Using bpstudy.sav, conduct an independent samples t test in SPSS with gender as the grouping variable (male = 1; female = 2) and HR1 (heart rate) as the outcome variable.
Paste the SPSS output and then report:
- The sample size for males ( n1) and sample size for females ( n2).
- The means for males ( M1) and females ( M2) on HR1.
- The calculated mean difference ( M1 – M2).
- The standard deviations for males ( s1) and females ( s2) on HR1.
- The Levene test (homogeneity of variance assumption) and interpretation.
- t, degrees of freedom, t value, and probability value. State whether or not to reject the null hypothesis. Interpret the results.
- Calculate Cohen’s d effect size from the SPSS output and interpret it. Specifically, if the homogeneity of variance assumption is met, divide the mean difference ( M1 – M2) by either s1 or s2. Violation of the homogeneity of variance assumption requires calculation of Spooled. Homogeneity assumed:
- Cohen’s d = ( M1 – M2) ÷ s1 or Cohen’s d = ( M1 – M2) ÷ s2
- To be comprehensive, report Cohen’s d based on a calculation with s1 and a calculation with s2. Round the effect size to two decimal places. Interpret Cohen’s d with Table 5.2 of your Warner text.
Section 2: Post-hoc Power Analysis
Open G*Power. Select the following options:
- Test family = t tests.
- Statistical test = Means: Difference between two independent groups (two groups).
- Type of power analysis = Post hoc: Compute achieved power.
- Tails(s) = Two.
- Effect size d = Cohen’s d obtained from Section 1 above (using either s1 or s2).
- α err prob = standard alpha level.
- Sample size group 1 = n1 from Section 1 above.
- Sample size group 2 = n2 from Section 1 above.
- Click Calculate.
Provide a screen shot of your G*Power output. Report the observed power of this post-hoc power analysis. Interpret the level of power in terms of rejecting a null hypothesis. Do you have sufficient power to reject a false null hypothesis? Interpret power in terms of committing a Type II error.
Section 3: A Priori Power Analysis
In G*Power, now select:
- Type of power analysis = A priori: Compute required sample size.
- Input effect size d from Section 1.
- Specify α err prob.
- Specify Power (1 – β) = .80.
- Set the Allocation ratio to 1 (i.e., equal sample sizes).
- Press Calculate.
Provide a screen shot of your G*Power output. Interpret the meaning of a .80 power value. Specifically, report the estimated n1, n2, and total N to achieve obtain a power of .80. How many total subjects ( N) would be needed to obtain a power of .80? Would you have expected a required N of this size? Why or why not?
Next, in G*Power, change the Cohen’s d effect size value obtained in Section 1 and set it to .50 (conventional “medium” effect size). Click Calculate. How many total subjects ( N) are needed to obtain a power of .80? Compare and contrast these two estimated Ns.
In conclusion, reflect on the importance of conducting an a priori power analysis in psychological research plans.