Inferential Statistics

As shown in the section on descriptive staistics, the mean aggressive-word priming effect was 0.7211. This value is significantly different from 0, t(31) = 2.2411, p = 0.0323 (Computed by the t-test/confidence interval procedure of the analysis lab). Therefore, preceding an aggressive word with a weapon prime decreases the time it takes to name the aggressive word.

More important is the finding that the"prime difference" score mean of 0.8432 is significantly different from zero, t(31) = 2.1724, p = 0.0376. (Again, computed by the t-test/confidence interval procedure of the analysis lab). Therefore, it can be concluded that weapon words prime aggressive words more than they prime non-aggressive words.

Since the distribution of prime difference scores is not normal, the effect of violating the normality assumption should be considered. Using the analysis lab, 50,000 simulated experiments sampling from a population with a mean of zero and the same shaped distribution as occurred in this sample were conducted. Of these, 0.051 rejected the null hypothesis at the 0.05 level. Therefore, the non-normality present in these data does not appear to matter.

t-test in SAS
t-test in SAS JMP
t-test in SPSS

Within-Subjects ANOVA
Analysis of variance (ANOVA) is an alternative computational approach that results in identical results. For this design there are two within-subject factors: prime type (whether a weapon word or a non-weapon word prime is used) and word type (whether the target word is an aggressive word or a non-aggressive word). The hypothesis is that the effect of prime type will be larger for aggressive than for non-aggressive words. In ANOVA terms, the hypothesis is that there will be Word Type x Prime Type interaction.

The output for SAS is shown below. Notice that the p value for the interaction (shown in red) is the same as the p value of 0.0376 computed using the one-sample t-test.

Within-Subjects ANOVA by SAS

How this was done

 
                      General Linear Models Procedure
                    Repeated Measures Analysis of Variance
                     Repeated Measures Level Information
 
 
          Dependent Variable         AW       AN     CXEW     CXEN
 
 
 
           Level of WORDTYPE          1        1        2        2
              Level of PRIME          1        2        1        2
 

         Univariate Tests of Hypotheses for Within Subject Effects
 
  
Source: WORDTYPE
                                                               Adj  Pr > F
 
     DF     Type III SS     Mean Square   F Value   Pr > F    G - G    H - F
 
      1      0.44281108      0.44281108      0.21   0.6468    .        .
  
 
Source: Error(WORDTYPE)
  
     DF     Type III SS     Mean Square
 
     31     64.12305123      2.06848552
 
 
Source: PRIME 
                                                               Adj  Pr > F
 
     DF     Type III SS     Mean Square   F Value   Pr > F    G - G    H - F
 
      1      2.87184578      2.87184578      1.97   0.1703    .        .
 
 
Source: Error(PRIME)
 
     DF     Type III SS     Mean Square
 
     31     45.17844844      1.45736930
 

 
Source: WORDTYPE*PRIME
                                                                Adj  Pr >  
     DF     Type III SS     Mean Square   F Value   Pr > F    G - G    H - F
 
      1      5.68603503      5.68603503      4.72   0.0376    .        .
  
 
Source: Error(WORDTYPE*PRIME)
  
     DF     Type III SS     Mean Square
 
     31     37.35101570      1.20487147
 
 

Within-Subjects ANOVA in SAS JMP

Within-Subjects ANOVA in SPSS