Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 4 down vote favorite
1

I ran an ANOVA finding for example an interaction between gender and grade than I want to know in what grades boys and girls differ, but in many cases I find (adjusted) p-values of 0 and 1. How / why is this possible? Doesn't seem right...

as.factor(gender)                     1     16    16.2    2.6377  0.104396    
as.factor(grade)                      7  50077  7153.9 1165.4184 < 2.2e-16 ***
as.factor(gender):as.factor(grade)    7    132    18.9    3.0795  0.003056 ** 
Residuals                          7747  47555     6.1                        
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

  Tukey multiple comparisons of means
    95% family-wise confidence level

Fit: aov(formula = rating ~ as.factor(gender) * as.factor(grade), data = users_c[users_c$grade %in% 1:8, ])

$`as.factor(gender)`
           diff        lwr        upr     p adj
m-f -0.09135851 -0.2016276 0.01891058 0.1043964

$`as.factor(grade)`
         diff        lwr       upr     p adj
2-1 0.3823566 -0.5454435  1.310157 0.9169296
3-1 1.9796023  1.1649854  2.794219 0.0000000
4-1 3.9558543  3.1534606  4.758248 0.0000000
5-1 5.7843111  4.9829529  6.585669 0.0000000
6-1 7.0752044  6.2708610  7.879548 0.0000000
7-1 8.4868609  7.6776332  9.296089 0.0000000
8-1 9.3867231  8.5626511 10.210795 0.0000000
3-2 1.5972457  1.0395026  2.154989 0.0000000
4-2 3.5734976  3.0337642  4.113231 0.0000000
5-2 5.4019544  4.8637616  5.940147 0.0000000
6-2 6.6928478  6.1502200  7.235476 0.0000000
7-2 8.1045042  7.5546625  8.654346 0.0000000
8-2 9.0043665  8.4329024  9.575831 0.0000000
4-3 1.9762520  1.6694948  2.283009 0.0000000
5-3 3.8047088  3.5006705  4.108747 0.0000000
6-3 5.0956021  4.7837806  5.407424 0.0000000
7-3 6.5072586  6.1830461  6.831471 0.0000000
8-3 7.4071208  7.0474558  7.766786 0.0000000
5-4 1.8284568  1.5588754  2.098038 0.0000000
6-4 3.1193501  2.8410202  3.397680 0.0000000
7-4 4.5310066  4.2388618  4.823151 0.0000000
8-4 5.4308688  5.0998193  5.761918 0.0000000
6-5 1.2908933  1.0155630  1.566224 0.0000000
7-5 2.7025498  2.4132612  2.991838 0.0000000
8-5 3.6024120  3.2738803  3.930944 0.0000000
7-6 1.4116565  1.1141985  1.709114 0.0000000
8-6 2.3115187  1.9757711  2.647266 0.0000000
8-7 0.8998622  0.5525763  1.247148 0.0000000

