Discovering statistics using IBM SPSS statistics (4th ed) London SAGE. Purchase the gradpack bundle.
Check or save the dates for the numeracy centre and PAL.
Pracs start next week, must answer questions prior to attending the prac. Must have a set of proposed answers. Strongly recommend revising first year notes.
They are due every Sunday night.
Recommendations to immerse yourself in SPSS, use the resources that are availble online.
There's an optional test, if you do better in the optional test these results will replace your final exam mark. Exams are also open book.
- Operate SPSS
- Interpret it using SPSS
z tables pp. 887-892
t tables pp. 893
f tables pp. 894-897
Bonferoni- adjusted f table
You design a study which is applicable to your research question. If your design is flawed the stats aren't going to be able to save you.
The best kind of research has a theory driven research question, which is either in favor or to disapprove a theory. It is rare to find a good theory based on a 'I jsut wanted to find this out.' Keep in mind that a theory which is not falsifiable is not scientific.
True (T) versus quasi (Q) experiments
(T): the potential to randomly allocate observations to conditions.
But what if this was the research question?
(RQ) Is there a difference between genders in the motor skills between genders?
The experiement is quasi if the separation of conditions is based on naturally occuring groups.
Between group design vs repeated measures
(RQ) Does spaced practice (5 hrs broken up) or mass practice (5 hours at once) affect performance?
(iv) The independent variable is practice schedule
(dv) level of motor skill.
The dependent variable isis not the number of balls they sucessfully get. This is becuase if a variable has 1 level it's a constant not a variable.
(RQ) Is there a difference between (m)/(f)'s?
(dv) in order for the exp to be good you want to isolate the relationship between the (iv) + the (dv). Once you do this you're permitted to make cause and effect statements.
Cause and effect statements
should be limited to true experiments.
Any other variable that isn't the (iv)/(dv) is called an extraneous variable.
Extraneous variables (ev)'s
They're always there, they are those awkward variables that hang around even if you're not testing them. Criteria for an (ev) is:
1. Correlated with the (dv)
2. Must vary systematically across the levels of the (iv)
(m) in the spaced condition
(f) in the mass condition
This means that the
(iv) has two levels, spaced and massed.
(dv) motor skill (by gender)
then what is the confounding variable?
The reason it's confounding is because the results are ambigous.…