11
$\begingroup$

I often find myself explaining (or wanting to explain but not wanting to be boorish) the basics of random sampling and the consequences of pseudoreplication, specifically the limitations and assumptions of the standard statistical methods that are often misapplied to such data.

As a student, I found a particular reference (Hurlbert 1984) to be clear and informative. I am considering sharing it with my colleagues.

This paper has been persistently cited over the last 30 years, so many in the field still consider it relevant.

enter image description here

For reference, here are a few figures that provide a flavor of the manuscript's contents. I suspect (project?) that these two figures represent a substantial fraction of information many readers take away.

enter image description here

I have two questions:

  1. Are any of the ideas in the paper 'out of date' to the extent that this is not a good introduction to the topic of pseudoreplication (and if so, which ones)?
  2. Are there any references that provide a similar information in a way that is more easily digestible (and requires less motivation) than this 25 page monograph.
  3. Is there a better overview of modern concepts or alternatives that allow valid interpretation of results but that don't require fundamentally different analyses?

References:

$\endgroup$
1
  • 4
    $\begingroup$ To my knowledge, it is the best explanation of pseudoreplication and definitely worth the read. I haven't read it in a while, but have you read the answer of Hurlbert to Oksanen ? It might summarize the info of Hurlbert 1984 Hurlbert, S. H. (2004). On misinterpretations of pseudoreplication and related matters: a reply to Oksanen. Oikos 104: 591-597. $\endgroup$ – Emilie Sep 18 '14 at 17:11
10
$\begingroup$

Well thanks for the nice comments, folks. Since I'm still kicking, I can let you know about several other paper of mine that deal with pseudoreplication, as well as ones that deal with other important statistical issues (multiple comparisons, pseudofactorialism, collapse of the Neyman-Pearson framework, terminological confusion promoted by statisticians, misuse of one-tailed tests) by referring you to my website at http://www.bio.sdsu.edu/pub/stuart/stuart.html

On topic, Hurlbert & White (1993) clarified the definitions of the different types of pseudoreplication, and Hurlbert('Ancient black art,' 2009; 'Affirmation', 2013) review the topic and related terminological issues.

Pdfs of all these papers are available on the website.

Keep calm and carry on! s.h.

I should add, in response to Emilie's Sept 25 comment, that pseudoreplication can never be "valid" as by definition it is a statistical analysis (and interpretation) that does not accord with the study design. It is not defined as conducting an experiment with only a single experimental unit per treatment or simply taking multiple samples from a single experimental unit.

$\endgroup$
2
  • 4
    $\begingroup$ Thanks for contributing! I am from a field where the term "pseudoreplication" is not commonly used. In your view, would it be fair to describe pseudoreplication as "violation of independence assumptions," or are there other important things that technically differentiate it from that? $\endgroup$ – Jake Westfall Dec 11 '14 at 21:04
  • $\begingroup$ Just saw you answer, years after. By valid, I meant it's still a valid concept. Not something that can be done. $\endgroup$ – Emilie Sep 21 '16 at 13:20
6
$\begingroup$

As nobody dared answered, I will expand my commentary.

I personnally think that Hurlbert 1984 is an essential article for all experimenters. Hurlbert coined the word, but also made a detailed article explaining the problem, with examples, potential sources of confusion and a litterature survey. Can we ask for more ?

It is indeed a long article. I am sure we can found in books, about statistics or experiments, good introductions of pseudoreplication. But Hurlbert is widely accessible and even tought not published in open access, I'm pretty sure anyone can find a copy (another advantage).

If someone want something shorter or want to go further in the subjet, I suggest looking at Hurlbert 2004, who resume misconceptions about pseudoreplication. In the last paragraph of the first page, Hurlbert cites many articles discussing the subjet and even suggest a good book. The main problem of the 2004 article (or the fun in it), is that it is a bit acrid in the tone, as it's an answer to Oksanen 2001. I would still recommand it.

References

Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54: 187-211.

Hurlbert, S. H. (2004). On misinterpretations of pseudoreplication and related matters: a reply to Oksanen. Oikos 104: 591-597.

Oksanen, L. (2001). Logic of experiments in ecology: is pseudoreplication a pseudoissue? Oikos 94: 27-38.

$\endgroup$
1
  • $\begingroup$ I realize I don't answer to the third question, but I am not sure I understand it. Pseudoreplication is still valid and you have to analyze data in a way that correspond to your design. $\endgroup$ – Emilie Sep 25 '14 at 14:37

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.