Research methods- Inferential statistics Flashcards
What are inferential statistics used for?
Used to determine the likelihood that an ‘observed effect’ is due to chance
What is probability?
The chance of something occurring
Why is probability important in psychology?
Interested in finding out if studies show real differences or relationships within the results or if they are due to chance factors. To establish this, a significance needs to be found.
What is a level of significance?
Whether results are significant and not due to chance factors, we use a level of significance. Expressed in bullet point values (e.g. P<0.00 - less than or equal to)
What significance is appropriate and why?
5% significance is appropriate
If P< 0.05 then alternative hypothesis is accepted and null is rejected
If P>0.05 the alternative is rejected and null is accepted
5% means probability of results occurring by chance are less than or equal to 5% (or 5 in 100) - 95% certain results are due to variables we are studying rather than chance
What is P<0.01?
Means that the level of significance would be 1% or less than 1%
Used when challenging well-established theory as it causes less extraneous variables as well as allows researcher to come to a more certain set of results
What is a Type 1 error?
When the levels of significance are too lenient causing ‘false positives’. Researcher would reject null hypothesis and accept research hypothesis
What is a Type 2 error?
When significance levels are too strict causing ‘false negatives’.
Researcher would accept null hypothesis and reject research hypothesis
What is the structure when choosing a test?
1-Association or difference
2-Related (repeated measures/ matched pairs) or unrelated (independent groups)
3-Level of measurement- Nominal (categories), Ordinal (ranked), Interval (universal measurements)
4-Name the test
What is a One-Tailed test?
-Directional hypothesis (predicts direction)
-Key words that indicate more, less, higher, lower
-Correlation, negative/ positive indicate direction
-Previous research
What is a Two-Tailed test?
-Predicts an effect but no specified direction of affect (Non-Dictional)
-If 5% probability is used, double the probability that differences could occur by chance
-Inconsistent previous research
How are the measures of central tendency linked to the levels of measurement and what are the limitations?
Nominal= Mode (least sophisticated as people just placed into categories)
Ordinal= Mean (based on opinion+ subjective judgement so not precise)
Interval= Median (very precise- more detail is preserved)
Which tests are more than or equal to and what are the 3 elements?
1) Related t test (parametric)- Difference, Related, Interval
2) Unrelated t test (parametric)- Difference, Unrelated, Interval
3) Chi Square test- Association, Unrelated, Nominal
4) Spearman’s rank order test- Association, Related, Ordinal
5) Pearson’s test(r) (parametric)- Association, Related, Interval
Which tests are less than or equal to and what are the 3 elements?
1) Wilcoxon signed ranks test- Difference, Related, Ordinal
2) Mann-Whitney U test- Difference, Unrelated, Ordinal
3) Sign test- Difference, Related, Nominal
What are the criteria for a parametric test and what does this mean?
Standard
1) Must be Interval
2) Score distribution must be normal or expected to be normal (Bell curve)
3) Variances should be homogenous (spread must be similar)- as seen in related design
How to work out a Sign test?
1- Work out sign of difference (+ and - signs)
2- Add up + and - symbols
3- Work out S value (the lowest of the two numbers found when adding)
4- Is there a significant difference? (Use critical values given). N value is number of participants- =. One tailed or Two = 0.05 significance
5- If calculated value is larger than critical, not significant.
S value < Critical value
What are the degrees of freedom for the other tests?
Related df= N-1
Unrelated df= NA+ NB- 2
Spearman’s Rho df= number of participants
Pearson’s r df= N-2
