R语言回归分析报错：coercion to 'logical(1)'问题排查求助

我基于调查结果进行回归分析，先通过以下代码生成衍生变量：

firstStepMutatedData <- mydata %>% 
    mutate(republican = ifelse((pid == 2), 1, 0),            
    conservative = ifelse((ideology >= 8), 1, 0),
    votedTrump = ifelse((Q38 == 2),1,0),
    white = ifelse((raceXethnic == "white & not hisp"),1,0),
    asian = ifelse((raceXethnic == "asian & not hisp" | raceXethnic == "asian & hisp"),1,0),
    otherRace = ifelse((raceXethnic != "white & not hisp" & raceXethnic != "asian & not hisp" & raceXethnic != "asian & hisp" ),1,0),
    wantMove1year = ifelse((Q17 <= 2),1,0),
    wantMove5year = ifelse((Q18 <= 2),1,0),
    jobProsp = ifelse((Q8 == 1), 1, 0),
    costProsp = ifelse((Q9 == 1), 1, 0),
    qualityProsp = ifelse((Q10 == 1), 1, 0),
    taxProsp = ifelse((Q11 == 2), 1, 0),
    schoolGovtProsp = ifelse((Q12 == 1), 1, 0),
    peopleProsp = ifelse((Q13 == 1), 1, 0),
    politicalProsp = ifelse((Q14 == 1), 1, 0),
    crimeProsp = ifelse((Q15 == 2), 1, 0),
    childrenLifeProsp = ifelse((Q16 == 1), 1, 0),
    buyHomeCA = ifelse((Q29 == 1),1,0),
    buyHomeElsewhere = ifelse((Q29 ==2),1,0),
    notBuyHome = ifelse((Q29==3),1,0))

随后运行回归并使用stargazer输出结果：

demoVote1 <- lm(votedTrump ~ republican + conservative, firstStepMutatedData)
demoVote2 <- lm(votedTrump ~ republican + conservative + white, firstStepMutatedData)
stargazer(demoVote1, demoVote2, type = "text")

运行后出现错误：

Error in (.format.s.statistics.list != "p25") && (.format.s.statistics.list != : 'length = 7' in coercion to 'logical(1)'

可能的问题及解决思路 #

• 版本兼容性冲突：这个错误大多是旧版stargazer与新版本R（尤其是4.0及以上版本）不兼容导致的。旧版stargazer在处理统计量格式的逻辑判断时，未考虑向量长度的问题，引发逻辑运算的长度不匹配错误。
  • 解决：执行install.packages("stargazer")更新到最新版本的stargazer包。
• 模型存在奇异拟合问题：如果回归模型中的自变量存在完全共线性（比如某个变量的取值完全由其他变量决定），会导致lm生成的模型对象结构异常，进而让stargazer无法正常解析。
  • 解决：先检查变量相关性，比如用cor(firstStepMutatedData[, c("republican", "conservative", "white")])查看相关系数；或者运行summary(demoVote1)、summary(demoVote2)，看模型是否有奇异拟合的警告信息。
• 数据存在缺失或异常值：若firstStepMutatedData中存在大量缺失值，或者因变量votedTrump的取值不符合预期（比如不是0/1），会导致lm模型的统计量输出结构异常，干扰stargazer的处理流程。
  • 解决：用na.omit(firstStepMutatedData)去除缺失值后重新运行回归；或者用table(firstStepMutatedData$votedTrump)确认因变量的取值是否正确。

内容的提问来源于stack exchange，提问作者Jessica Yu