ANOVA - Science method

Results show that these tests are not significant in every personal and functional data even though logically there is difference!

I have gender (male, female)

Position (3 choices)

Experience (4 choices)

How to treat this and did i use the wrong anova test?

Wir haben folgendes Design:

Gewünscht ist eine Auswertung, wie sich die Hinweistexte auf die Akzeptanz auswirken unter der Moderation von Risiko und Kreativität des Szenarios.

Der Faktor Hinweistext ist offenbar between subject mit drei Ausprägungen. Aber wie gehe ich mit den Faktoren Risiko und Kreativität um? Ist das eine Messwiederholung (ANOVA)? Wie berechet sich die benötigte Stichprobengröße?

Ich würde mich über sachdienliche Hinweise freuen.

I can't really wrap my head around this and maybe somebody can help me.

I conducted an experimental study with a 2x2 between-subjects design (n = 50 per condition). I have the following variables:

I first ran the base ANOVA model with the following terms:

Nothing was significant. Out of curiosity, I included the moderator by adding the following terms in addition to the above:

What happened is that Factor 1 * Factor 2 became significant (after being non-significant in the base model) as well as the three-way interaction Moderator * Factor 1 * Factor 2.

So the 2 issues I am struggling with are the following:

For the two-way interaction, I guess that I would report the marginal means of the DV in the 4 conditions and plot them with lines. This option is provided by SPSS.

But for the three-way interaction, if I understand correctly, the above two-way interaction between experimental conditions is not uniform, but differs across various levels of the metric moderator (e.g., the two-way interaction might be reinforced/attenuated/reversed/non-significant at certain values of the moderator). But SPSS doesn't seem to offer an option to plot this in the ANOVA menu.

I know that the PROCESS macro by Hayes features conditional effects at -1 SD, mean, and +1 SD values of the moderator as well as Johnson-Neyman areas of significance. So I guess I would need something in this direction. Or am I mixing up things here?

Hey,

I am wondering which tests I can do to analyze my results and if there is a better way to analyze my results. If I have multiple treatments in columns and I have just one donor in a row, should I use an unpaired ANOVA test with Tukey's post-hoc analysis to compare my treatments to one another or can I use an unpaired student t-test and then manually compare each treatment alone to my control group. Also, if I have multiple treatments but then I have two donors, can I perform a paired ANOVA or is it ok to do a student's t-test and compare each treatment individually to my untreated samples? Which test in both cases is more reliable, and should I assume normal distribution and sphericity? I have no idea about specific things and I am wondering which one is the best to present in my thesis. Besides, If I do a t-test and ANOVA on the same data, some treatments are significant in one test but not in the other, what is the reason behind that? Furthermore, why do we perform statistics on raw data not normalized data?

Thank you, your inputs are well appreciated

Software procedures:

Make ANOVA 1 Analysis data with R software.

Find out "P" and "F" with values during an ANOVA 1 Analysis of experimental data.

I want to research on the moderating role of anxiety trait levels (continuous) in the effects of type of reinstatement (independent variable, categorical) on fear avoidance behaviors. I cannot seem to find how to work with a categorical independent variable in a moderation analysis on SPSS and I was wondering whether I need to create dummy variables and run the typical moderation analysis or whether I should conduct a factorial ANOVA.

Thank you!

Using a questionnaire, I analyse the affective well-being after three different dance lessons at three points in time (before, during, after) in the same subjects. Background: the social interaction is changed in each of the three dance lessons, while movements, sequences, etc. are kept the same.

On the one hand, I would like to investigate whether well-being generally increases as a result of dancing, as well as whether well-being changes differently depending on the dance lesson content (social interactions). For this I would use a repeated measures ANOVA.

I would also like to investigate whether the experience of competence surveyed has an influence on the change in well-being (co-variate). I have read that an ANCOVA obviously does not work with a repeated measures ANOVA. Which statistical analysis should I use instead to analyse the influence of a co-variate (or moderator)? Thank you very much for your help

Translated with DeepL.com (free version)

Hi everyone,

I have data available for 4 years. To Compare weight-loss outcomes between the four medication groups for whole years, I will apply one-way ANOVA.

but I will be asked to compare weight-loss outcomes between the four medication groups for each year of the study. My question is,

should I apply one-way ANOVA 4 times as I have four years?

Thanks

I wanted to ask for clarification on statistical approaches for a classic 2x2 pre-post design where I measure an outcome in a pre-intervention and post-intervention phase in two distinct groups.

Typically, what we want to investigate in these cases is the interaction, which tells us if the changes over time in a specific outcome differ between the two groups. To do this, we have three approaches available: Repeated Measures ANOVA (RM ANOVA), ANCOVA, and Linear Mixed Models (LMM).

RM ANOVA: It allows us to study the interaction, as well as the main effects of the between-subject factor (group) and the within-subject factor (time). So, besides telling us if there's a different change over time for the two groups, RM ANOVA also informs us if there's a general change over time (disregarding groups) and if there's a general difference between the two groups (disregarding time). However, the issue with RM ANOVA is that it doesn't allow us to covary the score of the outcome in the pre-intervention phase. This means that any significant pre-intervention difference between the two groups could influence the results and make their interpretation much less robust. The solution in this case would be to verify (with an independent samples t-test) that there are no pre-intervention differences between the two groups in the outcome.

