Biostatistics
1
0
Entering edit mode
2 days ago
RCMC • 0

enter image description here

enter image description here

Good day, apologies if this is not so much of bioinformatics but biostatistics

Levene's test shows significant difference in homogeneity suggesting Welch ANOVA should be used.

The welch ANOVA shows significant difference (>0.05) so I used the Games-Howell Post hoc test to determine which groups are significant. But here is the problem, none of the groups seems to be significant to each other while the ANOVA shows there is significance.

I also tried to dissect the data using descriptives and I think I am confident that Place 3 should be statistically different to Place 1 and 2.

How do I interpret this data on results and discussion and is this kind of cases common in games-howell? should standard post hoc like scheffe be used?

ANOVA Post Biostatistics Hoc • 287 views
ADD COMMENT
0
Entering edit mode

Just to reinforce @gordonsmyth's message. You are lucky that he answered. I think in future you are unlikely to get an answer on here without setting your question in the biological context. Just place a load of statistics down and asking what test to use is unlikely to elicit useful responses.

ADD REPLY
4
Entering edit mode
2 days ago
Gordon Smyth ★ 7.6k

IMO, you relying much too much on statistical tests to decide what analysis to do, whereas you should be plotting the data and doing some basic QC. A simple scatter plot or beeswarm plot would give complete information. Judging from the descriptives, the data seems to show the standard sort of mean-variance relationship (with SD proportional to the mean) characteristic of strictly positive data. Almost certainly, you should apply a variance-stabilizing transformation (log, started log, or sqrt) and then do ordinary anova.

In future, you might consider sending purely statistical questions like this to the Cross-Validated forum on stackexchange. Statistical analysis requires thinking about the measurement scale of the data. If your data are measurements on a strictly positive scale, that is something that should be explained as part of the question.

ADD COMMENT
0
Entering edit mode

This is noted and suggestion is taken

This will be a great help

ADD REPLY

Login before adding your answer.

Traffic: 1596 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6