The document discusses hypothesis testing and the steps involved. It defines key terms like the null hypothesis (Ho), alternative hypothesis (Ha), type I and type II errors, significance levels, one-tailed and two-tailed tests, and when to use z-tests and t-tests. It provides formulas and guidelines for traditional hypothesis testing, including setting the significance level, determining the test statistic, comparing it to critical values, and deciding whether to accept or reject the null hypothesis. Examples are given for paired t-tests, independent t-tests, and one sample t-tests.
The document discusses hypothesis testing and the steps involved. It defines key terms like the null hypothesis (Ho), alternative hypothesis (Ha), type I and type II errors, significance levels, one-tailed and two-tailed tests, and when to use z-tests and t-tests. It provides formulas and guidelines for traditional hypothesis testing, including setting the significance level, determining the test statistic, comparing it to critical values, and deciding whether to accept or reject the null hypothesis. Examples are given for paired t-tests, independent t-tests, and one sample t-tests.
The document discusses hypothesis testing and the steps involved. It defines key terms like the null hypothesis (Ho), alternative hypothesis (Ha), type I and type II errors, significance levels, one-tailed and two-tailed tests, and when to use z-tests and t-tests. It provides formulas and guidelines for traditional hypothesis testing, including setting the significance level, determining the test statistic, comparing it to critical values, and deciding whether to accept or reject the null hypothesis. Examples are given for paired t-tests, independent t-tests, and one sample t-tests.
The document discusses hypothesis testing and the steps involved. It defines key terms like the null hypothesis (Ho), alternative hypothesis (Ha), type I and type II errors, significance levels, one-tailed and two-tailed tests, and when to use z-tests and t-tests. It provides formulas and guidelines for traditional hypothesis testing, including setting the significance level, determining the test statistic, comparing it to critical values, and deciding whether to accept or reject the null hypothesis. Examples are given for paired t-tests, independent t-tests, and one sample t-tests.
- (or α – error) **A test of hypothesis is a procedure used to - when we reject the null hypothesis (action) substantiate or invalidate a claim which started as a null when in fact the null hypothesis or Ho is true hypothesis.** (actual condition) and therefore the alternative hypothesis or Ha is false. Hypothesis testing Type II error - A statistical method that is used in making - (or β – error) statistical decisions using experimental data. - when we accept the null hypothesis (action) - Basically an assumption that we make about when in fact the null hypothesis is false (actual the population parameter. condition) and therefore the alternative hypothesis or Ha is true. Methods used to test hypothesis: 1. Traditional method - the most widely used especially Significance level of a test in statistics books and instructional manuals. - The probability of making a type I or alpha 2. The p-value method - becomes popular since the error in a test advent of modern computer and high-powered - is the maximum value of the probability of statistical software’s. rejecting the null hypothesis Ho when in fact it 3. The confidence interval method - explained after the is true. p-value and some illustrative examples were discussed to show its relationship with hypothesis testing. Level of significance - For tests involving statistical decision, it is **These hypotheses are then subjected to testing. If customary to use a level of significance of 5% or they are found to be statistically true, they are 1%. accepted; if they are found to be false, they are 5% significance level rejected** - means that we can accept about 5 chances in 100 that we would reject the null hypothesis **The basis of progress in the physical and social when it should be accepted (i.e., we are willing sciences consist in large part in the rejection of false or to commit an α error or type I error of 5%). unworkable hypotheses and the acceptance of true - implies that we are 95% confident that we and valid ones.** have made the right decision. 1% significance level Null Hypothesis - often denoted by α = .01, means that we could - Ho be wrong with a probability of 0.01. - is a statement that there is no difference - implies that we are 99% confident that we between the parameter and a specific value, or have made the right decision. that there is no difference between two parameters. One-tailed test - shows that the null hypothesis be rejected Alternative Hypothesis when test value is in the critical region on one - Ha side of the mean. - is a statement that there is a difference - Right-tailed test if may more than, less than, taller than etc. between the parameter and a specific value, or - Left-tailed test that there is a difference between two Two - tailed test parameters. - the null hypothesis should be rejected when the test value is in either of the two critical region. if may not equal Steps in the traditional method of hypothesis testing What type of t-test should I use? Step 1. Formulate the null hypothesis (Ho) that there is When choosing a t-test, you will need to consider two no significant different between items being compared. things: whether the groups being compared come from State the alternative hypothesis (Ha) which is used in a single population or two different populations, and case Ho is rejected. whether you want to test the difference in a specific Step 2. Set the level of significance, α. direction Step 3. Determine the test to be used. Use z-test if population standard deviation is given from the Paired T-test samples. - If the groups come from a single population Step 4. Determine the tabular value for the test. - Ex. - For a z-test, use the table of the critical values - Measuring before (pre-test) and after (post- of z based on the area of the normal curve. test) an experimental treatment. - For a t-test, first compute for the degrees of Two sample t-test freedom; then look for the tabular value from - Independent t-test the table of t-distribution. - If the groups come from two different For a single sample, populations df = n – 1 - Ex. For two samples, - Two different species or people from two df = n1 + n2 – 2 separate cities n = number of items One sample t-test Step 5. Compute for z or t. - If there is one group being compared against a **Usually statisticians test hypothesis at α level standard value of 0.05 and 0.01 and occasionally at 0.10.** - Ex. **Remember that the level depends on the - Comparing the acidity of a liquid to a neutral pH importance of the Type I error.** of 7 Step 6. Compare the computed value with its corresponding tabular value, then state your conclusion Summary of Steps for Testing of Hypothesis: based on the following guidelines: 1. Identify Ho and Ha Reject Ho - if the absolute computed value is equal to 2. Identify if one or two-tailed or greater than the absolute tabular value. 3. Determine Tabular Answer (if no level of significance Accept Ho if the absolute computed value is less than given, use 0.05) the absolute tabular value. 4. Determine Computed answer (Z-Test of T-Test, just look at the given) T-test 5. Compare tab and com - is a statistical test that is used to compare the 6. Decision Rule (Accept or Reject Ho) means of two groups. It is often used in 7. Conclusion hypothesis testing to determine whether a process or treatment actually has an effect on The difference in the grade point average of the BSA the population of interest, or whether two students between the first semester (mean = 90.3; SD = 5.06) and second semester (mean = 88.9, SD = 4.63) was groups are different from one another. not significant (t(18) = 0.65; p=0.26) When to use a t-test? - A t-test can only be used when comparing the means of two groups (a.k.a. pairwise comparison). - If you want to compare more than two groups, or if you want to do multiple pairwise comparisons, use an ANOVA test or a post-hoc test.