$`as.factor(gender):as.factor(grade)`
                diff         lwr        upr     p adj
m:1-f:1  0.005917865 -1.77842639  1.7902621 1.0000000
f:2-f:1  0.318074165 -1.28953805  1.9256864 0.9999988
m:2-f:1  0.442924925 -1.11597060  2.0018205 0.9998619
f:3-f:1  1.769000750  0.35262166  3.1853798 0.0020136
m:3-f:1  2.174229216  0.76569156  3.5827669 0.0000147
f:4-f:1  3.738998543  2.34268666  5.1353104 0.0000000
m:4-f:1  4.163719997  2.77146170  5.5559783 0.0000000
f:5-f:1  5.769586591  4.37599400  7.1631792 0.0000000
m:5-f:1  5.816721075  4.42497532  7.2084668 0.0000000
f:6-f:1  7.169439003  5.77317769  8.5657003 0.0000000
m:6-f:1  7.000924045  5.60308216  8.3987659 0.0000000
f:7-f:1  8.330142924  6.92683436  9.7334515 0.0000000
m:7-f:1  8.674488370  7.26930678 10.0796700 0.0000000
f:8-f:1  9.535307293  8.11198164 10.9586329 0.0000000
m:8-f:1  9.251081088  7.82191240 10.6802498 0.0000000
f:2-m:1  0.312156300 -1.12690148  1.7512141 0.9999959
m:2-m:1  0.437007060 -0.94741539  1.8214295 0.9995001
f:3-m:1  1.763082885  0.54136279  2.9848030 0.0000892
m:3-m:1  2.168311350  0.95569081  3.3809319 0.0000001
f:4-m:1  3.733080678  2.53468294  4.9314784 0.0000000
m:4-m:1  4.157802132  2.96412989  5.3514744 0.0000000
f:5-m:1  5.763668726  4.56844048  6.9588970 0.0000000
m:5-m:1  5.810803210  4.61772882  7.0038776 0.0000000
f:6-m:1  7.163521138  5.96518233  8.3618599 0.0000000
m:6-m:1  6.995006180  5.79482611  8.1951862 0.0000000
f:7-m:1  8.324225059  7.11768240  9.5307677 0.0000000
m:7-m:1  8.668570505  7.45984987  9.8772911 0.0000000
f:8-m:1  9.529389428  8.29962271 10.7591561 0.0000000
m:8-m:1  9.245163223  8.00863850 10.4816879 0.0000000
m:2-f:2  0.124850760 -1.02282435  1.2725259 1.0000000
f:3-f:2  1.450926585  0.50586965  2.3959835 0.0000172
m:3-f:2  1.856155050  0.92289131  2.7894188 0.0000000
f:4-f:2  3.420924378  2.50621691  4.3356318 0.0000000
m:4-f:2  3.845645832  2.93713824  4.7541534 0.0000000
f:5-f:2  5.451512425  4.54096139  6.3620635 0.0000000
m:5-f:2  5.498646910  4.59092496  6.4063689 0.0000000
f:6-f:2  6.851364838  5.93673457  7.7659951 0.0000000
m:6-f:2  6.682849880  5.76580854  7.5998912 0.0000000
f:7-f:2  8.012068759  7.08671595  8.9374216 0.0000000
m:7-f:2  8.356414205  7.42822339  9.2846050 0.0000000
f:8-f:2  9.217233128  8.26179669 10.1726696 0.0000000
m:8-f:2  8.933006923  7.96888762  9.8971262 0.0000000
f:3-m:2  1.326075825  0.46649985  2.1856518 0.0000150
m:3-m:2  1.731304290  0.88471145  2.5778971 0.0000000
f:4-m:2  3.296073618  2.46998162  4.1221656 0.0000000
m:4-m:2  3.720795071  2.90157332  4.5400168 0.0000000
f:5-m:2  5.326661665  4.50517434  6.1481490 0.0000000
m:5-m:2  5.373796150  4.55544575  6.1921465 0.0000000
f:6-m:2  6.726514078  5.90050756  7.5525206 0.0000000
m:6-m:2  6.557999120  5.72932364  7.3866746 0.0000000
f:7-m:2  7.887217999  7.04935402  8.7250820 0.0000000
m:7-m:2  8.231563445  7.39056617  9.0725607 0.0000000
f:8-m:2  9.092382368  8.22140761  9.9633571 0.0000000
m:8-m:2  8.808156163  7.92766524  9.6886471 0.0000000
m:3-f:3  0.405228465 -0.13578346  0.9462404 0.4221367
f:4-f:3  1.969997793  1.46166478  2.4783308 0.0000000
m:4-f:3  2.394719246  1.89762897  2.8918095 0.0000000
f:5-f:3  4.000585840  3.49977062  4.5014011 0.0000000
m:5-f:3  4.047720325  3.55206739  4.5433733 0.0000000
f:6-f:3  5.400438253  4.89224417  5.9086323 0.0000000
m:6-f:3  5.231923295  4.71940255  5.7444440 0.0000000
f:7-f:3  6.561142174  6.03389412  7.0883902 0.0000000
m:7-f:3  6.905487620  6.37327442  7.4377008 0.0000000
f:8-f:3  7.766306543  7.18788499  8.3447281 0.0000000
m:8-f:3  7.482080337  6.88942637  8.0747343 0.0000000
f:4-m:3  1.564769328  1.07871270  2.0508260 0.0000000
m:4-m:3  1.989490781  1.51520464  2.4637769 0.0000000
f:5-m:3  3.595357375  3.11716862  4.0735461 0.0000000
m:5-m:3  3.642491860  3.16971239  4.1152713 0.0000000
f:6-m:3  4.995209787  4.50929846  5.4811211 0.0000000
m:6-m:3  4.826694830  4.33626022  5.3171294 0.0000000
f:7-m:3  6.155913709  5.65010831  6.6617191 0.0000000
m:7-m:3  6.500259155  5.98928021  7.0112381 0.0000000
f:8-m:3  7.361078078  6.80213257  7.9200236 0.0000000
m:8-m:3  7.076851872  6.50319055  7.6505132 0.0000000
m:4-f:4  0.424721453 -0.01192015  0.8613631 0.0668946
f:5-f:4  2.030588047  1.58971048  2.4714656 0.0000000
m:5-f:4  2.077722532  1.64271796  2.5127271 0.0000000
f:6-f:4  3.430440460  2.98119847  3.8796825 0.0000000
m:6-f:4  3.261925502  2.80779484  3.7160562 0.0000000
f:7-f:4  4.591144381  4.12045589  5.0618329 0.0000000
m:7-f:4  4.935489827  4.45924616  5.4117335 0.0000000
f:8-f:4  5.796308750  5.26892973  6.3236878 0.0000000
m:8-f:4  5.512082545  4.96913148  6.0550336 0.0000000
f:5-m:4  1.605866594  1.17800058  2.0337326 0.0000000
m:5-m:4  1.653001078  1.23118920  2.0748130 0.0000000
f:6-m:4  3.005719006  2.56923916  3.4421989 0.0000000
m:6-m:4  2.837204048  2.39569420  3.2787139 0.0000000
f:7-m:4  4.166422928  3.70789927  4.6249466 0.0000000
m:7-m:4  4.510768373  4.04654394  4.9749928 0.0000000
f:8-m:4  5.371587296  4.85503631  5.8881383 0.0000000
m:8-m:4  5.087361091  4.55492128  5.6198009 0.0000000
m:5-f:5  0.047134485 -0.37906079  0.4733298 1.0000000
f:6-f:5  1.399852412  0.95913504  1.8405698 0.0000000
m:6-f:5  1.231337454  0.78563790  1.6770370 0.0000000
f:7-f:5  2.560556334  2.09799705  3.0231156 0.0000000
m:7-f:5  2.904901779  2.43669086  3.3731127 0.0000000
f:8-f:5  3.765720703  3.24558412  4.2858573 0.0000000
m:8-f:5  3.481494497  2.94557538  4.0174136 0.0000000
f:6-m:5  1.352717928  0.91787572  1.7875601 0.0000000
m:6-m:5  1.184202970  0.74431204  1.6240939 0.0000000
f:7-m:5  2.513421849  2.05645683  2.9703869 0.0000000
m:7-m:5  2.857767295  2.39508230  3.3204523 0.0000000
f:8-m:5  3.718586218  3.20341827  4.2337542 0.0000000
m:8-m:5  3.434360013  2.90326187  3.9654582 0.0000000
m:6-f:6 -0.168514958 -0.62249009  0.2854602 0.9968060
f:7-f:6  1.160703921  0.69016548  1.6312424 0.0000000
m:7-f:6  1.505049367  1.02895400  1.9811447 0.0000000
f:8-f:6  2.365868290  1.83862318  2.8931134 0.0000000
m:8-f:6  2.081642085  1.53882109  2.6244631 0.0000000
f:7-m:6  1.329218879  0.85401081  1.8044269 0.0000000
m:7-m:6  1.673564325  1.19285330  2.1542753 0.0000000
f:8-m:6  2.534383248  2.00296656  3.0657999 0.0000000
m:8-m:6  2.250157043  1.70328327  2.7970308 0.0000000
m:7-f:7  0.344345446 -0.15203755  0.8407284 0.5648416
f:8-f:7  1.205164369  0.65953016  1.7507986 0.0000000
m:8-f:7  0.920938164  0.36023867  1.4816377 0.0000022
f:8-m:7  0.860818923  0.31038540  1.4112524 0.0000101
m:8-m:7  0.576592718  0.01122178  1.1419637 0.0401330
m:8-f:8 -0.284226205 -0.89329509  0.3248427 0.9688007
share|improve this question
7747 residual degrees of freedom is a lot; is it possible that your data set has multiple responses per individual? If that's the case, you might want to either collapse each person's responses to a mean (automatically done by ezANOVA from the ez package), or use something like mixed effects models, which permit you to account for the repeated measurements (check out ezMixed from the ez package). – Mike Lawrence Sep 5 '11 at 19:17
I meant to say "or use something more powerful like mixed effects models". Also, for the latest version of the ezMixed code (which permits powerful evaulation of possibly non-linear effects of continuous variables like grade, not to mention visualization via ezPlot2), source and run this ezDev function while connected to the internet: raw.github.com/mike-lawrence/ez/master/R/ezDev.R – Mike Lawrence Sep 5 '11 at 19:41

