我正在使用bonferroni adj在R中进行成对t检验。除了一个我的数据的所有不同类别运行测试没有问题。某些测量与植物类别（如食者或群体）之间的t.tests结果是正常的。当我运行测试以按指定的植物类型比较该测量时，我得到了这个结果：
body {
font: 28px 'Cardo', 'Georgia', 'Times New Roman', Times, serif;
color: #272F32;
text-align: left;
background: #FFFFFF;
}
/* Main CSS & Header */
section {
max-width: 60%;
margin-right: 3%;
float: right;
text-align: left;
}
a {
color: #b10c00;
text-decoration: none;
}
a:hover {
color: #345059;
}
article a:visited, ul#main-menu a:visited {
color: #272F32
}
.group:after {
content: "";
display: table;
clear: both;
}
ul#main-menu {
display: block;
margin-top: 3%;
}
ul#main-menu li {
font: 24px 'Playfair Display', 'Georgia', 'Times New Roman', Times, serif;
display: inline-block;
float: right;
text-align: right;
margin-left: 40px;
text-decoration: overline;
}
header#main-header {
margin-bottom: 40px;
}
/* Articles Page */
article {
margin-bottom: 20px;
}
h1 {
font: 72px/1.2em 'Playfair Display', 'Times New Roman', Times, serif;
margin-bottom: 10px;
}
small {
font-size: 20px;
font-family: 'Playfair Display', 'Georgia', 'Times New Roman', Times, serif;
font-style: italic;
}
hr {
border: 0;
height: 0;
border-top: 1px solid rgba(0, 0, 0, 0.1);
border-bottom: 1px solid rgba(255, 255, 255, 0.3);
margin-top: 30px;
}
nav#pagination a{
font: 24px 'Playfair Display', 'Georgia', 'Times New Roman', Times, serif;
display: inline-block;
width: 45px;
float: right;
text-align: right;
}
nav#pagination a:last-child {
float: left;
text-align: left;
}
footer {
margin-bottom: 20px;
}
/* Single Article */
.single-article header h1{
position: absolute;
left: -25px;
top: -50px;
max-width: 1050px;
font-size: 150px;
margin: 0;
padding: 0;
z-index: -1;
}
.single-article header {
margin-bottom: 310px;
}
.single-article p {
font-size: 24px;
line-height: 1.4em;
margin-bottom: 36px;
}
article.single-article p:last-child {
background: green;
margin-bottom: 10px;
}
img {
position: absolute;
left: 0;
max-width: 100%;
height: 20%;
background-size: cover;
align-content: center;
}
/* Media Queries */
@media (max-width: 1600px) {
body, ul#main-menu li, nav#pagination a {
font-size: 24px;
}
h1 {
font-size: 48px;
}
}
@media (max-width: 1300px) {
section {
min-width: 60%;
margin-right: 4%;
}
body, ul#main-menu li, nav#pagination a {
font-size: 20px;
}
ul#main-menu {
margin-top: 5%;
}
h1 {
font-size: 48px;
margin-bottom: 10px;
}
small {
font-size: 18px;
}
}
@media (max-width: 700px) {
section {
min-width: 65%;
margin-right: 5%;
}
body, ul#main-menu li, nav#pagination a {
font-size: 18px;
}
ul#main-menu {
margin-top: 7%;
}
hr {
margin-top: 25px;
}
h1 {
font-size: 36px;
margin-bottom: 10px;
}
small {
font-size: 16px;
}
}
@media (max-width: 430px) {
section {
min-width: 80%;
margin: 10%;
}
body, ul#main-menu li, nav#pagination a {
font-size: 16px;
}
ul#main-menu {
min-width: 100%;
margin: auto;
}
ul#main-menu li {
float: none;
margin: 0 auto;
padding: 0 15px;
text-align: left;
}
hr {
margin-top: 20px;
}
h1 {
font-size: 26px;
margin-bottom: 5px;
}
small {
font-size: 14px;
}
}我的代码是：
<!DOCTYPE html lang="es">
<html>
<head>
<title>Article Proof</title>
<!-- Asset links -->
<link href='https://fonts.googleapis.com/css?family=Playfair+Display:700,400italic|Cardo' rel='stylesheet' type='text/css'>
<link href="art/assets/reset.css" media="screen" rel="stylesheet" type="text/css" />
<link href="art/assets/style.css" media="screen" rel="stylesheet" type="text/css" />
</head>
<body>
<section class="group">
<header id="main-header">
<nav>
<ul id="main-menu" class="group">
<li><a href="index.html">Blog</a></li>
<li><a href="work.html">Work</a></li>
<li><a href="about.html">About</a></li>
</ul>
</nav>
</header>
<article class="single-article">
<header>
<a href="#try"><h1>The coming of age of the century</h1></a>
</header>
<p>Est multos in quis est tempor ea incurreret. Senserit amet laborum, singulis eu
vidisse. Ingeniis tamen iis tempor relinqueret de cernantur est velit vidisse,
dolor ex nostrud, officia ipsum officia tempor, deserunt relinqueret et
ingeniis, sed pariatur transferrem eu admodum legam ipsum aut eram, id ad varias
litteris.
Sint consequat do officia te incurreret fugiat laborum appellat, lorem
excepteur reprehenderit quo anim commodo tractavissent ubi ingeniis lorem
vidisse nostrud a o quibusdam efflorescere sed legam quamquam in singulis aut
offendit firmissimum o laboris. Hic quem aliquip eu nulla litteris exquisitaque
se nam si philosophari an expetendis tamen summis te legam ex ex fugiat quid
quae singulis se duis singulis sed reprehenderit, iis summis multos ita litteris
eu sunt officia te eiusmod. Non legam eiusmod hic nam fabulas efflorescere,
culpa consequat si arbitror si se se legam tempor. Mentitum quis amet o legam.
