Hypothesis Testing Comparisons Among Two Samples Hypothesis Testing
- Slides: 19
Hypothesis Testing Comparisons Among Two Samples
Hypothesis Testing w/Two Samples n By far a more common statistic to use than any other covered so far n n If we have access to an entire population to calculate its mean (μ), why do we need to get a sample to infer its characteristics? We can just measure them directly 99% of the time we don’t have this kind of access
Hypothesis Testing w/Two Samples n Also, most experiment employ two groups – a treatment group and a control group n n n The Treatment Group – gets the IV The Control Group – identical to the Tx Group, minus the IV Since Control = Tx + IV, if our Tx Group has a different mean than our Control Group, we can attribute this to the IV
Hypothesis Testing w/Two Samples n Ex. If we want to know what effect fire has on water, we have two groups, one that has water over an open flame (the Tx Group), and one that has water unheated (the Control Group). If all other factors that could influence the heat of water are kept constant between the two groups (i. e. the water in the two groups has the same salinity, the air pressure is the same, etc. ), if the water in the heated condition is hotter, we can conclude that the difference is due to the fire, and not the other factors that we kept constant
Hypothesis Testing w/Two Samples n Assumptions of the Two-Samples TTest: n 1. Normally-distributed data n Like all other t-tests, and according to the Central Limit Theorem, as long as sample sizes are large, you can ignore this requirement
Hypothesis Testing w/Two Samples n 2. Homogenous Variances (s 2) n n n As long as your samples have similar sample sizes (n) and your data is not significantly nonnormal, you’re OK If, however: you have unequal sample sizes or non-normal data then you have to use a slightly different procedure for calculating t We’ll go into more detail about how to quantitatively detect, and cope with, nonhomogenous variances in the next SPSS lab
Hypothesis Testing w/Two Samples n 3. Independent Samples n As stated previously, your groups have to be independent – i. e. a subject can only be in one group and the group cannot be yolked n Yolked Groups = you assign subjects to groups that look similar on a variable or variables that you’re interested in n Ex. You’re interested in sociability ratings, so since subject 1 in Group A has a sociability rating of 45, the next subject that has a similar rating you assign to Group B n Here, membership in Group B is dependent on the subjects in Group A, i. e. the groups are not independent
Hypothesis Testing w/Two Samples n How to compute a Two-Samples TTest? n n n Instead of subtracting μ from subtract from +t = < -t = > , you
Hypothesis Testing w/Two Samples n Also, instead of using s 2, we use n n n for grp 1 variance for grp 1 n for grp. 2 variance for grp 2 n This is what is called our Pooled Variance
Hypothesis Testing w/Two Samples n This formula averages the variances from our two samples, however… n First it multiplies the variance by the sample size [s 2 x (n-1)], which gives more importance to variances from larger samples
Hypothesis Testing w/Two Samples n This is an example of what is called a weighted average n Weighted Average = average where our values to be average are multiplied by a factor that we think is important (in this case, n is this factor)
Hypothesis Testing w/Two Samples n Therefore, our formula for our Two. Samples T-Test is:
Hypothesis Testing w/Two Samples n n Since we’re using two samples, our df = Also, the form of our hypothesis changes: n For a One-Tailed Test: n n n H 0 = (μ 1 μ 2) (or visa-versa) H 1 = (μ 1 < μ 2) (or visa-versa) For a Two-Tailed Test: n n H 0 = (μ 1 = μ 2) H 1 = (μ 1 ≠ μ 2)
Hypothesis Testing w/Two Samples n We would apply the Two-Samples TTest in the same way as previous tests: n n n 1. Identify H 0 and H 1 2. Calculate df and identify the critical t 3. Determine whether to use one- or twotailed test, determine what value of α to use (usually. 05), and identify the rejection region(s) that the critical t is the boundary of
Hypothesis Testing w/Two Samples n n 4. Calculate the variances for both samples (s 1 and s 2), and use them to calculate the pooled variance 5. Calculate the mean of both samples 6. Utilizing this information, calculate t 7. Compare your value of t to your critical value and rejection region to determine whether or not to reject H 0
Hypothesis Testing w/Two Samples n Confidence Intervals: n The formula is the same as the one-sample t-test, once again all that is different is that we use two sample means instead of a sample mean and a population mean, and a pooled variance instead of a regular, sample variance n CI = ± Critical t (two-tailed at p=. 05) x pooled standard deviation
Hypothesis Testing w/Two Samples n Example: n Much has been made of the concept of experimenter bias, which refers to the fact that for even the most conscientious experimenters there seems to be a tendency for the data to come out in the desired direction. Suppose we use students as experimenters. All the experimenters are told that subjects will be given caffeine before the experiment, but half the experimenters are told that we expect caffeine to lead to good performance, and half are told that we expect it to lead to poor performance. The dependent variable is the number of simple arithmetic problems the subject can solve in 2 minutes. The obtained data are as follows:
Hypothesis Testing w/Two Samples Expect Good Performance 19 15 22 13 18 15 20 25 Expect Poor Performance 14 18 17 12 21 21 24 14 n n n What are the Ho and H 1? What is the df and critical t? What are your alpha, type of test (one- vs. two-tailed), and rejection region(s)? What is your t? Will you reject or fail to reject the null hypothesis? 22
Hypothesis Testing w/Two Samples n n n Assuming a one-tailed test, H 1 = (μ 1 > μ 2); H 0 = (μ 1 ≤ μ 2), where μ 1 = “Expect Good Performance” df = 15; critical t(15) = 1. 753 t =. 587 Fail to reject the null hypothesis Our data do not support theory that experimenter bias influences data.
- A dietitian wishes to see if a person's cholesterol
- Figure of speech comparison
- Inferences based on two samples
- Rumus uji median
- Null and research hypotheses
- Developing null and alternative hypothesis
- Null and alternative hypothesis statistics
- Protoplanet hypothesis weakness
- Comparisons of equality spanish examples
- Bad的superlative
- According to the pictures
- Comparisons ejemplos
- Comparative equality ejemplos
- Comparisons of equality ejemplos
- The comparative of equality
- Cmpt 225
- A priori and post hoc comparisons
- Tan tanto tanta tantos tantas
- Which modifier could be used to make a comparison?
- Modifiers of comparisons