migrated from stackoverflow.com Sep 5 '11 at 17:03

1 Answer

up vote 12 down vote

All that the 0 and 1 mean are that they are very very close to 0 or 1. If you look carefully you'll see that when the adjusted p is 1 then the effect is almost 0 and when the adjusted p is 0 the nearer bound of the effect is very far away. Therefore, there's nothing "wrong" per se. Now look at how many significant digits you have. The 1 or 0 just means that it's closer to that value than can be represented by a number with that many digits. Feel free to report something like < 0.0001, or > 0.9999.

share|improve this answer
+1 - Those are just arbitrary rounding thresholds. And one of the reasons I really hate * based significance reporting. – EpiGrad Sep 5 '11 at 22:12
2  
With that large a sample size it is not surprising to find really small p-values. I think it raises the question of practical vs statistical significance here and I would be more interested in the confidence intervals than p-values. – Glen Sep 6 '11 at 1:45
@John, do you mean to imply that there would be a problem with reporting a p-value as 1.00 or 1.000? I would see nothing wrong with doing this. – mark999 Sep 6 '11 at 1:57
Glen, I agree... – John Sep 6 '11 at 10:07
mark999, then you should report them that way. The only issue that I'd have with that is that such numbers tend to be interpreted special. We all know any value would be an estimate but 1.0 and 0.0 might be considered special or confusing to statistical novices just as they were to this questioner. The confusion that prompted this question would then be in readers of the report. – John Sep 6 '11 at 10:11
show 3 more comments

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.