Malis excepteur firmissimum, nescius iis nostrud, magna ullamco cohaerescant te
irure appellat possumus an fugiat consectetur probant ipsum expetendis qui si in
malis velit quid, tamen iis de malis cernantur ad ad anim deserunt deserunt.</p>
<img src="art/assets/image.jpg">
<p>Sed o nulla sint quorum. Nisi e ne quid laboris, fabulas cillum fugiat
expetendis aliqua se ab velit sed summis, an de varias quibusdam, litteris aut
cillum pariatur an in fore tractavissent nam laboris ubi offendit ex dolore
mentitum reprehenderit.
Ad dolor occaecat, officia multos possumus non eu legam
et quem qui te se sunt minim tamen. Excepteur si amet proident, aut commodo
distinguantur non et hic nulla dolore varias ut irure ea ingeniis e illum, an
ipsum irure qui cernantur, mandaremus ad senserit, lorem doctrina sed
instituendarum non dolore ne singulis ea culpa. Excepteur se fugiat, enim id
probant, ne nisi in anim iis irure occaecat ex nostrud, incurreret ad cillum
tempor, consequat sempiternum ad admodum ubi nulla pariatur et velit amet, ab
est domesticarum.
De dolore transferrem id te malis a amet, an ubi eram officia,
labore expetendis philosophari. Ne quem aliquip, arbitror ad duis o quae
cernantur non graviterque ab do aut nulla excepteur si quem possumus est
nostrud, minim relinqueret officia nulla cernantur ab nam culpa vidisse
consectetur ut sed qui quae mentitum.</p>
<p>Proident summis ingeniis officia. Do quis iudicem fidelissimae. Admodum se
mentitum. Si aute tempor laboris. Ita duis offendit admodum ab est aliqua irure
ubi laborum id eiusmod quae aut possumus firmissimum ex te nostrud a laborum et
se sunt aute de cupidatat ubi laboris lorem magna mentitum fore eu deserunt ubi
veniam ut magna quamquam praetermissum. Fabulas amet consequat. Ad duis
exercitation, mandaremus malis arbitror deserunt, iis eram summis aut senserit
se duis litteris reprehenderit.</p>
<p>Commodo legam quorum si dolor si incididunt fore dolore incididunt enim. Tamen
domesticarum doctrina cillum ingeniis hic se fugiat amet ne appellat, excepteur
de sint aliquip hic tamen sed litteris, aliquip eram an proident voluptatibus
non se sint aliquip familiaritatem ab id amet tempor senserit, irure relinqueret
excepteur aute litteris. Hic esse occaecat fabulas, illum ita aut irure
quamquam. Quorum se officia a multos o iis quem incurreret cupidatat. A cillum
tamen quorum commodo, in nescius illustriora, ea litteris praesentibus non quid
fidelissimae expetendis culpa offendit iis eram voluptate despicationes eu nam
tamen lorem ne doctrina non culpa admodum eiusmod ne quae proident commodo. Se
se dolor quorum irure, cupidatat quid tamen nescius enim, ita dolor mentitum, ad
culpa malis id incididunt. Ea incurreret ita iudicem iis mandaremus quo quae
offendit, ne aliqua nescius, ad si amet malis minim e est quid ut quis, culpa
consequat ne esse quis, ea irure fabulas doctrina, quem proident sed pariatur.</p>
<p>Te varias consectetur, sed labore ea irure. Quem quo pariatur. E possumus
coniunctione, iudicem et aliqua e id quae cohaerescant nam occaecat velit a
arbitror graviterque. Quorum ubi ut cillum ullamco ita quamquam arbitrantur si
senserit. Do nulla doctrina concursionibus nam duis eu eiusmod si minim, lorem
relinqueret senserit varias tempor ubi se veniam fabulas illustriora. Cernantur
a nulla. Culpa cernantur o iudicem.</p>
<small>written 4 days ago</small>
</article>
<footer>
<hr/>
<small class="credit"><a href="https://github.com/ferpalaciosd/artanchor" title="Art: GitHub">Art</a> for <a href="http://anchorcms.com" title="Anchor CMS Homepage">Anchor CMS</a>, Created by <a href="https://twitter.com/ferpalaciosd" title="Fernando Palacios Dueso">Ferpalaciosd</a></small>
</footer>
</section>
</body>
</html>这与用于按组进行t检验的代码完全相同。
我的数据看起来像
Pairwise comparisons using t tests with pooled SD
data: `plants$dC and plants$type`
dg gr hb li sh
gr - - - - -
hb - - - - -
li - - - - -
sh - - - - -
tr - - - - -
我已经仔细检查了我的数据是否有NAs，空值等等。
任何想法是什么问题？
答案 0 :(得分：0)
当存在只有一个项目的组时，就会出现此问题。例如：
pairwise.t.test(1:7, c("A", "A", "B", "B", "C", "C", "D"))
返回：
Pairwise comparisons using t tests with pooled SD
data: 1:7 and c("A", "A", "B", "B", "C", "C", "D")
A B C
B - - -
C - - -
D - - -
P value adjustment method: holm
解决方案是在进行t检验之前过滤掉仅出现一次的所有组（无论如何这些检验都没有意义）。例如，可以在基数R或dplyr中完成此操作：
# Base R
plants_filtered <- plants[tapply(plants$type, plants$type, length), ]
# dplyr
plants_filtered <- plants %>%
group_by(group) %>%
filter(n() >= 2) %>%
ungroup()