Dependent-samples t test
Formate of reporting t tests
(text set up here), t(df) = t value, p = significance value Can also report: mean of each group Cohen's d
Assumptions of the dependent-samples t test
1) Assumption of independence 2) Assumption of normality 3) Missing: Assumption of homogeneity of variance
Assumptions of one sample t test
1.) Assumption of independence 2.) Assumption of normality 3.) Missing: Assumption of homogeneity of variance
Null hypothesis significance testing (NHST) One sample t test (part 2)
2. Collect data Error term is now simply the standard error of the mean.
Formate of reporting independent sample t test
An independent-samples t test indicated that males (M = 24.00, SD = 3.21) reported significantly stronger JWB than did females (M = 20.00, SD = 3.16), t(16) = 2.65, p = .02, d = 1.26.
Reporting significant results (for t test results)
Can report exact value: p = .007 Can report a criterion: p < .05, p < .001
Reporting nonsignificant results (for t test results)
Can report exact value: p = .48 Can report a criterion: p > .05, "all ps > .10" Can use abbreviated "not significant": t(16) = 0.89, n.s.
Dependent-samples t test
Compare two means The two samples are related in some way: -One sample with repeated measures -Two samples that are matched in some way.
one sample t test similarities
DV must be continuous Still have effect over error
What are the similarities between independent and dependent sample t test?
IV must be categorical with two categories DV must be continuous Still have effect over error
one sample t test differences
No IV - no groups, no group means No need for pooled variance
What are the differences between independent and dependent sample t test?
No longer comparing group means. We are looking at difference scores now. No longer need pooled variance to get standard error of the difference.
Null hypothesis significance testing (NHST) One sample t test
Null: H0: M = u Alternative: H1: M does not equal u It is acceptable to perform one tailed and two tailed test when comparing a sample mean to a population mean.
4. Null hypothesis significance testing (NHST) One sample t test
Our obtained t 3.26 was higher than the critical t 2.09
Example: Comparing self-esteem scores form pre- to post-tornado
SE1: M = 23.1 SE2: M = 21.6 NHST steps 1) Assume there is not effect - null hypothesis 2) Collect data / calculate a test statistic 3) Determine probability of the result if the null is true 4) Decide to reject or retain null Our obtained t (1.65) was not higher than the critical t (2.09) Retain the null hypothesis It does not appear that self-esteem changed from before to after tornado
one sample t test
Used to determine if a single sample mean is different from a known population mean Main difference: We just have one sample, one measurement We compare out sample values to a know population mean or historical mean.
Reporting t tests
When reporting t test results, you must include: Obtained t value Degrees of freedom df Significance value for the test p
