Summary of Statistical Tests (03 25 08) (sample results below were fabricated)

Return to Opening Page of PSYC 4650

 

T Test

To compare the means on a dependent variable of two independent groups.  E.g., mean mental health scores of ADD vs. non ADD groups of children. 

 

Sample results:  The mental health score of children with ADD (mean= 3.34) was significantly lower than that of children without ADD (mean=4.45), indicating that ADD children have more mental health problems than nonADD children,  t (25)= 2.43, p<001 .

 

Correlation

To compare the relationship between two sets of measures on individuals to see if the scores are related.  E.g., relationship between mental health scores and birth weight of ADD and nonADD children.

 

Sample results:  There was a small (r=.14) but statistically significant (p<.001) positive correlation between birth weight and mental health scores, indicating that the lower the child’s birth weight, the more mental health problems the children had.

 

One Way ANOVA

To compare the means on a dependent variable of several independent groups.  E.g., mean mental health scores of ADD vs. LD vs. non ADD/non LD.

 

Sample results: Mental health scores of ADD children (mean=3.51) and LD children (mean =3.47) were lower than mental health scores of non ADD, nonLD children (mean=4.43), and the differences were statistically significant, F (2, 469) = 57.23, p<.001.

 

Two-Way (multi factor) ANOVA

To compare means of a dependent variable among two or more levels of two or more factors (independent variables). Includes study of interactions of the independent variables.

E.g., comparison of mean mental health scores of ADD vs. LD vs. nonADD/ nonLD and Boys vs. Girls.

 

Sample results: Two-way ANOVA showed that mental health scores of ADD children (mean=3.51) were significantly lower than nonADD children (mean= 4.32), F (1, 359) = 23.91, p<.01.  Also, mental health scores of boys (mean= 3.01) were significantly lower than mental health scores of girls (mean=4.21), F (1, 359) =15.92, p<.05.  Furthermore, there was a significant interaction between the gender and ADD/nonADD factors, F (1, 359) = 6.45, p <05. Post hoc analysis using the Tukey test showed that mean of ADD boys was significantly lower (p<.05) that the means of the other groups, which did not different among themselves.

 

Cross Tabs, Chi Square

Cross tabs examines the frequency or percentage of pairings between two categorical variables. Chi Square provides a statistical measure of two variables which are theoretically expected do be independent vs. whether they are actually observed to be independent of each other. E.g., % boys vs. girls with ADD vs. nonADD.

 

Sample Results:  A cross tabs analysis paired percentages of  boys vs. girls,  with percentages of ADD vs. nonADD cases.  The overall chi square was statistically significant, Χ2 (1) =149.04, p<001.  Follow up with a nonparametric chi square test showed that the observed percentage of boys with ADD (73.9%) vs. girls with ADD (26.1%) was significantly different, Χ2 (1) = 158.37, p<.001, from the expected (49% vs. 51%).  This result indicated that boys had ADD at a significantly higher rate than girls had.

Return to Opening Page of PSYC 4650