# Comparing more than two means of continuous variables

Cross Validated Asked by kapetantuka on August 28, 2020

I have a small sample of 11 people and their results on 5 intelligence test subscales. My hypotheses are that subscale’s 1 mean will be significantly different than subscales’ 3, 4 and 5 means and that subscale’s 2 mean will be different than subscales’ 3, 4 and 5 means. Levene’s test says variances of subscales are equal (p>0.05).

What is the best way to test these hypotheses, which test and how to do it in SPSS? Is there a need for a e.g. Bonferroni correction?

You could use ANOVA, which will tell whether there is a difference between any of the means. If there isn't then you are done. If there is, then you proceed to figure out which means are different. Read up an ANOVA, and post ANOVA.

Answered by user293102 on August 28, 2020

The easiest option will be a linear mixed model, if you are willing to use it.
Put all of your 5 independent variables in one column, create two new categorical variables, one for test id and one for individual id.
Here is how you would do it in R (don't know SPSS):

> dat=data.frame(
>   "score"=unlist(mtcars[,c("disp","hp","drat","wt","qsec")]),
>   "test"=factor(rep(c("disp","hp","drat","wt","qsec"),each=nrow(mtcars))),
>   "individual"=factor(rep(1:nrow(mtcars),5))
> )
>
> library(nlme)
> summary(lme(score~test,random=~1|individual,data=dat))

Fixed effects: score ~ test
Value Std.Error  DF    t-value p-value
(Intercept)  230.72188  11.19880 124  20.602372       0
testdrat    -227.12531  14.47111 124 -15.695083       0
testhp       -84.03437  14.47111 124  -5.807043       0
testqsec    -212.87313  14.47111 124 -14.710212       0
testwt      -227.50462  14.47111 124 -15.721295       0


Answered by user2974951 on August 28, 2020

## Related Questions

### ANOVA complete block design, more units per block than treatments

1  Asked on December 15, 2021

### How can proper scoring rules optimize the probabilistic prediction compared to improper scoring rules?

1  Asked on December 15, 2021

### Calculating the variance of dice rolls?

3  Asked on December 15, 2021

### Multiple comparisons for not normal and heterogeneous data

2  Asked on December 15, 2021 by ana-hernandez

### R-squared is equal to 81% means what?

2  Asked on December 15, 2021 by f-c-akhi

### Time series forecasting: from ARIMA to LSTM

3  Asked on December 13, 2021

### Is it bad to have error bars constructed with standard deviation that spans to the negative scale while the variable itself shouldn’t be negative?

1  Asked on December 13, 2021

### Can I exclude outliers when calculating mean or standard deviation (small-sample)?

1  Asked on December 13, 2021

### Estimating kappa of von Mises distribution

3  Asked on December 13, 2021 by swiss-army-man

### Proper way to combine conditional probability distributions of the same random variable conditioned on a discrete variable ? (based on assumptions)

3  Asked on December 13, 2021 by brainpermafrost

### Von Mises distribution to detect outliers

1  Asked on December 13, 2021 by velvetshelter

### What transformations preserve the von Mises distribution?

1  Asked on December 13, 2021

### what is the name of distribution similar to von mises distribution

0  Asked on December 13, 2021

### Count Panel Data Event Study

1  Asked on December 13, 2021 by econstat

### Expectations of cosine under von Mises distribution

1  Asked on December 13, 2021

### Significance Levels, Confidence Intervals and P-Values

1  Asked on December 13, 2021

### Backpropagation on a convolutional layer

1  Asked on December 13, 2021

### Mixed Models: How to derive Henderson’s mixed-model equations?

2  Asked on December 13, 2021 by domb

### Working with Time Series data: splitting the dataset and putting the model into production

2  Asked on December 13, 2021 by fernando-camargo

### Changing the reference level for contrasts on glmer (lme4) changes the output in anova

0  Asked on December 13, 2021 by evy