When all other factors are held constant, is there a non-chance difference in response latencies between grammatical (1) and ungrammatical (0) sentences elicted from 21 native speakers of English? Answer this question based on the data below. 

Use the ASCII data in "j:\ling\msdos\otherdata\anovastudy1.sav" for a complete data list.

Data file:  anovastudy1.sav
Variables: GRAM, LATENCY

ITEM    ORDER  GRAM    MATCH  RESPONSE CORRECT  TRIES   LATENCY
1       47        1        1        1        1      1      1641
2       48        1        1        1        1      1      1306
3       31        1        1        1        1      1      1543
4       3         1        1        1        1      1      1661
5       61        1        1        1        1      1      1693
(data continue)

Answer the following questions:

1. Do these data meet the assumptions of ANOVA or Kruskal-Wallis? Explain in full.
 
 
  

2. What kinds of data do the two variables (i.e., GRAM and LATENCY) represent?
 
 

3. How many independent variables and levels do the data have?
 
 
 

4. Explain which statistic you ran and why.
 
 
 

5.  Copy the F-statistic or the H-statistic (whichever is relevant) and detemine if the .05 level of significance is attained.
 
 
 
 

6. What is the direction of the difference? That is, if there is a significant difference between the grammatical and ungrammatical sentences, which is higher?