1
$\begingroup$

In my study, we investigated the effect of sex and location on fish length taken from 3 different regions: A,B, C. From each site we randomly took 10 females and 10 males. Now I am going to test whether sex and/or region affect fish length. I am not really sure which method should I take to analyse my data. 2 way ANOVA with randomized block design? Please guide me on this.

  A       B       C
50.1    53.4    54.0
52.8    55.2    49.1
50.8    51.0    60.5
58.8    59.3    57.8
59.7    61.5    48.7
49.0    61.2    57.0
58.8    57.8    61.1
62.2    50.1    62.8
57.8    56.0    59.8
61.2    56.5    60.3
        Female
46.5    57.5    49.1
44.4    59.3    51.8
42.0    62.4    55.3
51.1    61.1    43.6
45.8    59.9    50.1
46.3    55.6    51.0
41.8    56.8    49.0
52.0    59.2    48.8
46.5    50.4    52.0
39.0    47.8    43.0
$\endgroup$
  • $\begingroup$ It looks like you have 2 factors of interest and no subplots, so a CRD makes sense. Is there some other variable you want to block on? Ideally, this would all be decided prior to running the experiment. $\endgroup$ – Joel Mar 3 '17 at 15:35
  • $\begingroup$ My supervisor asked me to test whether sex and/or region affect fish lenght. He did not say anything about a nuisance factor!! I just assume that region might be. I dont know what I should do $\endgroup$ – Non Mar 3 '17 at 16:13
2
$\begingroup$

Assume you're happy with two-way ANOVA assumptions.

There is no reason why you can't do RBC here, if the goal of your experiment is to have ANOVA between fish length and gender while keeping location as nuisance factors.

You can also do a two-way ANOVA with interaction if location is a variable that is interesting in your analysis (I doubt).

You'll need to ask yourself whether location is something you care, it should be part of your experimental design.

$\endgroup$
  • $\begingroup$ +1 makes sense, although the OP did say investigated the effect of sex and location on ... Perhaps that's not actually what they intended. $\endgroup$ – Joel Mar 3 '17 at 15:56
  • $\begingroup$ My supervisor ask me to test whether sex and/or region affect fish length. I am not really sure which I have to consider the region as nuisance or not? what do you suggest? and whether RCB is different from RBD? $\endgroup$ – Non Mar 3 '17 at 16:09
  • $\begingroup$ @Non then region is a factor you are interested in analyzing and you do not want to use it as a blocking factor. $\endgroup$ – Joel Mar 3 '17 at 19:08
  • $\begingroup$ @Non That depends whether you want location in your analysis. $\endgroup$ – SmallChess Mar 4 '17 at 3:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.