peso_in= c(runif(120, 1.5,2.0))
hum= c(runif(120, 40, 70))

#Variedad 
genotipo= gl(2, 60, 120, c('varied1', 'varieda2'))

tiempo= gl(2, 30, 120, c('10', '12'))

df=data.frame(genotipo, tiempo, hum, peso_in )
library(ggplot2)
library(viridis)
## Loading required package: viridisLite
ggplot(data= df, aes( x= genotipo , y=tiempo, color=hum))+
  geom_point(size =15, shape=20)+
  scale_color_viridis(option = "A")

tapply(df$peso_in, df$tiempo, mean)
##       10       12 
## 1.740066 1.740146
tapply(df$peso_in, df$genotipo , mean)
##  varied1 varieda2 
## 1.716191 1.764021
boxplot(df$peso_in ~ df$tiempo)

lattice::bwplot(df$peso_in ~ df$tiempo | df$genotipo)

modav= aov(peso_in ~ tiempo * genotipo , df)
summary(modav)
##                  Df Sum Sq Mean Sq F value Pr(>F)  
## tiempo            1 0.0000 0.00000   0.000 0.9975  
## genotipo          1 0.0686 0.06863   3.628 0.0593 .
## tiempo:genotipo   1 0.0004 0.00039   0.021 0.8854  
## Residuals       116 2.1945 0.01892                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
TukeyHSD(modav, 'tiempo')
##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = peso_in ~ tiempo * genotipo, data = df)
## 
## $tiempo
##               diff         lwr        upr     p adj
## 12-10 8.027455e-05 -0.04965731 0.04981786 0.9974549
res1 = modav$residuals
shapiro.test(res1)
## 
##  Shapiro-Wilk normality test
## 
## data:  res1
## W = 0.9697, p-value = 0.008262
trat=interaction(df$tiempo, df$genotipo)
bartlett.test(res1, trat)
## 
##  Bartlett test of homogeneity of variances
## 
## data:  res1 and trat
## Bartlett's K-squared = 1.0885, df = 3, p-value = 0.7798

Las varianzas de los tratameintos no son diferentes ya que son de 6%

plot(res1)

plot( modav$fitted.values, res1)

Ajustando el modelo

modavb= aov(peso_in ~ tiempo+genotipo, df)
summary(modavb)
##              Df Sum Sq Mean Sq F value Pr(>F)  
## tiempo        1 0.0000 0.00000   0.000 0.9974  
## genotipo      1 0.0686 0.06863   3.658 0.0582 .
## Residuals   117 2.1949 0.01876                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
shapiro.test(modavb$residuals)
## 
##  Shapiro-Wilk normality test
## 
## data:  modavb$residuals
## W = 0.96962, p-value = 0.008126
bartlett.test(modavb$residuals, trat)
## 
##  Bartlett test of homogeneity of variances
## 
## data:  modavb$residuals and trat
## Bartlett's K-squared = 1.0885, df = 3, p-value = 0.7798

Analisis de covarianza

ANCOVA <- aov(df$peso_in ~ df$hum+ df$tiempo)
summary(ANCOVA)
##              Df Sum Sq  Mean Sq F value Pr(>F)
## df$hum        1 0.0006 0.000623   0.032  0.858
## df$tiempo     1 0.0000 0.000004   0.000  0.989
## Residuals   117 2.2629 0.019341
shapiro.test(ANCOVA$residuals)
## 
##  Shapiro-Wilk normality test
## 
## data:  ANCOVA$residuals
## W = 0.965, p-value = 0.003256