Flashcards in Quant: Lecture 4 Deck (11):
Is a research question conceptual or operational?
What about a hypothesis?
What's the aim of a hypothesis?
Should you use a one-tailed or two-tailed hypothesis?
To state the expected result and to indicate the cause and effect.
Two-tailed, unless you have research two back up a one-tailed hypothesis.
What is the aim of a null hypothesis?
To predict no effect between two variables and any effect is due to random error. It's a statistical throwback.
When testing for significance, what does the expected hypothesis predict?
It predicts that the results won't be in the same scale as the normal distribution.
What is the hypothetico-deductive method?
It's when you check whether your results are due to random error or not by calculating a statistical test. This finds the probability of whether the results are due to sampling error.
How do you calculate probability?
Number of ways an event can occur divided by the number of possible outcomes. For example, flipping a coin for heads: there is one way the event can occur but there are two possible outcomes. 1/2=0.5. Probability will always be between 0 and 1 and the smaller the value, the less likely the event will occur.
What is the number that psychologists use as a limit for random error explaining results?
0.05. If it's larger than this then random error explains the results and they aren't statistically significant. Then we would accept the null. Non-significant results can still be important.
Can results ever be 100% confident?
No, because you use a sample, not the whole population, so the results are only supported, not proven.
List one disadvantage of an unrelated sample and list the solution for it
Do the same for a related sample
Individual differences as different people are in each condition.
Solution: Randomly allocate.
Order effects, e.g. fatigue or learning
When should you use a matched-pairs design?
When shouldn't you use a related sample?
When a related sample would reveal the aim of the study.
When the pre-condition can't be identified or caused, e.g. the effects of car accidents on depression.