Exploring the Relationship between Race, School Skipping, and Substance Use

ANOVA Test for School Skipping and Race

1.Are students of different races more likely to skip out on school? Criminologists are interested in this question because school acts as a form of social control; if you’re at school, you won’t have as many chances to engage in crime or interact with the justice system! We  can test this question using a one-way ANOVA in SPSS with the variables “SchoolSkip” and “V1070” 
 
a. First, state the null and research hypotheses for this ANOVA.
b. Second, determine the critical value of F you will need for your results to be significant at p < .01. You will need to calculate your within-group and between-groups degrees of freedom in this case. 
 

c. Estimating an ANOVA model in SPSS: Doing this by hand would take weeks, but SPSS does it in seconds. To do this, select analyze->compare means->one-way  ANOVA. Place the variable “V1070” in the “factor” slot and “SchoolSkip” in the dependent list: 
 
d. What is your F statistic? What do you conclude about the null hypothesis in this case?

2.You may have noticed that an ANOVA test on its own isn’t a very precise assessment of our research question. Let’s follow this up by estimating the difference between pairs of means and the strength of the association: 
 
a. Post Hoc Tests in ANOVA: In the one-way ANOVA menu, you will notice an option to “post-hoc” tests. Click that button and select your post hoc test 
 
i. The output provided compares one racial category with all the others, providing the statistical significance of each difference.  
 ii. Interpret the remaining differences between racial groups. Are all groups significantly  different from one another? Which group means are closest/furthest from one another? 
iii.  What new information does this test add to our ANOVA interpretation?  

3.Do students who hang out with individuals who drop out of school also tend to skip out on school themselves? Run a one- way ANOVA test with pairwise comparisons using the variables “V7253” and “SchoolSkip”: 

a. Conduct the hypothesis test fully, stating your null hypotheses and identifying test statistics. What do you conclude about the null hypothesis? 
b. What do the pairwise comparisons in this case tell you? What groups are most prone to skipping school? 
 
4. Often researchers talk about the importance of socioeconomic status in producing heightened risk for criminal and delinquent activity. The education of a student’s parents is one way to look at this difference. Let’s see whether the variable “LowParentEduc” is associated with higher levels of student smoking and drinking: 

a. Conduct a hypothesis test to examine whether there is a relationship between parent education and child alcohol and tobacco use. Fully state your null hypotheses and identifying test statistics. What do you conclude about the relationship?

b. What do the pairwise comparisons in this case tell you? Use an alpha of .001 for this test. Are any groups not statistically different from one another? 
 
c. Take a moment to consider the “why” part of this result; what mechanisms might explain why the results you have found.