Greetings!
G*Power indicates that 128 is the target sample size for a one-way ANOVA (effect size = .25; significance = .05; power = .8). For my study, I've got one categorical independent variable comprising two groups: a treatment group and a control group. With a sample size of 128, I'd need 64 participants for each group. However, my treatment group number is 72, and my control group number is 58 (which is smaller than 64). Combined, there are 130 total.
Is it ok to run the analysis with these numbers for my groups? What do you guys recommend that I do to mediate this discrepancy?
If I bump the effect size to .4, then my sample only needs to be 106, and my numbers would be good. But I really don't want to do this. :-/
Dear professor, as you must be knowing that the majority of statistical programs can automatically handle the issue of sample size. It is thus just to run ANOVA as usual and do not worry about sample sizes, i think so.
Dear professor, as you must be knowing that the majority of statistical programs can automatically handle the issue of sample size. It is thus just to run ANOVA as usual and do not worry about sample sizes, i think so.
just check the normalcy and just go ahead. ANOVA will not be affected. only thing u need to check is can it be applied.
Dear Shannon McFarlin ,
Yes, ANCOVA can be run as along as the same KEY assumptions for ANOVA are met AND as long as your data further satisfies the following additional three assumptions required for ANCOVA:
(1). For each independent variable, the relationship between the dependent variable and the covariate should be linear,
(2). Lines expressing linear relationships are all parallel, ie. homogeneity of regression slopes, and
(3). The covariate is independent of treatment effects; ie the covariate and independent variables are independent
Shannon McFarlin ,
I have just noted something of concern to apply ANOVA in analysing your dataset.
Please remember that Analysis of Variance (ANOVA) is a parametric test which compares means of three or more independent groups to assess whether there is statistical evidence that population means are significantly different.
In your dataset, you indicated that your one independent categorical variable has two groups (treatment group and control group). However, the assumption or requirement for ANOVA that there should be at least 3 groups is being NOT met, which may render the application of ANOVA inappropriate. In addition, the measurement level of your dependent variable must be continuous - another key requirement/assumption for ANOVA.
If the measurement level of your independent variable is continuous, with a categorical independent variable having at least three (3) mutually exclusive groups, then the nature of such data can best be analysed by testing for mean differences of the dependent variable across groups. Application of ANOVA requires fulfillment of the following assumptions:
(a). Dependent variable must be continuous (interval or ratio level).
(b). Independent variable must divide cases into three or more mutually exclusive groups.
(c). Independent samples or groups (independence of observations)
(d). There is no relationship between the subjects in each sample.
Please advise on the following:
(1). What the measurement level of your dependent variable is?
(2). What is the main objective of your study? -
(a). to test for group mean differences? - [paired sample t-test may be suitable], or
(b) to evaluate the effect of a particular intervention in the pre-intervention phase and post-intervention phase between the control group and intervention group?
(3). What is your research design?
Regards
Analysis of variance (ANOVA) is a statistical technique that is used to check if the means of two or more groups are significantly different from each other. @Elvis, was wrong and seriously on that issue.
A main practical issue in one-way ANOVA is that unequal sample sizes affect the robustness of the equal variance assumption, except weights are incorporated,( and power only is affected for the smaller sample) but you don't need to worry about that since softwares use the right approach for the analysis. Professor, @Shanon if you are convinced there is a third variable involved(a Covariate), you can have an issue with Simpson’s Paradox. You may or may not have collected that third variable but that could be good for an ANCOVA as alternative. Best wishes, Ibiloye A. C.
Abiodun Christian Ibiloye ,
I NEVER wrote "two" or more groups, but rather "three" and more. - An independent-samples t-test is commonly more applicable and used for two groups. The only assumption you are focusing on relates to unequal variance? What about the following assumptions (a). the experimental errors of data should be normally distributed, (b). Equal variances between treatments (homogeneity of variances, and homoscedasticity), and (3). Independence of samples (each sample is randomly selected and independent). To say the least about you clearly ruling out that I am wrong, may you PLEASE read as much as you can on ANOVA assumptions. Thank you.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running