I am currently working on trend analysis of time series data. For that purpose I have used Mann Kendall and Sen's Slope techniques. Some of the results show insignificant Z scores with higher Q score than the Q scores of significant Z scores.
My question is, do Z and Q scores not related in a sense that significant Z score should have high ±Q scores ?
(the bold values in the attached results show significant trend at alpha=0.05)
It gives you a Z statistic when you use the Mann Kendall test.
In this test, we examine the existence of a long-term trend in the data and the significance of this trend. In this analysis, two types of trends (decreasing or increasing) are tested.
I can imagine what are your Z scores but what do you mean by the Q scores? Can you give more information on your notations?
I should add that when several tests exist for a given hypothesis, there is no guarantee that they will all lead to the same conclusion. Each test is based on some assumptions (in your case, there may be something else than a trend and a random component) and is based on some statistic. Some of these tests can be more powerful than others against specific alternatives.
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series