I am doing a study based on the speech measures with gender, health condition, different passages read and different number of sentences from each passage. The number of levels in these factors are 2 for gender, 2 for health condition, 2 for different passages, 4 (for first passage) or 5 (for second passage) for number of sentences. My data set is unbalanced.
I just want to know what kind of ANOVA to use and how to do it on SPSS. Any help is much appreciated.
It depends on your hypothesis. you can use One way, two way or ANOCOVA
Thank you for the response. If it helps, my primary goal is to see if there is a significant difference between the healthy and control groups.
What is your response variable?
I would code each factor which will require 1 dummy variable for gender (d1), 1 for health condition (d2), and for the two remaining factors 3 (d3, d4, d5) and 4 (d6, d7. d8 and d9), respectively. Then use regression to compare and contrast the various levels and determine if there are any significant interactions.
if response is numberic..
y = beta0 +beta1d1+beta2d2+...+beta9d9+APPROPRIATE INTERACTIONS
otherwise, if the response categorical (health and control group) then using logistic regression (logit or probit - I prefer logit)
Logit y = ln( p / (1-p)) where p is the probability of being in either the healthy group or the control group
not only depends on the research question, but also you have generated the corresponding experimental design
also by what I see your variables are discrete ...
Hello everyone,
Thank you for the valuable inputs. Based upon the answers and some research on the internet, I figured that I need to use the "Repeated Measures" in the "General Linear Model" in SPSS. To be clear, I have speech measures evaluated on a total of 9 sentences spoken by a same set of people who can be divided into 4 groups based on the gender and health condition. In the process, I saw there are options to choose "Between-Subject Factor(s)" and "Covariates". I would like to know the difference between those two types, so that I can decide where to place the two variables (gender and health condition) in the analysis.
Thanks in advance
what type is your dependent variable, because is Ordinal or DIscrete, you can`t apply General Linear Model you should look for a non-parametric test
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA