I understand that they are looking for randomized studies, but there are some other criteria that I am less clear about.
The studies have to be about humans, is that correct?
Also, in the training it says that they accept studies that don’t meet the criteria of a true experiment exactly. For example, they accept quasi-randomized studies. I know what a randomized study is, but I’m a little less clear about just exactly what a quasi-randomized study is all about. If you could provide some clarification, that would be helpful. That they are looking for randomized studies is reasonably clear, but it gets tricky when they start adding in all the exceptions that they are making to their rule.
Hi Mike,
Emily from Cochrane Crowd here. First of all, it’s great that you’re checking out Cochrane Crowd! Welcome.
And they are great questions.
Yes, you are right that the studies have to be about humans.
In terms of quasi-randomised studies, sometimes in studies the form of randomisation used is not truly random. For example, treatment may have been allocated by date of birth or day of the week. These are what we think of as CCTs but you could also call them quasi-randomised controlled trials.
I’d recommend checking out the ‘Quick reference’ and ‘FAQs’ which you can access through your Cochrane Crowd dashboard, they are handy as a Cochrane Crowd newcomer.
Please feel free to keep the questions and feedback a-coming, we welcome it all! The best way to get hold of us is to tweet us @cochrane_crowd or email through to crowd@cochrane.org AND anna.noel-storr@rdm.ox.ac.uk. Anna and I work in opposite timezones so when the questions are time critical it’s best to email both of us!
Thanks again for connecting in with Cochrane Crowd and hope to see you at our CitSciMed Blitz challenge, which starts at 3pm GMT on Friday the 2nd.
Thank you for the reply, and email addresses! I find it interesting that the project accepts the quasi-randomized control trials as strict enough to draw conclusions from, but not the other types of studies that are not random at all. I guess it all depends on what type of conclusions one hopes to draw from the studies. I would think that anything less than true randomization would severely limit the conclusions that could be drawn from the studies, but I guess that’s not true in this instance.
I have thought of two questions while I’ve been doing this. One of them is some studies are described as randomized, single group assignment. If there’s only a single group, how can it be randomized? I thought you can only have randomization if you have at least two groups.
The other question I have is, some of these studies have not yet taken place, so I do not know if they are acceptable. Are studies that have not yet taken place acceptable to your project, or not? Sorry did I did not email you these questions, but I thought maybe others might have them as well. Also, by the time you reply, the challenge will probably be over anyway, so I didn’t see that sending an email was all that important.
Hey, Mike, don’t know about you, but I don’t expect to be very competitive in this challenge - many of the studies sound interesting and I keep reading them well after the classification is evident. Well, time to dodder on. Best of luck!
Mike C.
Best of luck to you too! I also find some of the articles interesting. I found one about a trial of a drug to treat frontal lobe dementia that I’ve never heard of before. The only thing about this challenge that’s kind of frustrating is that they don’t tell you whether or not you’re correct about your judgments for quite some time, so I have absolutely no idea how well I’m doing. I’m hoping that they’re just counting the number of documents completed, and not whether or not I’m getting them correct. I currently have absolutely no idea how I’m doing, or how I am stacking up against the other people who are playing. I guess I will only know after this competition is over.
Mike, if you go back to the Dashboard and click on “History and Settings” and then on “My Statistics,” you should find some accuracy info. Almost three-quarters of my records have already been compared with the “final decision.” (My stats were better than I expected, but that’s probably because I’m slow.)
Mike
You’re right that you can only have randomization if you have at least two groups. Quite of a lot of records that we’ll be screening have conflicting information. This means that the researchers entered the information incorrectly into the database that feeds Cochrane Crowd. One of the most common types of conflicting info is where it says randomised and single group assignment. On the whole, I think single group assignment has been selected in error, and that what the trialist actually meant was that the randomised trial was a single centre trial.Often you can spot this as you read the more detailed information in the record. But if you’re unsure it’s perfectly fine to classify the record as unsure.
In terms of trials that haven’t happened yet - yep, we’ll have those! : )
Thank you for your reply! If I say that I am unsure about a study, how does that affect my score vis-à-vis the final decision? If I say that I am unsure, that cannot be scored is either as I agree, or disagree, so how is the unsure classification handled when comparing it to the final decision?
