No P Value given - can we form a conclusion

[psychlaw]

Can we form a conclusion with no P Value given?

This was an IARTV question from the 2008 exam. A study, to test the effect of alcohol consumption on driving ability was conducted using a driving simulator. University sudents from two Research Methods classes were randomly allocated into two groups and then tested.

- Class A, of 20 third year uni students consumed more than two alcoholic drinks in an hour
- Class B of 20 third year uni students consumed a number of non-alcholic drinks over an hours period
- After three hours, both groups of participants used the driving simulator
- All errors were recorded andused to measure the driving skill of each participant

Results:
- Class A had an average number of 20 driving errors
- Class B had an average number of 5 driving errors




Question: Can the researcher conclude that alchohol consumption causes impairment in driving ability? Explain.



Also, is this an appropriate operational hypothesis for this question

It was hypothesised that third year university students would have a higher average number of driving errors if they consumed more than two alcoholic drinks in an hours, then third year university students who consumed a number of non-alcholic drinks in an hour
Can anyone give me any tips if I'm doing anything wrong


thanks heaps

[Lulu]

No conclusion can be formed as we don't know whether or not the difference in those scores for both classes is a significant difference, that is, whether it is due to the IV rather than chance.

Only thing I would add to your hypothesis is some like "as measured by a driving simulator", other than that its fine

[arthurk]

yeah im confused at this myself
if there are no p values can we give a conclusions, this sorta thing was found in a previous exam before the current study design where a p value was not given, due to my learning of the current study design i thought "you only gave me descriptive statistic how can i make a conclusion" then the answers are like it can be concluded that... and i was like <.< but i asked the teacher this and she was unsure but she assured me that the current exams will not contain only descriptive statistic as there should be a big research investigation with all the question in research investigation section devoted to it, and itll ask you if conclusions can be made given a p value etc.

and for your hypothesis i dont think saying 'average' is necessary. besides that it seems good

also ive seen some questions where the sample is representative and the p value is significant but there are extraneous variables that made the answers begin with no conclusions can not be made etc. because of this extraneous variable blah
so what amount of extraneous variables will cause an experiment to be unable to make conclusions based on results?

[psychlaw]

No conclusion can be formed as we don't know whether or not the difference in those scores for both classes is a significant difference, that is, whether it is due to the IV rather than chance.

Only thing I would add to your hypothesis is some like "as measured by a driving simulator", other than that its fine

I thought it was no conclusion can be formed, but the answers says yes, you can form a conclusion...[

[arthurk]

yeah its gay, ignore it, teacher says there wont be something like that on our exam

[melaniej]

say "even though the results indicate BLAH BLAH, no conclusion can be....etc etc

[Lulu]

I've seen answers which do a sort of thing like...

It is likely that the hypothesis was supported as class A had a high average score than class B. However without statistical analysis of the data, it is not possible to determine whether the observed differences are statistically significant.

That's from a question which is a similar situation like your one, its from the NEAP smartstudy book.

I think you'd be safe if you put something like that down on tomorrow's paper

[psychlaw]

Official answer given:

"For one mark, students need to answer 'yes'.
For another mark, students need to explain that the results of the research demonstrated that alcohol increased the number of driving errors and that a generalisation can be made, based on the research conducted, that consuming alcohol can have a negative affect on one's driving ability".

[Nick]

Obviously in this case, a conclusion based on statistical evidence cannot be made. Despite this, given the overwhelming trend in the results towards alcohol instigating more errors, it is possible to make a conclusion by generalising the results.

You wouldn't get a question like this on the exam so don't even worry about it.

[arthurk]

psych 45 nick
nice work
thats my goal. do you know how high ur A+'s were for exams to get that mark