ANCOVA with the pre-intervention outcome score as a covariate: It allows us to see if there are differences between the two groups post-intervention. It doesn't tell us if there's a different change over time between the two groups, if there's a general change over time, or if there's a general difference between the two groups. It doesn't evaluate the interaction, the main effect of time, or the main effect of group, unlike RM ANOVA. The advantage of ANCOVA is that it enables us to study the differences between the two groups ONLY in the post-intervention phase, accounting for pre-intervention differences in the outcome. This means that this model works well even in cases of large outcome differences between the two groups in the pre-intervention phase.

Therefore, RM ANOVA and ANCOVA provide quite different information. If there are differences in the outcome in the pre-intervention phase, it's advisable to use ANCOVA (less informative, as it considers fewer effects); otherwise, it's better to use RM ANOVA (more informative, as it considers more effects). Right?

The solution to all of this is LMM because, in addition to considering all the effects of RM ANOVA (interaction + the two main effects), it also models any differences in the outcome pre-intervention. So, is it always better to use an LMM, right? Is everything correct, or am I missing something?

If the Box's Test yields a significant p-value (p < .001), indicating unequal covariance matrices of (X) across (3 variables) between (2 groups), it raises concerns about the assumption of homogeneity of covariance matrices. (ratio between 2 groups 1.21) In such cases, should we still rely on the results of the multivariate test, or should we consider applying Mauchly's Sphericity Test?..

I have performed a fairly standard kinetics experiment, in which I investigated the stability of various oligonucleotides in an enzyme solution, by incubating the oligonucleotides with the enzyme solution over the course of 1 hour. I took aliquots at 7 time points, added a stopping buffer and performed a denaturing urea PAGE electrophoresis.

I expressed stability as a general proportion of (top band)/(all fragments+top band).

I performed a two-way ANOVA (edit. RM ANOVA) test in GraphPad Prism (the factors being time and the different oligos) and a Dunnett multiple comparisons test (as the first (0) time point was being used as a control). Following the recommendations on the GraphPad website I used a Geisser-Greenhouse approximation, as the sphericity assumption was violated (the epsilon estimate was equal to 0,2855).

I have doubts that I have perfomed the analysis correctly as the statistical significance between the control and some of the points seems too small (*) between points that visually are miles apart and which don't have large SD's. Conversely it gives the same significance to points very close to a control (though in these cases the SD's are extremely minor, so this part makes sense to me).

I have read that the GG approximation can underestimate p-values. I can't just assume sphericity regardless, as it overestimates significance and puts (****) pretty much everywhere. Hence I assume performing a Mauchly test on some other software would be redundant at this point, its seems very clear that a correction is required.

Should I use some other p-value correction? The epsilon estimate seems too small for Huynh and Feldt, and, from what I can tell, noboby uses lower bound estimates anymore. Perhaps take I should take a different approach to analysis altogether (possibly not on GraphPad)?

Any suggestions will be appreciated.

Thanks in advance!

1. Can I do one way ANOVA to know significant difference between my 3 groups on individual Likert items( each measured on 5 point scale). These Likert items are part of my unobserved latent variable. I have read that Likert items measured on atleast 5 point scale can be treated as interval / continuous data. Therefore I don't want to analyse it using non parametric Kruskal wallis test. 2. Am I correct in my approach?

In a clinical study, after applying the clustering method, several classes are identified. We perform ANOVA or Kruskal Wallis for between-class comparison. The class has a different sample size (unequal). I hope that is fine. For a post hoc test, it has been suggested that effect sizes are better than the p-value. Can you please suggest what type of effect size is expected? If there is any other method, please let us know.

1. Is Smartpls the only way to deal with formative constructs?

If we have formative constructs, cant we use SPSS and AMOS? If not, that means that all the studies done on SPSS and AMOS are having only reflective constructs (having similar meaning items)?

2. What to do if i have formative constructs, but want to do EFA & CFA?. Can I do ANOVA, MANOVA and other multivariate tests on formative constructs?

Can I do ANOVA, MANOVA & Multiple regression with formative constructs?

If I do ANOVA after calculating the MEAN (which will form my continuous dependent variable) of all 6 items ( MOT 1, MOT 2, MOT 3, MOT 4, MOT 5, MOT 6), then how can i find out difference between my three groups (independent categorical variable), on the basis of any specific motivation item out of these: ( MOT 1, MOT 2, MOT 3, MOT 4, MOT 5, MOT 6) ?

Please suggest the right way to do it. I know ANOVA can give me significant difference between the three groups with post hoc analysis, but how can I ascertain the difference on a specific item level, if I am transforming the 6items to a mean value? Is it better to do MANOVA or apply any other technique for this purpose??

I am running a two-way mixed anova analysis with one IV is a between-subjects variable and another is a within-subjects variable (each participant was measured six times) and find a significant interaction effect. What should I do to interpret the interaction? I have find similar papers and it says that we should calculate the slope for each participant as shown in the picture, but how?...I am really confused...

I have two groups (A and B) each going through 5 repeated measures. I want to know how I can determine eta squared so that I can use it to calculate an a-priori sample size.

Thanks in advance

Dear experts,

Now I am writing my PhD research proposal, and I have a question regarding variables.

