post hoc in Results--new answer

We were wondering that after contrasting groups and running appropriate post hoc analyses, what exactly we should include in the results section. We guessed simply the p value associated with the groups that are significant, because everything else is redundant. However, are these even necessary or is something like "___ post hoc comparisons show that the interaction is of main importance in that only Group __ is significantly larger than Group __.." ok? Thanks!

Viewing older posts

Just a little heads up if people were having this problem...

I'm not sure if this is an issue with blogger or my Firefox browser (or if I'm completely missing it), but I was trying to look back at the previous posts that were not listed on this page, but I couldn't find the "older posts" link at the bottom like we could for the Regression blogger. So at the moment it looks like the only way to access older posts is to go to the "archives" at the right side of the page and click on "January 2009" which would show all the posts from this month.

Assumptions: Independence of Observations vs Independence of Errors

Chapter 3 in T & F refers to "Independence of Errors" in its summary of analysis assumptions rather than independence of observations. Are these the same thing or different? And if they're different, is the assumption of "independence of errors" relevant to ANOVA or something else? I thought that independence of observations referred more to collinearity/orthogonality, which T & F discuss in a separate section, so just looking for clarification. Thanks!

Levene's test

This was Marisol's question posted to the regression blog by mistake...

I have three questions:1) Should we report the F and p value under the Levene's test in the write up for t-test2) When we are checking for the assumption of homogenity of variance, do we have to calculate F max or we can say that Levene's test shown that there is equal variances. Is Levene's test related to the assumption of homogenity of variance? If we have to calculate F max, from where in my printout do I get the S squared (or the necessary information to calculate it)?3) In class it was mentioned that it is worthless to report degrees of freedom as (1, 145) for t-test and that the appropiate way is to say just t(145), does that apply to one-wayANOVA when you are reporting the F?
Posted by Marisol De Jesus at 3:45 PM

Table

If a table is needed for data that has more than 3 numbers, do we count both columns and rows? If so, then is a table not needed for the t-test part of the homework, since there are only 2 items on the IV? Or do we still need a table since there are 3 items on the descriptive statistics?

a little heads up on how nit picky APA formatting can be...

Hey y'all,

As I've been grading article reviews, I've noticed a few APA formatting mistakes that most of you probably are not aware of making.

1. When it comes to seriation (i.e. making a list), APA requires that you do so by lower case letters in full parentheses, not numbers. e.g. GOOD = "Panera serves (a) bagels, (b) coffee, and (c) soup." BAD = "Panera serves 1) bagels, 2) coffee, and 3) soup."
2. Generally, numbers under 10 should be written as text, except for 6 types of exceptions, see slides 50-52 of the first link below). Numbers 10 and over should be written in numerical form (except for 3 types of exceptions, see slides 53).

All of this is spelled out in your APA manual. I also found some pretty good online resources that spell a lot of these issues out, www.buffalostate.edu/library/research/apa.ppt and http://www.docstyles.com/apacrib.htm. Just know that APA has an opinion about EVERYTHING, so always double check every stylistic effort, even if you think it's right.

Don't feel discouraged or too overwhelmed in trying to keep straight all of these little details. The more research type writing you do, the more you'll pick up on these. The only reason I came to know these little tidbits is after my advisor edited my Master's thesis 8 times :P and from an older student in my lab. So, seek out the edits of more experienced APA writers, and they will point you to the light of the APA format. Good luck and stay strong.

One-way homework

When selecting cases I clicked over the IV then typed in "It 3" and nothing happened. How do you selcet cases? This is causing an error message.

Thanks

Library Computer Lab

I have been informed that the 9 computers that have SPSS loaded on them have been moved to the room on the third floor that has no windows. My informant was pretty sure that this is room 309, but not 100% certain.

Hope this info is helpful to you in completing your homework assignment.

Table Example (click on image to enlarge)

Note use of italics, capitalization, lines, spacing, and alignment of decimals.

Results Example (click on image below)

Article Reviews at the Front Desk

The first set of article reviews are available for you to pick up at the front desk.

Please look at them before submitting article reviews 3 & 4 this week so that you can correct any errors you may have made. Note that it is fine to pencil in corrections that need to be made if you do not have time to re-print the article reviews today.

Article Review Question

For the research I am interested in 2 articles have been submitted for publication and I have access to them. I am doing my reviews on those articles (in press), is that okay?

Valerie

Qualitative Analyses for Article Reviews

I have a question related to the article review format we should be using for qualitative or phenomenological analyses in journal articles. I am aware there is a theoretical article format provided for us when reviewing those types of articles, but if I have a relevant qualitative analysis (e.g., an analysis with a formal hypothesis, method, result and discussion, but no empirical design or statistical tests) should I use the original (e.g., quantitative) format?

Thanks,

Chris

Power Analysis Software

This is anecdotal, but for those of you who are getting a step ahead on your research or are at least a little bit interested in stats, I wanted to provide a very useful link for a free power analysis program created by Faul & colleagues specifically for behavioral research. You may or may not note that (maybe Mari can prove me wrong here) that there is no power calculator in the actual SPSS bundled statistical package, but it is provided as separate stand-alone software.

