PG851: GLM--ANOVA

This blog is for the use of students enrolled in either section of PG851. It is intended to be a forum for asking, discussing, clarifying, and helping.

Friday, March 19, 2010

reporting sig vs. insignif findings

While usually inappropriate to look at p comparisons/post hocs if there are no significant interaction effects/main effects with more than two levels, I was looking through my notes and wrote something down during last class about still running them on the final project if it is in your specific hypothesis to compare groups (even if you dont get a significant effect) -? Does that sound right or did I write it down wrong? If we get an insignificant effect for interaction/main effect with more than two levels, should we still run and report insignificant post hocs along with the other insignificant effects for the project? Thank you

Thursday, March 18, 2010

If you haven't yet talked to me about your analysis...

...you are rapidly running out of time to do so.

In the event that you are unable to formulate your hypothesis by noon Friday, I will not be able to provide feedback to you.

To that end, here are the common issues students have been running into.

1. When I say "two or more of the constructs," what I mean is either (a) two predictors (i.e., factorial ANOVA), (b) a predictor and a covariate (i.e., ANCOVA), or (c) two outcomes (i.e., repeated measures). As I said in class, a simple oneway ANOVA will not be acceptable for this project. Some of you have spoken to me about specific variations on a oneway, and if I have approved your analysis, that is fine.

2. Make ABSOLUTELY certain that you are using the correct types of variable (i.e., categorical or continuous) in the correct places in your analysis (i.e., IV, CV, DV).

3. You must observe the ANOVA bare minimum of 10 per cell. This means that if you cross two variables, every cell created by the crossing must have at least 10 participants.

4. You must detail every change you made to the data set, including data transformations, variable recoding, and participant dropping, in a Preliminary Analyses subsection of the Results. If I cannot replicate your findings with the data set, this will be a very bad thing.

5. Your writeup must be prepared to APA style requirements.

4. Finally, the project MUST be submitted electronically, through turnitin.com, by midnight Friday. No exceptions. Failure to submit the project will result in a Hold for the class (if and only if you were prevented from submitting the project because of a true, verifiable emergency) or a score of no credit for the assignment.

means

do we report the interaction means (as in not the total means) in the preliminary analysis section?

wording

If we only made one change/transformation to a variable, do we title the section "Preliminary Anaylsis" or "Preliminary Analyses"...is the heading "analyses" used regardless of how many were done?

Final Project Reminder

As I said in class and in another posting on the final project, the final project is due on turnitin.com by midnight Friday evening.

Do NOT turn them in to my box or the front desk. I will not grade final projects submitted in hard copy form.

If you have already submitted your project in hard copy form, please return to the front desk and pick it up.

hypothesis

should we state our hypothesis in the results section? for our past HW assignments, we just reported our findings in the results section.

-sarah

Wednesday, March 17, 2010

write ups

The final day of class Mari went over what was involved in the write ups for our result section. She gave an example results section with 2 parts. Can you remind me why we would need to divide our results section?

additionally, what point/part in the results do we state that we deleted cases or excluded variables?

Thursday Final Exam

The exam on Thursday will be in Room 314. (We'd be way too crowded in 126!)

Monday, March 15, 2010

Final Projection Questions

Someone asked...

1) I have dichotomized the values for DV for the final project. When I run assumptions, am I suppose to run the DV from the original data or the dichotomized one?

2) I ran box and whisker's plot for dichotomized one and it's absurd looking.

Sunday, March 14, 2010

ANCOVA grades & feedback up on turnitin.com

Your grades are up on turnitin.com. I would encourage you to (as always) look over the comments made on your paper.

This is later than planned because on Thursday, I came down with the same illness that my son and husband have had for weeks. Sigh.

Thursday, March 11, 2010

Powerpoint

Hey all...

FYI I put the powerpoint from the review session on Moodle under Forum. Just make sure to read the Warning on the first slide.

Dichotomizing

Hi!
I am getting a "no valid cases found" message after I attempted to dichotomize one of my IVs. I thought I followed the steps carefully. Do you know what this could mean or what I am doing incorrectly?

Thanks so much!

Final Project Deadline

Final projects must be uploaded to turnitin.com by midnight on Friday, March 19.

Monday, March 08, 2010

Non-significant result in final project

For the final project, I wonder if it is acceptable when I can't establish a significant result based on my original proposal? This is after I have performed the necessary assumption checks and any transformation of data necessary. Thanks.

***Review Session***

Hey All!

You're almost done with the quarter! Congrats!

Well, our third review session is going to be held this thursday, March 11, from 6-7:30. We will be in the psych auditorium again (Travis). Please look over the review and come with any questions!!!

See you then!

Wednesday, March 03, 2010

when did intervention occur?

For our hw, when was the intervention done in relation to the collection of sensxf1m and sensxf2m?

That is, did the intervention take place between the two measurements, or did it happen before both?

HW question

I'm sorry as I feel like this may be obvious, but I'm confused on what Q#5 is looking for from the ANCOVA HW? Should we include stats...are we supposed to reiterate what we spoke of on question 4? Can we just say they're slightly different or similar in these ways but not these...?

Tuesday, March 02, 2010

Last Exam Scheduling

All students in both sections are welcome to take the third ANOVA exam EITHER on Monday, 3/15, at 8 a.m. in Psychology 130 or on Thursday, 3/18, at 8 a.m. in Psychology 126.

Monday, March 01, 2010

ANCOVA question

Below the Estimated Marginal Means table on SPSS output, it says that "CVs appearing in the model are evaluated at the following values..."

I'm a little confused on whether that value is the mean for the CV or the DV...in other words, it is the value that everyone is assumed to be the same on, on the DV?

Sorry, that's confusing...I guess I'm asking...what exactly is that value?

Assumptions for ANCOVA

I'm wondering which assumptions we should have plots, etc for shown on our HW for ANCOVA? Do we need Cook's d, linearity, and homogeneity of regression, more specifically?

Thanks so much!