You see, I have two dependent variables in my research and one independent variable. My question is, is it ok to have two DVs in my research? If so, what kind of methodology should I use to justify it? (I think maybe AnoVa is cool?)

Thank you. I am looking forward to hearing from you.

My research scenario is as follows:

I have ppts in three different intervention groups (independent variable one, between groups)

I have two time points (independent variable two, within groups) where scores on psychometrics will be recorded first at month 0 and month 12

My dependent variable is treatment change from month 0 to month 12

Am I right in thinking that the appropriate statistical analysis would be a mixed factorial ANOVA? As I have two IV's, one between, one within, and want to measure their effect both together and separately on my DV.

In my study, I have a within-subject independent variable (the DV was measured twice for each participant), and I am also interested in examining the effect of participants’ demographic characteristics (between-subject) on the dependent variable. Therefore, I am running a mixed model ANOVA using SPSS.

Here are my questions:

Thank you in advance for your advice.

For my research, I have to compare multiple groups (between) and some factors within those groups. I will try to explain it as good as possible.

Independent variable: group. There are two groups: experts (N =13) and non-experts (N=13). These two groups evaluated 6 different robot gestures, representing emotions. They evaluated these via 6 videos. The 6 emotions represented were: anger, surprise, happiness, fear, sadness and disgust.

They received this statement for every emotion: it is feasible for a child with the Autism Spectrum Disorder to recognize this emotion.

They answered via a 7-points Likert item (1 = strongly disagree to 7 = strongly agree).

I have to compare the feasibility of the gestures within the groups (within) and for every emotion between the groups (between). I was thinking about a mixed ANOVA. Unfortunately, some data is not normally distributed (I used Shapiro-Wilk in SPSS for this because of my sample size). The emotions anger, sadness, disgust (only for non-experts) and fear (only for non-experts) are not normally distributed. I tried to fix this with transforming (log) but this did not work.

I saw something about the Generalized Estimating Equations (GEE) and the Friedman test (non-parametric tests) but I am still not really sure what to do.

I hope someone can give me some tips.

I have the survey among three groups of health professionals to answer the survey.

The survey is a Likert scale for seven different treatment options across different disease conditions (6 conditions).

The goal is to assess if there's a statistically significant difference between the mean scores of these different groups of health professionals for various treatment options.

I have 6 different disease condition (or 6 different question) each question with 7 different treatment option.

I am thinking to run ANOVA works here, but if I want to run ANOVA, then will need to run that 42 times. Because (6*7=42)

Then Applying corrections like the Bonferroni correction to adjust for the increased risk of Type I errors (false positives) when conducting multiple statistical tests, such as multiple ANOVAs. But still will need to apply ANOVA 42 times, is that correct method to do for this? Is there Any better method?

if I want to calculate the overall mean score for one specific treatment option across all disease conditions, then will run the ANOVA just seven times.

If I want to go with the first approach, what method do you suggest?

Am I correct to apply anova?

Thanks

Hello,

I conducted a lexical decision task using two distinct talkers (within subjects), to determine if three different types of particpants would respond differently (between subjects). Groups were counterbalanced as words 1-20 were spoken by a male talker, and words 21-40 were spoken by a female talker for half of participants (and vice versa).

I was guided to perform three separate ANOVAs on

-reaction time (ms)

-accuracy (0-1, means hovering around 93%)

-recall (0-20, means hovering around 2)

I have performed a log transformation on the reaction times for analysis. What transformations would be recommend for accuracy percentages and for the number of words recalled?

Hi everyone, I would appreciate to tell me you suggestion:

I’m going to compare just two independent group. The outcome variables are continuous -one of them has 5 subclass and another one has 4 subclass-

I am thinking to apply running separate t-test for each subclass of these outcome variables (to compare total score and subclass score).

Then will consider adjusting for multiple comparisons to control the Type I error rate by Bonferroni correction.

But as I have one confounding variable, so will ignore T-test and apply ANCOVA?

Am I correct, ANCOVA is the best for my goal?

Dear Community,

I have recently used Aligned Rank Transform (ART) ANOVA for my dataset, due to its non-normal distribution and heteroscedasticity. As you might know, ART ANOVA is performed on aligned ranks rather than the original data.

I need to visually represent the main and significant effects of the factors studied. However, I am uncertain about the most appropriate method for plotting these effects. My question is:

Should I plot the differences in levels of a factor (e.g., showing mean values and error bars) using the aligned ranks generated from ART ANOVA? Is there a recommended best practice for visually summarizing the ART ANOVA results to illustrate the significant effects clearly and accurately?

I would greatly appreciate any insights or references to literature that could guide the appropriate visualization of ART ANOVA results.

Thank you

I performed swelling experiments three times for each of five different samples, recording measurements every hour for 24 hours. How do I apply ANOVA to analyze this data?

I can select one of these method:

Thank you in advance

Greetings, I am final year BSc.(H) Psychology. I used two-way, repeated measures ANOVA for a 2x2(time x intervention) pre and post-test model for my study. Now, within-subject effects of time and intervention are significant. However, the significance level of time*intervention is (0.67). How do I interpret these results?

We are a group of students working on a grant poposal as part of our veterinary degree but are struggling with choosing the right statistical test.

