如何修复线性混合模型中的错误？

如何解决如何修复线性混合模型中的错误？

我正在运行一个线性混合模型。 我不知道为什么会出现以下错误。 固定效应是 ageroup 和 Phase。 随机效应为 username。

mixed.lmer <- lmer(compound ~ Phase_2 + agegroup + (1|username),data = df23)
summary(mixed.lmer)

model.frame.default 中的错误（数据 = df23，drop.unused.levels = TRUE，公式 = 复合 ~ ： 可变长度不同（找到“Phase_2”）

数据在 NA 和 agegroup 中有一些 score，因为并非所有用户名在所有阶段 (1,2,3) 中都具有 score 值。

> str(df23)
tibble [941,448 × 4] (S3: tbl_df/tbl/data.frame)
 $ username  : chr [1:941448] "____________bug" "____________bug" "____________bug" "____________Nia" ...
 $ Phase_2   : num [1:941448] 1 2 3 1 2 3 1 2 3 1 ...
 $ agegroup  : chr [1:941448] "YA" NA NA "YA" ...
 $ mean_score: num [1:941448] 0.735 NA NA 0.67 NA ...

structure(list(username = c("mimi__ramos","_LucyParker","bttrhmsndgrdns","chicken__lady","TheSolape","theg0ldengirl","RoyalFlushpokr","svnsettown","Abbeyleighsmith","12voltman60","MartiTweetsTV","alienrobotXperC","ChanIncognito","WoW_Chicken_","KayJ12","Abhisheksing164","tayyy_712","wiinterwiidow","ngyntlinh","theonlysesmo","LenairH","ImomoTimipa","ayjmnssigan","TheEdiCorona","makeyoubillions","jamescat79","Jro9902","_prettybrwn_","LilDisFan","yoavkaufman","FenrirCe","LaVidaBoricua","KHoey3000","Jessica_Wicket","spha305","selenasgardenn","ListElaine","emilyannleivas","dojae1s","tpwk_aiyana","davidalmond86","mmahalwy","danielandjay","ullyabigail","pawncrackers","ibsroyalty","coolbhopz","PolsSteph","practicalbob","Soccermom1714","jodimediocre","_RayGervais","A_Omodu","EdoCrypto2","nathan_hauk","asl3676","beckytaylorgill","lalisaglowz","pinkpluswhite__","a_time_in_grey","Ruth_fielamor","Besaid_Aurochs","FlannaganSonia","mo_onstarry","thisisnayeon","emnelder","Kate_Travels","pizza_fucker69","alexrubner","kissphoria","AquafabaMcGee","jahintheshammer","Kate_in_Guelph","tatertooots","Gillbaba302","guillotines_rus","Sara_Rossetti4","Sentry_23","imashvelasco","chinifromdabloc","seeonbothsidez","IamShainaJaye","ggrovyle","MsMicheleNicole","sire_gabriel","thafoodist","AudryFerraris","aksainsbury","alyakhairina_","csmitroyalfan","richardtheap","jkhuggins","nogumbofornazis","Filemon_Fly","LiamBeh98157012","P_Luxolo","simmonneee","GivHerMigraines","__ninarose"),compound = c(-0.765,0.743,0.1901,0.2732,-0.2263,0.3612,0.9906,0.4404,0.5106,-0.7533,-0.2942,0.7351,-0.9505,0.755,-0.4886,0.9266,0.7199,-0.4959,-0.0772,0.3818,0.8074,-0.34,-0.5983,0.6124,0.4648,0.7717,0.6369,-0.8176,0.9134,0.6239,0.7667,0.9817,0.6988,0.296,0.431,-0.5057,-0.4184,0.936,0.7346,0.6808,0.128,0.6249,0.2905,0.891,0.6597,0.5916,-0.4215,-0.0798,-0.0258,0.5684,-0.8122,0.5193,0.7579,0.8225,-0.0926,0.3515,0.7269,-0.6679,0.8622,-0.8126,0.2263,0.4588,-0.683,-0.3415,-0.3182,0.9819,-0.2732,-0.453,0),agegroup = c("YA","YA","MA","OA","YA"),Phase_2 = c(1,1,1)),row.names = c(NA,-100L),class = c("tbl_df","tbl","data.frame"))

我使用以下代码生成 df23。似乎这段代码没有生成正确的“年龄组”

df23 <- data_d %>%
  group_by(username,agegroup,Phase_2,) %>%
  summarise(mean_score = mean_(compound),.groups = 'drop') %>%
  complete(username,Phase_2)

find("Phase_2") 显示

> find("Phase_2")
[1] "df23" "df23" "df23" "data" "data"