Redeem results in r

Question

It is possible to rescue only p-value from a simple linear regression in R?

I have the following code:

reg<-lm(abscissa~ordenada)

summary(reg)

the answer is:

Call:
lm(formula = abscissa ~ ordenada)

Residuals:
    Min      1Q  Median      3Q     Max 
-746.77 -308.91  -62.22  285.36 1236.04 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) 2163.1769   161.4879  13.395  < 2e-16 ***

ordenada       1.7729     0.2735   6.482 8.95e-09 ***

Signif. codes:  0 *** 0.001 ** 0.01 * 0.05 . 0.1 '' 1

Residual standard error: 452.9 on 74 degrees of freedom
Multiple R-squared:  0.3621,    Adjusted R-squared:  0.3535 
F-statistic: 42.01 on 1 and 74 DF,  p-value: 8.951e-09

i need to call only p-value (8.951e-09)

I tried to use the $ operated but unsuccessfully.

Grateful

Answer 1

The p-value does not exist in the summary object. What the summary object offers are the conditions of calculating the p-value, with the statistic F and the two degrees of freedom.

reg <- lm(mpg ~ wt, mtcars)
sumario <- summary(reg)
estatisticas <- sumario$fstatistic
estatisticas
#    value    numdf    dendf 
# 91.37533  1.00000 30.00000 

With this information in hand we can calculate the p-value used the package stats, which already starts loaded in a standard session of R.

pvalor <- stats::pf(estatisticas[1], estatisticas[2], estatisticas[3], lower.tail = FALSE)
#        value 
# 1.293959e-10

If you want the p-value to be nicely formatted, as when the summary print occurs, you can

format.pval(pvalor, 3)
# [1] "1.29e-10"

Answer 2

Despite the calculations of answer by Tomás Barcellos are right, the object sumario already has the p-values:

reg <- lm(mpg ~ wt, mtcars)
sumario <- summary(reg)

sumario$coefficients[, 4]
# (Intercept)           wt 
#8.241799e-19 1.293959e-10

To have only the second p-value, run the above command but this time with the line index.

sumario$coefficients[2, 4]
#[1] 1.293959e-10

See what sumario$coefficients has. To see the object sumario complete the best option is to run str(sumario) and then spend some time reading the output of this command. (I never remember exactly where the p-values are.)