Our project compares 2 premedication protocols in horses undergoing GA (60 horses in each group) where multiple parameters will be measured during the surgery (ie heart rate, blood pressure among other) every 10min. We are expecting the surgeries to last a maximum time of 120mins, which will create a lot of data to analyse.

Would a 2 way anova test be more accurate or should we consider running multiple T test (for each paramters measured) ? Our goal is to determine if one premedication protocol is better at safeguarding cardiovascular properties of enrolled animals vs the other premedication protocol.

Thank you any help !

Q1

We have animal behavior scores of 4 group, Normal+ctrl virus, Normal+down-regulation virus, Model+ctrl virus and Model+down-regulation virus. It has two factors(Independent Variable): Model and virus. Editors suggested we use two way ANOVA to analyze, and now we obtained main effects of Model (F(1, 56)=201.18, P<0.0001) and virus (F(1, 56)=11.17, P=0.00427), as well as Model × virus interactions (F(1, 56)=16.13, P=0.0007).

If we should continue to calculate? For example, Model+ctrl virus vs. Model+down-regulation virus. We want to confirm the role of virus in Model animals.

Q2

Next, we used chemical drug to treat the Model animals and Normal animal. It has 4 drug concentration. Should we still use two way ANOVA to analyze the behavior scores? We want to know the role of different drug concentration in Model animals. And what do we do after two way ANOVA？

Thanks very very much!!!

As a Computer Science student inexperienced in statistics, I'm looking for some advice on selecting the appropriate statistical test for my dataset.

My data, derived from brain scans, is structured into columns: subject, channels, freqbands, measures, value, and group. It involves recording multiple channels (electrodes) per patient, dividing the signal into various frequency bands (freqbands), and calculating measures like Shannon entropy for each. So each signal gets broken down to one data point. This results in 1425 data points per subject (19 channels x 5 freqbands x 15 measures), totalling around 170 subjects.

I aim to determine if there's a significant difference in values (linked to specific channel, freqband, and measure combinations) between two groups. Additionally, I'm interested in identifying any significant differences at the channel, measure or freqband level.

What would be a suitable statistical test for this scenario?

Thanks in advance for any help!

SPSS doesn't offer Aligned Rank sum test in version 17-21? If one were to report a multiple time data where continuous variable werent normal, would normalizing and using anova not amount to simple split plot anova? what of if the data is ranked originally, how can one handle it possible with these older SPSS versions?

What are the possible ways of rectifying a lack of fit test showing up as significant. Context: Optimization of lignocellulosic biomass acid hydrolysis (dilute acid) mediated by nanoparticles

I performed the mixed model ANOVA analyses and repeated measures analyses.

I need answer for this question:

What is considered small, medium or large effect size and cite reference as needed? Also explain, how should I calculate or examine effect size. I used SPSS.

I am analysing a set of data which have both within-subject and between-subject variables. However, since my assigned task is just simply focusing on part of this data set, I am not interested in examining the within-subject variable.

The original study is to examine how the branding of the product (independent variable) affects participants' happiness (dependent variable), and all participants are required to try both brand A and brand B product. However, extending from this original study, I was asked to examine how participants' demographic may actually affect their happiness instead.

How can I solve these problems using SPSS? Can anyone help? Thank you

Hello, I am looking for an advice regarding my experimental design and what statistics I should do.

My experimental design:

I have 3 cell lines (A, B, C) and from each I do 3-4 differentiations into cardiac myocytes (A1-A4, B1-B4, C1-C3). Then each differentiation is treated with the same 3 concentrations of a chemical (T1, T2, T3). And for each treatment, I measure the calcium concentration (y). So I have one continuous dependent variable (y), two categorical independent variables (x1 for cell line and x2 for treatment) and a random error which I want to correct for (e for differentiation).

Many thanks in advance !

One of the many assumptions in ANOVA is that our data follows a normal distribution. Usually in biology experiments are carried out in triplicates and the avalilable data is very small, not > 30 which I think is standard sample size for assessing normality. Any one who is familiar with use of statistics in phytochemistry, microbiology, pathology or any relavant field kindly answer the following questions.

1. Is it necessary to carry out tests of normality such as Shapiro-Wilk normality test to confirm if our data follows normal distribution?

2. With such a small sample size is it possible to carryout these tests?

3. Is it true that test for normality is unnecessary in biological experiments?

4. Can I safely assume that my data will follow normal distribution wiithout any of these tests?

5. Which statistical software is best for a beginner?

I would like to utilise the correct regression equation for conducting Objective Optimisations using MATLAB's Optimisation Tool.

When using Design Expert, I'm presented with the Actual factors or Coded factors for the regression equation. However, with the Actual Factors, I'm presented with multiple regression equations since one of my input factors was a categoric value. In this categoric value, the factors were, Linear, Triangular, Hexagonal and Gyroid. As a result, I'm unsure which Regression equation to utilise from the actual factors image.

Otherwise, should I utilise the single regression equation which incorporates all of them? I feel like I'm answering my own question and I really should be using the Coded Factors for the regression equation, but I would like some confirmation.

I used one of the regression equations under "Actual Factors" where Linear is seen, but I fear that this did not incorporate all of the information from the experiment. So any advice would be most appreciated.

Most appreciated!

I am working on plant breeding and need to compare many genotypes in the breeding process. What is the best post-hoc test to compare differences between 10, 20 or 30 genotypes?

