Miles has e-mailed Ren to come to a non-descript, out-of-the way pub on a Monday night. Ren assumes this is to do with Lucius – he told Miles on Friday his discoveries about Lucius.
When Ren gets there he finds Miles with two younger guys. Miles introduces them as Ken and Halberd. They’re social psychology postgraduates. They’re not Lucius’s students but they’ve been working with him on the analysis of some of his studies.
After getting drinks in, they all sit down to talk.
‘So,’ says Ren. ‘Is this about…?’
‘It’s about Lucius, yes,’ says Miles. ‘I’ve told these two what you’ve discovered. Or possibly discovered, we don’t want to jump to conclusions just yet. But Ken and Halberd have been suspicious of Lucius for a while, as I discovered a couple of weeks ago, when we were having a drunken conversation in the pub about fraud in psychology. They’ve been analysing his work, and they think there are very suspicious things in Lucius’s data.’
‘Like data that’s too good to be true?’ says Ren.
‘Yes,’ says Ken. ‘Sometimes, anyway. In the earlier work it looked more credible. Then it started to look more perfect as it went on. But recently it’s got more credible again. This suggests, assuming he’s massaging or even inventing his data, that he was careful early on to make the data look realistic, then he got blase about it for a while, and then he started being careful again, after realising that he was being too risky.’
‘Something else we’ve discovered is the Cronbach Alpha scores on some of his studies are very low,’ says Halberd.
‘It’s a measure of the internal consistency, or reliability, of the responses made. Like if a respondent says on one question that they’re an atheist, then they shouldn’t say on another question that they believe in God. If the internal consistency is low then that’s a problem. It means you may have participants who are just answering anything to the questions, and don’t care, in which case you can’t take the answers as confirming anything.’
‘But it seems very unlikely that Lucius would get so many participants who are just answering the questions randomly,’ says Ken.
‘Low Cronbach Alpha scores can also be caused by questionnaires that have been poorly designed,’ says Halberd.
‘Lucius’s questionnaire’s, however, aren’t poorly designed,’ says Ken.
‘So let me guess,’ says Ren. ‘Another thing that can explain low Cronbach Alpha scores is that the analyst has changed some responses without checking how this affects the overall consistency of that participant’s answers.’
‘Exactly,’ says Ken.
‘Hmm, not good for Lucius,’ says Ren. ‘Did Miles tell you that I noticed that a lot of Lucius’s studies were done at Lancedown University, years after he left it, yet no-one from Lancedown was listed as a co-author?’
‘He told us, yeah,’ says Ken. ‘We hadn’t noticed that, but you’re right, it’s suspicious. Not in itself, perhaps, but along with everything else it’s adding up to… well, I hesitate to use too strong a word here. Let’s just see how we go with some more analysis. There’s various other weird things to do with his work which we’re taking a closer look at, and we’ve been setting up some replications of his studies, which we were hoping to do over summer.’
‘He heard that they want to replicate his studies,’ says Miles, ‘and he tried to discourage them. And now he’s subtly trying to sideline them from his analysis team.’
‘So what has he been giving you to analyse?’ asks Ren. ‘Do you ever see, like, responses written in different handwriting? Videos of people giving answers? Anything like that?’
‘No, it’s always just computer files of results.’
‘Do you know anyone who has ever worked on one of his studies, and actually seen it happen?’
‘No. These are further reasons why we were suspicious.’
‘He loves the limelight, does Lucius,’ says Miles. ‘He isn’t going to go down without a fight, and he’ll try to take you all down with him. You guys sure you’re prepared to face that?’
‘I am a little scared,’ says Halberd.
‘I’m not,’ says Ren. ‘Bring it on, data-fucker.’
‘I’m worried too,’ says Ken. ‘Perhaps we should have stayed out of it. But once we joined his analysis team, we had no choice. Imagine if he got exposed as a fraud later on, and it was known that we were on his team. How would that look?’
‘So what about this Tobias Woolley guy?’ says Ren. ‘Do you think he’s in on it?’
‘We’re not sure, but we suspect not,’ says Halberd. ‘If Lucius is faking everything then Tobias won’t be involved in any studies, because there are no studies. And if Lucius just sends him data he’s going to assume that it’s kosher.’
‘Tobias is very political, in a very admirable way, in my view’ says Ken, ‘but he’s a bit gullible, I’d say, having met him at a conference once. Like, you could tell him anything, as he’d believe it, as long as it fits in with his politics.’
‘Or maybe he suspects that Lucius is making it all up,’ says Ren, ‘but he’s keeping quiet because the conclusions are politically agreeable to him.’
‘Well, who knows?’ says Ken. ‘This is all speculation. All we can do is dig deeper.’
‘It’s depressing,’ says Halberd. ‘I never knew academia could be so corrupt and political.’
‘Oh, I knew that,’ says Ren. ‘What I never realised was that it could be so much fun.’