14 replies to this topic

[Mac2012]

What type of testing would i use to conclude my experiments validity for my IA?

[blindpet]

I assume by validity you mean statistical testing, in that case we need to know the following:
Are you HL or SL?
What exactly did the participants do?
How are your data measured? (numbers, yes/no answers etc)

If you mean something else then please clarify what kind of validity you want to verify.

[Shahad]

Hey, i'm currently working on my Psych IA Loftus and Palmer SL.
I was wondering, what exactly does the results section in SL require?
My teacher told me not to do standard deviation and now she says i need to. But she is even questioning her own judgement.
In the assessment rubric it says that descriptive statistics are required.
What exactly does that mean?
Like, percentages of how many people said yes to broken glass or like the mean of the speed estimation?
or does it mean standard deviation?

[CUB^3D]

I did loftus and palmer and I have HL.

They way I did to show my results was by using x^2 (chi square) with that you can calcullate the probability that one predicted the velocity over or under the average, hence if participants over or under estimated the velocity.

[Yap]

Currently doing my Psychology IA, Loftus and Palmer 1974

[Shahad]

Is the chi square for descriptive or inferential statistics though?
And i did standard deviation for my results just to see how it turned out.
It was a really large value, in the 30s actually.
Yap, did you do your experiment yet?
My results coincided by the original Loftus and Palmer study, but the results i got were farily higher.
Like the speed estimations were pretty high, but they were still higher for smashed and lower in bumped.

[CUB^3D]

Well, don't mind your results just discuss them later on. Why it might be different and so on. My results showed that it was the opposite to what Loftus and Palmer said. But it was beacuse of uncontrollable factors.

and for the Chi square, I'm pretty sure it's inferential. I did it a while ago so you might wanna check that up

[Shahad]

Fair enough. My results would have opposed what Loftus and Palmer said if i took out the outliers.
And inferential statistics aren't required at SL.
My math teacher knew of the Chi square but he didn't know what inferential statistics was lol.
Standard deviation is supposed to be large for car speed estimations as all participants were shown the same car and accident.
But i guess that can be attributed to the inexperience of the participants in terms of driving.
Honestly, if the teacher isnt sure what I'm supposed to do for descriptive statistics, how is the student supposed to know?!

[CUB^3D]

Hm, that's not good. The teacher should know, the student shouldnt need to learn it by him/herself. Try to speak with your headmaster about it.

[Shahad]

Okay, so it turns out that standard deviation is required for descriptive statistics.
But honestly, it is a bit pointless in terms of the Loftus and Palmer study.
The speed estimations really do not matter.
They were just used to put in a question with different verbs.
But the speed estimations is the only thing one can do standard deviation for.
You cant really do descriptive statistics for yes or no answers about broken glass.

[Jyrgen]

lol I did mine on Loftus & Palmer too (HL) Seems like a popular topic.

AFAIK, the descriptive statistics are Mean, Mode, Median, and StdDev. I'd recommend showing both the Mean and the Standard Deviation.

You're right about the broken glass, if it's not possible, then don't do it. You have already fulfilled the statistics criteria with the speed estimations so move on

Edited by Jyrgen, Apr 26, 2012 - 11:06.

[Shahad]

Thanks for all the advice, really appreciate it.
Does anyone know how long the abstract is supposed to be?

[kim luffy]

if i'm not mistaken, around 150-200 words

[dechengirl]

Hi! wow, I am also starting my IA on Loftus and Palmer. I am not thinking of doing the second part of the experiment about the broken glass. So is it okay? I have started writing a bit, but havent really started the experiment. thanks

[Shahad]

It should be fine. IB psychology experiments are supposed to be simple.
Are you doing Psychology SL or HL?
Because at HL you are allowed to change the method of the original experiment.
I only did the broken glass part because that was how i measured the dependent variable, the memory recall of the pariticpant.
As long as you can justify the change you made appropriately, it should be acceptable.