1. As I navigate through the complexities of miRNA expression analysis using qPCR, my study involves a panel of 20 plates, each containing 96 wells—18 dedicated to patient samples and 2 for controls.

In categorizing my samples into newly diagnosed lymphoma/leukemia, those in remission, and those exhibiting resistance to treatment, I aim to calculate essential parameters such as ΔCT, ΔΔCT, and fold change. Is it appropriate to determine the average CT values for each subgroup to obtain a representative measure of miRNA expression within these distinct clinical states? keep in mind each group contain 3 samples from different individuals. Also is't acceptable to take multiple reference genes and take average of them for normalization?

Additionally, I encounter undetermined CT values; what would be the most judicious approach to handle these values? Should I assign them as 35?

Similarly, CT values exceeding 35 pose a challenge. How can I establish thresholds for further analysis in order to maintain data accuracy? because I cant delete any thing from genes as they are panel of miRNA

Moving into the statistical analysis phase, which methods, like ANOVA or t-tests, would be most effective in discerning significant differences between the categories of my samples?

Finally, in presenting my findings, how can I ensure clarity and transparency, incorporating well-organized tables and figures to visually convey the intricate dynamics of miRNA expression?

Moreover, how should I adeptly discuss the biological implications of my results while addressing potential limitations in my study?"

It's worth noting that the source of my samples is plasma, and they are derived from patients with hematological malignancies at various stages of the disease. Furthermore, each sample has been processed only once without any technical repeats.

Hi everyone! I tried to perform a classic One Way Anova with the package GAD in R, followed by a SNK test, which I always used, but it didn't work with this dataset, and I got the same error for both tests, which is the following:

"Error in if (colnames(tm.class)[j] == "fixed") tm.final[i, j] = 0 :

missing value where TRUE/FALSE needed"

I understand there is something that gives NA values in my datatset but I do not know how to fix it. There are no NA values in the dataset as itself. Here is the dataset:

temp Filtr_eff

gradi19 11.33

gradi19 15.90

gradi19 10.54

gradi26 11.01

gradi26 -1.33

gradi26 9.80

gradi30 -49.77

gradi30 -42.05

gradi30 -32.03

So, I have three different levels of the factor temp (gradi19, gradi26 and gradi30) and my variable is Filtr_eff. I also already set the factor as fixed.

Please help me, how do I fix the error? I could do the Anova with another package (library car worked for example with this dataset) and I could do tukey instead of SNK, but I want to understand why I got this error since it never happened to me..thanks!

PS: I attached the R and txt files

Mean, median, Range, one way ANOVA, Pearson correlation, spatial distribution with scatter plot, species richness and Shannons index are applied to the data. Is there any other tools are to be included for the data?

I am trying to report the results of a study that employs a 2 (between subjects: tx vs ctrl) X 4 (within subject: time :T0,T1, T2 and T3) mixed factorial ANOVA. I have analyzed the data using both SPSS and JAMOVI.

Personally, I find the JAMOVI results cleaner and the graphs of estimated marginal means much better because SPSS is not graphing with error bars at all. and JAMOVI tables are much simpler.

Question is: why does SPSS provide a multivariate table here for an analysis that is essentially univariate (One continuous DV), then follows it up with separate tables for within and between subjects factors?

Second, from the JAMOVI output, I have been using the interaction term posthoc testing to breakdown the significant interaction- I wanted to verify this is essentially doing the same thing as a simple main effects analysis. conceptually this is testing each level of each factor within each level of the other factor- so this should be okay, right?

I am doing a two way ANOVA in spss 27. However, the homogeneity of variances assumption is violated. Thus, I considered doing a Welch's factorial Anova but I cannot seem to find the way to do this in spss. How can I do this? And is there anything else I can do in spss, as it is the only program available to me, to solve this problem? I tried transforming my DV with the square root method, it did not work.

Thank you

Hello everyone,

I am performing multiple comparisons at the same time (post hoc tests), but among all the possible p value adjustments available (Bonferroni, Holm, Hochberg, Sidak, Bonferroni-Sidak, Benjamini-Hochberg, Benjamini-Yekutieli, Hommel, Tukey, etc.), I don't know which one to choose... And I want to be statistically correct for the comparisons that I am making in my experiment.

In my experiment, there are 4 groups (let say A, B, C, D), but I want to compare A vs B, and C vs D. That's all. So, after performing wilcoxon tests, the non-parametrical equivalent of a t test (because I have such a low amount of repeat per group (n=6) + non-normality for some groups), for A vs B, and C vs D, I don't know which p value adjustment should be performed here.

I would like to understand 1. which adjustment I should perform here. 2. how to decide which test I should perform for any other analysis (what is the reasoning).

Thanks in advance for your response,

Dear all,

I am currently exploring Latent Profile Analysis (LPA) and have noticed a trend in some of the papers I've read. After completing LPA to categorize subjects into groups, these studies often proceed directly to further analyses such as ANOVA, without conducting tests for normality or homogeneity of variance. I am curious if this approach is considered methodologically sound.

For context, in my own research, I have divided participants into three groups and performed the Shapiro-Wilk test, which indicated that none of the groups conform to a normal distribution. Despite this, I am interested in conducting an analysis of differences between groups. Consequently, I opted for the Kruskal-Wallis test. Is this approach justified in this context?

