significant difference between expected value and observed value
What to do when p≤0.05 for T test
- There is less than/equal to 5% chance that the probability is due to chance
- Reject null hypothesis
- There is a significant difference between the 2 means
What to do when p≤0.05 for Correlation coefficient test
- There is less than/equal to 5% chance that the correlation is due to chance
- Reject null hypothesis
- There is a significant correlation between x & y
What to do when p>0.05 for Chi squared test
- There is more than 5% chance that the probability is due to chance
- Accept the null hypothesis
- There isn’t a significant difference between expected values and observed values
What to do when p>0.05 for T test
- There is more than 5% chance that probability is due to chance
- Accept null hypothesis
- There isn’t a significant difference between the 2 means
What to do when p>0.05 for Correlation coefficient test
- There is more than 5% chance that correlation is due to chance
- Accept null hypothesis
- There isn’t a significant correlation between x & y
Degree of freedom for Chi squared test
n - 1 -> n = number of categories
Degree of freedom for T test
(n+n) - 2 -> n = total of one set
Degree of freedom for correlation coefficient
n -> n = number of pairs
What to do when test results > critical value for Chi squared test
- Chi squared value is greater than critical value
- Probability that the difference is due to chance is less than 5%
- Reject null hypothesis
- There is a significant difference between expected values and observed values
What to do when test results > critical value for T test
- T test value is greater than critical value
- Probability that difference in means is due to chance is less than 5%
- Reject null hypothesis
- There is a significant difference between the 2 means
What to do when test results > critical value for correlation coefficient
- Correlation coefficient value is greater than critical value
- Probability that correlation is due to chance is less than 5%
- Reject null hypothesis
- There is a significant correlation between x & y
What to do when test results < critical value for Chi squared test
- Chi squared value is less than critical value
- Probability that the difference is more to chance is less than 5%
- Accept null hypothesis
- There isn’t a significant difference between expected values and observed values
What to do when test results < critical value for T test
- T test value is less than critical value
- Probability that difference in means is due to chance is more than 5%
- Accept null hypothesis
- There isn’t a significant difference between the 2 means
What to do when test results < critical value for correlation coefficient
- Correlation coefficient value is less than critical value
- Probability that correlation is due to chance is more than 5%
- Accept null hypothesis
- There isn’t a significant correlation between x & y
