Independent Study Results?

[danzgymn86]

Advertisement - Members don't see this ad

So, I have some preliminary pilot data (granted, I only have 5 per condition) but so far, it's looking like the exact opposite of my prediction is going to happen. I know it's not a lot and I'll have 75 per condition when all is said and done, but I'm getting really nervous. Is that a really bad thing?

 

[RayneeDeigh]

Of course not! Data is data, and finding SOME effect is arguably a lot better than finding no effect at all.

It's only bad if you were hell bent on your hypotheses being correct. I think it's pretty exciting.

 

[danzgymn86]

Yea I basically had a 50-50 shot at predicting and may have screwed it up! "This is going to happen for condition x but not for condition y." It's looking like it might happen for condition y...

 

[Ollie123]

Happens all the time. Better to find the opposite effect than no effect.

Your predictions were wrong. I promise you it won't be the last time🙂

 

[JockNerd]

Data are always friendly.

I'm not clear on what you're doing, but n = 5 isn't really enough to make a judgment about anything, much less give up hope for your hypothesis (especially since I'm guessing your power analysis or the effect size you're looking for requires the 75 per condition).

Even if it doesn't work out, you can take comfort in knowing that senior thesis research doesn't typically go too far 🙂

 

[danzgymn86]

But how would I address this in the final paper. (If my results actually do show this.) Do I basically say, "I was wrong, but there is other data supporting why I was wrong?" (Obviously, not in so many words.)

 

[RayneeDeigh]

My predictions have always been shots in the dark anyway. I mean sure you can do your lit review and get a good educated opinion about what might happen, but often you can find support in the literature for an effect either way and you have to pick one that you think your study will get and back it up. Just think, when you write up this study you'll have a fascinating discussion section. I always love reading "this is why I found the opposite results" more than "look at me, everything I predicted came true!"

 

[JockNerd]

But how would I address this in the final paper. (If my results actually do show this.) Do I basically say, "I was wrong, but there is other data supporting why I was wrong?" (Obviously, not in so many words.)

Say it didn't work out, try to find out why, try to think of a way to make it work in the future.

But don't worry about it till you get a bigger sample!

 

[RayneeDeigh]

Oh and don't get tempted to change your hypothesis post-hoc. I know it's tempting, but don't do it.

Plus, it's just 5 people in each group so far. The first ten surveys I gave out for my undergrad thesis came back with the craziest data that completely went against my hypotheses. Then after I had 200 participants my hypotheses ended up being mostly true and pretty significant. So never fear, the beginning is not always indicative of the end.

 

[Ollie123]

Seriously, it will be fine.

I've read Journal of Abnormal papers where this has happened. I THINK if JAP editors can accept it, an undergraduate thesis committee will manage to do so.

 

[psychanon]

First, don't try to interpret your data at this point. If you're planning to have 75 p's per condition, that must mean that you think you need a lot of power to find your effect. So, your results thus far are completely uninterpretable, and there's no need to think about them. It could definitely go either way. I know it's tempting to peek at the data before it's ready-- I've been there-- but don't get stressed out about it.

Second, if you can come up with an interpretable explanation, it could be really interesting even with opposite results. People change their hypotheses post-hoc all the time-- I know it's "against the rules", but things don't always go by the book. Probably the best thing you can do is to set up your intro as if you're testing two competing hypotheses..."according to theory x, this might happen, but according to previous research y, this might happen instead." In fact, as a general rule, if you can design your experiments so that you are testing two competing hypotheses, it's ideal, because no matter what the results, you end up being right. If that does work for you, though, you can just say that the results were in the opposite direction as excepted, and speculate why. The important thing (in terms of publication) is that you can tell a coherent story with your data, in a way that meaningfully adds to the literature.

EDIT: Oh, is this just for a senior thesis? Then don't even worry about it. Even if it were your dissertation, they couldn't not pass you for not getting predicted results-- the spirit of science is that we go in with questions, not with answers.

 

[danzgymn86]

^^ Yes, it's for undergrad. However, I obviously still want to get results!