6.5.2. Jonckheere’s Trend

This test is used when there are three or more conditions and it is possible to predict the ordering of results. A predicted trend across the scores of different groups is evaluated without ranking the data. Data should be arranged in predicted order, from lowest to highest.

Data entry is in multisample format (see 6.0.4. Multisample Tests). Each sample should be entered in a separate column (not necessarily of equal length). Missing values are omitted by case.

The test statistic displayed is corrected for ties. The one-tail probability is reported using the normal distribution.

Example

Table 112 on p. 275 from Cohen, L. & M. Holliday (1983). Correctly spelt words by pupils of three different ages are given. The null hypothesis “spelling ability is the same for all age groups” is tested.

Open NONPARM1, select Statistics 1 → Nonparametric Tests (Multisample) → Jonckheere’s Trend and include Seven Year, Eight Year, Nine Year (C16 to C18) in the analysis by clicking [Variable] to obtain the following results:

Jonckheere’s Trend

	cases
Seven Year	10
Eight Year	10
Nine Year	10
Total	30

 

P =	198.0000
Q =	50.0000
P-Q =	148.0000
variance P-Q =	2635.5501
Z-Statistic =	2.8829
1-Tail Probability =	0.00197

 

This result is significant at the 1% level. Hence reject the null hypothesis.