I am not sure what you mean by "two sets of the same variable," but if your comparison is between two subgroups on the same variable (e.g., men and women), then you can use a t-Test, which is the two-category equivalent of an ANOVA.

Alternatively, if you have the same variable measured twice on the same sample (e.g., before and after some event), then you would use a paired-comparisons t-Test.

I think before answering the question, it would be useful to know what you mean by "how can I find a significant mean difference," and say a bit more about the variable (including what the scatterplot looks like and the distribution of the differences). If you don't know what these mean, I would suggest teaming up with someone at your university with appropriate background, since producing statistical output is the easy task, properly interpreting it requires more knowledge.

Use multiple comparison test. Two-way ANOVA works fine.

Two-way ANOVA

David L Morgan Daniel Wright Musa Saheed Sana Noreen Thank you for your responses! I will explain my variables a little better. Let's say I measured a component for an output current. I measured a batch of this component last week, and I measured a batch of this component this week. I have a slight suspicion that these batches are different, but not by chance. I suspect some part of the process that is used to make these components was slightly altered between last week and this week. That's when I googled and I found out I could do a one-way ANOVA test. And then I googled some more and saw that a lot of explanations for the one-way ANOVA test require a independent and dependent variable. My study tries to use two batches of the same (dependant) variable.

What I am really trying to find out is, are the differences in the current of these batches by chance or not?

WELCOME

U.w.

Kindly note, what ever you are measuring must actually have the dependent and independent variables. You are actually measuring component of an output current against what? What is ur X and Y axis? I mean for the week 1 measurement and week 2 measurement.

Jenelysc Martinez Paez , are the measurements independent, or is this essential compare 1 batch with 1 batch, so an n=2?

Because you measured two separate batches, you should use an ordinary t-Test.

Since you have received a variety of advice here, let me clarify a few things.

The independent variables is your predictor, in this case the difference between the two batches. You are getting advice to use a t-Test or ANOVA because you have a categorical predictor variable, and the t-Test is the special case of ANOVA when you only have two categories in your predictor variable.

Your dependent variable is the outcome variable you are trying to predict, i.e., whatever it is that differs between your two batches. For ANOVA or a t-Test, this should be a continuous variable.

Based on your secondary response, this might help clarify your independent (IV) and dependent variables (DV).

The IV here is whether the measurement was taken in week 1 or week 2 (binary/categorical variable)

the DV here is the component for the output current (Continuous variable)

Use a t test to see whether the mean of the DV measured for each IV are statistically different from each other. (An ANOVA and a regression can also technically do this, but keep it simple and use the t test)

Dear friend

Please use the MANOVA to find it.

Amazing, we still dont know the design and yet we have a mingle-mangle of answers: ANOVA, MANOVA, two-way ANOVA..... doesnt seem correct so far.

If I've understood correctly you have the same component that generates numeric outputs of the same kind (i.e. like you are measuring 10 blood pressure from an individual in the same week) and then, after some time you are measuring exactly the same kind of outputs. We are assuming that averaging the first 10 measurements and then the second series of measurements makes sense.

If all these things are true you can do a paired t test comparison between means.

I should add that, if the variance between output of the same set is too high, the test could results not significant event if there is a visible difference in the means.

Since you have two sets of the same variable, you should consider the following:

1 whether the observations in the sets are independent or in some way associated with each other.

2 whether the observations in the sets are drawn from a normally distributed population with a known standard deviation.

If the observations in each sets are independent and drawn from a normally distributed population with a known standard deviation and the sample size (n>30) then you use z-test statistics for your analysis, otherwise you resort to t- test statistics.

Rainer Duesing I agree that this tangle of answers is distressing. So, let's be clear: the answer is a simple t-Test.

2 x 2 ANOVA is wrong because it requires two independent variables

MANOVA is wrong because it requires two or more dependent variables

ANOVA and regression are not technically wrong but they are equal to the t-Test

Rainer Duesing and David L Morgan , I think the mangle is in part the typical misleading answers that RG (I think) encourages by not being able to vote down wrong answers, and not enough information from the questioner to provide an answer. I am waiting to hear if the data are independent from each other within the batch and what she means by find a significant difference (which sounds like how to p-fish, or she might be using "significantly" as a measure of the size/importance of the effect, so might want an interval estimate of some sort ... I don't know). So my view is wait for the questioner, and if they don't want to share these and I am sure other people have questions, then it should just stay unanswered. One of the problems with many RG questions is that very little information is provided, and if some one provides an answer without worrying about this information, this can be misleading because others might think this is unimportant.

Before you applying statistics method, you should make it clear what your hypothesis is, that is what factors cause you think the defference of the dependent variables.

As you mentioned you have 2 sets of data with same variables, then I would guess your dataset is not longitudinal (not repeated measure. but I am not sure, and you should figure it out, it is important).

So I would explain assuming your dataset is crosssectional, for example the 2 dataset were taken in different locations, different races..... if you think your dependent variables are affected by factors mentioned above(location/ race), and you can directly use t-test, one-way anova, two sample non-parametric test (based on your dataset's properties) ...... pls remember that this depends on "no confunding factors" in your datasets.

for more, I would recommend you to take a look at a serials of tutorials :

https://www.youtube.com/watch?v=nk2CQITm_eo&list=PLblh5JKOoLUIzaEkCLIUxQFjPIlapw8nU&index=2

I guess you can learn the statistics basics from above.