I appreciate any insights or advice on this matter.

Using GraphPad Prism, when I carry out a 2-way ANOVA with an interaction term, data between two groups is not significant. The post-hoc tests shows non significant interaction between my column and row data. So I did redid the analysis using a 2-way ANOVA without an interaction term, and now I get significance between two groups of data, which is more like what I think happens biologically speaking. But why does the p values change when I use an interaction term versus when I don't use an interaction term, given that the interaction term is not significant anyway? Which is the correct way to analyse the data?

Dear professors and readers,

I have data that is measured every day in two groups. The first group is shift workers (2 weeks data which divided into the first week is day shift and the second week is night shift). Then, the second group is non shift workers (1 week data which only do day shift).

Workers of the shift work and non shift work are different.

I want to know:

- if there is a group difference between shift work and non shift work

- what is the weakest working shift type condition (day shift or night shift or non shift)

My dependent variables are blood pressure, heart rate, etc (continuous data). I already checked the data distribution and most of them are not normal and also not homogen.

I tried repeated one way Anova for comparing between day shift and night shift. (Because the subjects are same people)

I also tried one way Anova for comparing between day shift - non shift and night shift - non shift. (Subjects between shift and non shift are different)

Other teacher said I can use GLMM (Generalized Linear Mixed Model), but I am still not understand the basic concept of it.

My questions are:

1. Was my statistical analysis correct?

2. Is there other statistical analysis that I can use for comparing those conditions in the same time? I wonder might be there is an interaction or interesting phenomenon between day shift, night shift and non shift.

3. Is GLMM suitable with those conditions?

Thank you very much for your kind help and support.

I set the eta square as my effect size.

There is only the standard of effect size regarding eta square for one-way ANOVA.

What about the two-way ANOVA and three-way ANOVA?

Is there any book or researcher paper I can refer?

I will conduct experimental research with two independent variables and one moderation variable so that the factorial design is 2x2x2; with the ANOVA interaction test, the research topic is accounting. What steps should I take to test the moderation role?

I am hoping someone can advise on the use of the Bonferroni correction. I am struggling to understand how and when to use it.

I have been part of a pilot study evaluating balance in older adults. We used several balance measures (sway area, sway velocity) with different parameters (eyes open/closed, solid/compliant surface), plus a timed functional test, measured on three occasions with the same subjects. I have performed separate repeated measures ANOVAs for each measure and parameter combination (7 in total) using SPSS, alpha set at .05. I understand the need to use the Bonferroni correction with the pairwise comparisons and that SPSS does this. However, I am unsure whether the correction needs to be applied across all the ANOVAs, namely whether I should set the corrected alpha level to 0.05/7 = 0.0071 when determining statistical significance for each individual ANOVA.

I understand there are debates around minimising Type I errors at the expense of Type II errors, but is the approach of using alpha of 0.0071 for each individual ANOVA fundamentally correct?

dear my fellow colleague. I have a question, I have 3 independent variables ( with 3 means) and 1 dependent outcome. I had conduct Anova and reveal significant differences in these 3 (p < 0.001) and also run post hoc for this 3 variables, and i find differences in 1><3 and 2><3. but I also have 3 nominal confounding variable that need to be clear for the effect on the outcome. what kind analyses I can run for?

I have a setting of 4 groups of normally distributed data and unequal sample sizes. Alpha-level is set at 0,05 in all analyses.

To see if there are any differences between the groups I've run (with SPSS) an ANOVA and Tukey HSD post hoc test. SPSS provides also the Welch test for cases with unequal variances (= significant Levene's test with p < 0,05), and with these I've used Games-Howell as a post hoc test.

However, there a few cases in which the Levene's test p > 0,05 and Welch test p < 0,05 (with GH finding significant differences between groups). In some of these cases the omnibus ANOVA p is significant but the Tukey HSD doesn't yield significant differences between the groups.

Welch test can be certainly used instead of ANOVA when the variances are unequal, but can it (with GH post hoc) also be used when the variances are somewhat equal according to the Levene's test?

During work for my undergrad thesis, I've examined and compared gene expression of stress-induced genes in plants challenged with a fungal infection.

I have calculated the relative fold change and wondered: How high does my fold change have to be for it to actually make a difference?

For example:

The log2 RFC between Group A and Group C is 0.63.

According to ANOVA, this difference is significant.

I'm wondering if this difference is enough to change the plants' stress response.

Is there a certain value that I can see as a "threshold" or is mere statistical significance enough to confirm the change in the plant?

Thank you in advance .)

I am conducting a study where 4 groups are measured over 4 different time points (t0, t30, t60 and t90). I am mainly interested in what happens within groups. I am carrying out RM_ANOVA, I got the mean differences between the groups but I want to represent the differences in percentage.

Any idea what is the best method to do this?

Is it the sum or average of all SDs?

Hello,

I am having problems on finding the right statistical test to use, hope you can help me. Essentially, I’m looking at the effect of a drug at reducing multiple sclerosis progression in a EAE mice model. I have two groups (treated and vehicle) and I have neurological scores for each mouse in each group for a period 16 days.

The neurological test are scored on a 5 point scale (0 to 5) where 0= normal and 5= paralysed (similar to a Likert scale), and are conducted daily (for the 16 days). compound is given to the animal daily, and so is the vehicle.

