R语言回归输出显示类别标签及模型输出格式不符问题

Hey there! Let's tackle your two issues with R's linear model output one by one.

First: Why does your output look different from the tutorial? #

Looking at your residual output (each observation's residual listed individually), that's not the default behavior of summary.lm()—usually it shows the 5-number summary of residuals (Min, 1Q, Median, 3Q, Max). Here are the most likely reasons:

• You've got a custom/replaced summary.lm function: If you loaded packages like arm, lme4 (with helper functions), or wrote your own summary.lm function, it might override the base R version. Check with this code:

  getS3method("summary", "lm")
  

  If the output isn't the base R function, restart your R session without loading extra packages, or force the base version with base::summary(xmdl).

• Odd output settings: It's possible you accidentally set an option that changes residual display. Reset all options to default with:

  options(reset = TRUE)
  

  Then re-run summary(xmdl) to see if it matches the tutorial.

Second: How to show clear category labels in regression output #

By default, R uses treatment contrasts for factors: it picks the first level as the reference group, and only shows coefficients for other levels relative to that reference (in your case, sexmale means male vs. female pitch difference). Here are ways to make labels clearer:

1. Set a clear reference group #

Relevel your factor to make the reference explicit, so coefficient labels tell the full story:

# Set male as the reference group
my.df$sex <- relevel(my.df$sex, ref = "male")
xmdl <- lm(pitch ~ sex, my.df)
summary(xmdl)

Now the coefficient will show sexfemale, clearly indicating female vs. male difference.

2. Use the broom package for clean, labeled output #

The broom package turns model outputs into tidy data frames with clear labels:

library(broom)
tidy(xmdl)

You'll get a table like this:

# A tibble: 2 × 5
  term        estimate std.error statistic p.value
  <chr>          <dbl>     <dbl>     <dbl>   <dbl>
1 (Intercept)     128.      7.53      17.0 2.24e-4
2 sexfemale       104.     10.7        9.70 1.04e-3

3. Use car::Anova() for factor-level effects #

If you want to see the overall significance of the sex factor (not just group vs. reference), use the car package:

library(car)
Anova(xmdl)

This output clearly labels the sex factor and its significance.

4. Manually rename coefficients #

For full control, rename the model's coefficients directly to human-readable text:

xmdl <- lm(pitch ~ sex, my.df)
names(xmdl$coefficients) <- c("Average Pitch (Female)", "Pitch Difference (Male vs. Female)")
summary(xmdl)

Now the coefficient table will have intuitive labels instead of the default shorthand.

内容的提问来源于stack exchange，提问作者RobS