Even if SPSS can provide power analyses, this is a useful stand-alone tool with good instructions and tutorials.

The program is called G*Power. Why it is useful:

1) It can calculate Ns and powers beforehand for research designs you may want to use to achieve an optimal effect size.

2) It is a very powerful program with relatively detailed instructions on its use, and can be used with t-tests, ANOVA, Regression, etc.

3) Once again, it is free.

You can download the program, as well as tutorials and the original published article on its creation here:

http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/

 

Cheers,

Chris.

Additional Review Sessions

We've added some review sessions before the three exams. The complete list is now:

Tuesday, January 13, 5 to 6 p.m.
Wednesday, January 21, 6 to 7 p.m.
Tuesday, January 27, 5 to 6 p.m.
Wednesday, January 28, 6 to 7 p.m.
Wednesday, February 4, 6 to 7 p.m.
Tuesday, February 10, 5 to 6 p.m.
Tuesday, February 17, 5 to 6 p.m.
Wednesday, February 18, 6 to 7 p.m.
Tuesday, February 24, 5 to 6 p.m.
Wednesday, March 4, 6 to 7 p.m.
Tuesday, March 10, 5 to 6 p.m.
Wednesday, March 11, 6 to 7 p.m.

All review sessions will be held in Room 314 Weyerhaeuser (a/k/a Psych Building).

Organization of Files on Portico

I noticed that with all the files for the quarter posted on Portico, the list was a little overwhelming. So I have created subfolders with different types of files in them. I hope that it is all relatively self-explanatory and clear, but in case you run into problems...

When you click "Files" under the "My Courses" tab on Portico, that will bring you to a page where the syllabus for the class is posted. You will also see subfolders entitled "Lecture Handouts," "Homework Assignments," "Readings," and "Review Sheets."

Under Lecture Handouts, the slides from all lectures throughout the quarter are posted. The only slightly tricky things about this are that (a) Portico alphabetized "week 10" right after "week 1" (not after "week 9") and (b) the "week" designator can be slightly misleading. That is, Oneway ANOVA really begins in week 2 and continues into week 3. All dates on your syllabus are correct, and the titles of the lecture handouts are rough approximations.

Under Homework Assignments, you will find the descriptions of all four homework assignments and the final project. Note that the necessary SPSS files for completing assignments 1 and 2 (oneway ANOVA and factorial ANOVA) are on the CD included with your text book. If you purchased the book used, it is fine to get a copy of those files from a classmate who has the CD. If you did not purchase the book, I strongly recommend that you do so, as it is a very readable and helpful stats book.

Under Readings, you will find supplemental materials including an additional reading on Latin Squares Designs, pointers for completing Article Reviews, and the listing of faculty recommended articles for review. Consulting these documents will save you time and help you correctly complete homework assignments and exams.

Under Review Sheets, you will find the questions for review for each exam.

PP Notes

Hello, I noticed that the power point notes were also posted on portico, but will we still get hard copies in class, or do we need to print them and bring them to class ourselves?

New Files Posted on Portico

There are four new files on Portico that you may find useful in completing your article reviews.

(a) The file, "fac student refs 2008-09 oct 28.pdf," contains faculty recommendations for student reading. It is the same file as last quarter.
(b) The file, "Corrected Article Summary Outline v.2.3.pdf," contains and example article review along with the required headings for both empirical article reviews and (at the bottom, theoretical article reviews). It has been corrected to show proper placement of the pseudonym, to illustrate how best to title the article review, and to fix the capitalization of one of the headings. Please attend carefully to these and all other formatting issues in article reviews.
(c) The file, "Article Review Pointers," outlines common errors in article reviews. it will probably be especially useful to those of you who did not complete article reviews last quarter, but it is worth reading for everyone.
(d) The file, "Pseudonym List all," includes all pseudonyms for all assignments throughout the quarter. I would encourage you to printout this list and keep a record of your pseudonyms as a backup to the class pseudonym list that will be generated for each assignment.

BTW

If you were wondering about the pretest answers, you can find them in the January 2007 archive.

Review Session Times

The review sessions are currently scheduled for

Tuesday, January 13, 5 to 6 p.m.
Wednesday, January 21, 6 to 7 p.m.
Tuesday, January 27, 5 to 6 p.m.
Wednesday, February 4, 6 to 7 p.m.
Tuesday, February 10, 5 to 6 p.m.
Wednesday, February 18, 6 to 7 p.m.
Tuesday, February 24, 5 to 6 p.m.
Wednesday, March 4, 6 to 7 p.m.
Tuesday, March 10, 5 to 6 p.m.

All review sessions will be held in Room 314 Weyerhaeuser (a/k/a Psych Building).

Congratulations, You Found It!

This is the blog for ANOVA, so if that's what you were looking for, you have come to the right place!