I am comparing the scores for the two groups (between subject design), but I also want to see how the scores change over time (within-subject design) i thought on using the Mixed ANOVA am I correct? However, since the neurological testing is scored as a Likert scale, the data should be non parametric, correct?

What equivalent test can I use on SPSS?

I appreciate your help in advance

Thanks

I have data for two seasons of tamarind genotypes, and I conducted the experiment trial using a Randomized Block Design (RBD). I need to analyze this data to estimate the variability. Can someone kindly provide an exact solution?

Counts data re-distribution.

I ran ANOVA however, the numbers don't make sense. My experiment was based on treating plant tissue culture explants with NaDCC and PPM. After a 30-day period, I had to count the number of aseptic and contaminated tubes from a batch of 10. So if 5 tubes were contaminated it means 50% were aseptic. On the other batches, I had O%-contaminated tubes and 100% clean tubes. Therefore, there's a huge variation between the numbers resulting in my data having a non-normal distribution or skewed distribution. For me to use ANOVA I assume my data must be normally distributed. The tubes were counted and values were converted to percentages. But when plotting graphs in Excel +/-SD extended passed zero to the negative quadrant. I am not familiar with statistics or ways to code, I am completely clueless on how to resolve this issue.

When i do regression analyze, in Model Summary Table, i found Rsquare is very weak like:0,001 or 0.052, and value of sig. in Anova table is greater than 0.05, how can i fix this?

Is there a need to check for Multi-collinearity before the ANOVA and ANCOVA?

Dear all,

as part of my Master thesis, I am conducting a two-way repeated measures ANOVA on data from a social evaluation paradigm. The purpose for running this analysis is to investigate whether social feedback prediction (like, dislike) moderates the impact of social feedback valence (negative, positive, neutral) on participants' self-rated state self-esteem (i.e. whether unexpected social feedback exerts a differential impact on state self-esteem depending on the feedback valence).

Participants undergo several trials in which they are instructed to make predictions as to the social feedback they will receive, after which they receive alleged social feedback on a personality profile. After a random number of trials, participants need to rate their state self-esteem with reference to the previous social evaluations.

The analysis revealed a significant interaction effect between social feedback valence and social feedback prediction on state self-esteem. I have already plotted this interaction effect using a bar chart, but I would like to quantify the mean differences in state self-esteem between levels of feedback prediction for each level of feedback valence such that I can clearly describe the underlying simple effects and, accordingly, interpret the cause of the interaction effect.

Which methods/procedures/strategies would you recommend using to probe the valence x prediction interaction effect?

I hope that I have provided sufficient information.

Thank you very much in advance.

With warm regards,

Marius

Hey everyone,

I have the following situation:

Two groups of mice wildtype (WT) and knockout (KO) in each group I have lets say 10 mice. I isolate cells and treat the cells from each mouse with a basal or insulin condition (these treatments are done on the same cell suspension) and I measure glucose uptake as an outcome.

From previous data, I know that with increasing glucose uptake, which is induced by the insulin treatment the mean of glucose uptake increases and with increase in mean my variance increases. I choose to log-transform the data and I get approximately normal distribution and equality of variances. Is this reasonable to do?

Then I would like to carry out a statistical analysis. From my n=10 sample I can say very little about the distribution and assumptions of the model I want to use and I would therefore base my assumption about this sample on my previously observed findings of about 200 observations. Is this okay to do? Assuming the assumptions of a parametric test are fulfilled, how would you analyse this data?

As basal and insulin are from the same cell suspion, I assumed these to be paired/dependent. While the other measurement, so between WT and KO should be independent. In each group, so WT Bas, WT Ins, KO Bas, KO Ins I have 10 biological replicates. I was thinking about a repeated measures ANOVA. Could someone help here?

Hello All,

So I am doing a two-way ANOVA. The within-subject factor is Ethnicity (3 levels). The between-subject factor is Gender (2 levels). I have been writing it interchangeably as either "3 x 2 mixed ANOVA" or "2 x 3 mixed ANOVA". Of course, it doesn't change the total number of conditions, but in terms of clarity and understanding, is there a standard way of writing the factors? Do you state the within-subject factor first (i.e. 3 x 2) or between-subject factor first (i.e. 2 x 3).

Thank you so much.

I know I can use a paired t-test or repeated measures ANOVA, but I want to run a series of paired t-tests--is it possible to do them all at the same time? I want to see if there is a greater difference for some pre/post tests than others.

Hello, I want to determine how much statistical power my study has in finding a significant 4-way interaction. The study uses a 2x2x2x2 mixed ANOVA design (1 between and 3 within factors all containing 2 levels each). Is there a tool that could be useful with such a design? Mixed ANOVA with g*power and superpower seems to be limited to 2x2x2 interactions. Perhaps someone can give a guide on a R packages and how to figure out the power.

Hi there,

I have qPCR results from 6 target genes and a housekeeping gene. Each different gene had its expression measured after 3 different treatment conditions and a non treatment control.

I have log 2 transformed the fold changes that were the result of normalising the different treatments to the housekeeping gene and non treatment control.

My question is, how does one present this on a graph. Is it necessary to leave a space for the non treatment control, given that all its values will equal to zero (log2(1) = 0)?

Also, does a one way ANOVA work for statistical analysis? But am I correct in saying that performing an ANOVA will only show statistical difference between different treatments, but not whether the treatment decreases or increases the expression?

I'm hoping to get some advice for my thesis research. I am running an RCT and due to having a large number of outcome measures (18 in total), I have conducted 18 4 (time: T1-T4) x 2 (group: intervention vs. control) mixed ANOVAs. For those with significant interactions, I am now thinking about the best approach to use for follow-up analyses.

Due to the vast number of comparisons that would be needed for post-hoc testing, there is a high risk of Type 1 error, and the required bonferroni correction would lead to a high risk of Type 2 error. It has therefore been suggested to me by an academic colleague that I could choose to not carry out post-hoc tests and, instead of looking at measures of statistical significance, focus on measures of effect size to interpret my results. This is not an approach I am familiar with and I have not yet been able to find examples of this in the literature.

I wondered whether anyone is aware of research where post-hoc testing has not been carried out in studies of this nature? Would be helpful to hear views on whether this could be an acceptable approach to analysis in this context or whether it would be frowned upon!

Thank you in advance.

Hello everyone,

I want to do a comparative analysis with an ANOVA, but my dependent variable is measured with a 5 point Likert scale. I am not sure if I can use ANOVA, as some people point out that it can be used because we treat the variable as continuous, but others recommend against it since it is ordinal.

Could you recommend me articles that justify the use of ANOVA with dependent variables measured with likert scales.

Thank you.

I have a question about what statistical tests I can do in R that would be most suited to analyse my enzymatic assay results.

My data

Dependent variable: Absorbance (its’ a colorimetric assay)

Factor 1: Protein (type of protein, wt and several mutants)

Factor 2: Substrate (several substrates)

For each assay, (e.g. Prot1 with Substr1) I have 3 – 6 data points (repeats), and it is not feasible to obtain more.

An example of my data would be something like fig1.

What I want to do:

1. Test if the different mutations have a different effect on the enzymes’ activity with the different substrates (basically test the dependence of Absorbance on the interaction between Protein and Substrate: Abs~Protein*Substrate)

Visually, if I plot my data on a bar chart (mean +/-SD), it appears to be the case, but I need to verify that what I see is significant.

Normally, I would do this with a two-way ANOVA, however:

my data is not normally distributed (according to Q-Q plot and skewness test, I have over-dispersed residuals (Laplace distribution), without any skew);

the variance is not homogenous (standardized residuals vs fitted values plot shows heteroskedasticity)

What sort of model could I use instead? Is there a way I can transform my data to allow a parametric test (the only transformations I found were against skewness, which I do not have)?

2. Test for each substrate, which mutations make the activity differ significantly from the wt (e.g. whether activity with Substr2 is significantly different for Prot2 and 3 from that for Prot1)

To avoid doing multiple pair-wise comparisons, I would normally do a pairwise.t.test with a Holm family-wise error rate correction.

For non-normally distributed data of non-equal variance, I saw it is recommended to do a Pairwise Wilcoxon rank sum test.

If I group my data by substrate, with one of the substrates (e.g. Substr2) it was normally distributed and of equal variance. I did the pairwise T test and it gave results consistent with what is seen on the graph. However, when I tried a Pairwise Wilcoxon rank sum test (holm correction) it showed no significant difference between any of the Proteins, which makes no sense (e.g. that there is no significant difference between Prot1 and 2, although one has an activity with the substrate and the other doesn’t). So it looks like the non-parametric test may not be powerful enough/ at all useful.

For the other substrates, either the distribution is not normal (again over-dispersed, Laplace distribution, with no skew), or the distribution is normal, but the variance is not equal (heteroskedasticity).

Just to note, one-way ANOVA or Kruskal-Wallis rank sum tests (where applicable) for Absorbance~Protein for each substrate individually, showed there is a significant variation in Absorbance depending on Protein.

What sort of pairwise comparison tests can I do in these cases? Or, again, how can I transform my data to be able to use a pairwise.t.test?

Thank you in advance!

I really appreciate any advice!

I have 4 groups in my study and I want to analyse the effect of treatment in 4 groups at 20 time points. Which test should I chose?

What is the appropriate ANOVA model for the following experimental design: the effect of four different concentrations of compound X on microorganisms? Each X concentration has three jars and three replicates are collected from each pot. Samples are drawn weekly for 18 weeks.

Hello everyone, does anyone know how to calculate the simple size for a 2(gender: male, female)*2(culture: Asian, European)*2(age: children, adult)*2(direction: back, front)*2(position: left, right)*2( condition: confort, non-comfort) repeated ANOVA? We have 6 factors, among these, gender and culture are between-factors, while age, direction, position, and condition are within-factors. I'd appreciate it if someone can help me.

I have two factors: time and intervention. Time has 4 levels (week 0, 4, 8, and 12), while the intervention consists of intervention group vs control group. The dependent variable is a test score. I have done two-way repeated measures ANOVA and found that there was significant time effect and interaction effect between time and intervention on the test score, but there was no significant main effect between intervention vs control. Now, I want to do a post-hoc analysis to see at which time point the score differences between intervention vs control group is significant. How should I perform the post-hoc test on SPSS? Please correct me if any of the steps that I have mentioned above are wrong. Thanks a lot.