Maths bit Flashcards
1
Q
When should the sign test be used?
A
- Needs to look for a difference rather than correlation.
- Data should be nominal (organized into categories)
- Needs to be related design (repeated measures)
2
Q
How would I work out the sign test?
A
-
Step 1
Convert the data to nominal data by working out whether the scores turned out lower or higher.
Do this by subtracting the score of the 2nd test from the score on the 1st test.
If the answer is negative, record a negative sign, if the answer is positive: record the plus sign. -
Step 2
From a table, add up the pluses and minuses.
Take the less frequently appearing sign and call this S value
! Remember, if the values are the same - ignore the data and adjust the N (total number) -
Step 3
Compare calculated value with the critical value.
Determine if the hypothesis is one or two tailed
Use 5%/0.05 level of significance, unless told otherwise.
Determine the row by N (number of participants). - The calculated value of S MUST BE equal to or less than the critical value at 0.05 level of significance.
3
Q
Why are statistical tests used in psychology?
A
- They are used to determine whether a significant difference or correlation exists
- This tests their hypothesis, allowing to see if the results have occurred by chance or otherwise.
- This allows us to accept or reject the null hypothesis
4
Q
What 3 things decide what statistical test to use?
A
- Whether looking for a difference or correlation
- Type of data (nominal, ordinal or interval)
- Level of measurement/experimental design (related or unrelated)
5
Q
What designs are related in psychology and which design is unrelated?
A
- Related: matched pairs and repeated measures
- Unrelated: Independent groups
6
Q
Why is a matched pairs design considered a related design?
A
Even though participants in each condition are not the same, they have been ‘matched’ on some variable that makes them ‘related’.
7
Q
When should a sign test be used? (practise on whiteboard)
A
- When it is looking for difference
- When the data is nominal
- When the design is related
8
Q
When should a chi-squared test be used? (practise on whiteboard)
A
- It can look for either an association or difference
- Data is nominal
- Design is either unrelated or just a correlation
9
Q
When should a Wilcoxon test be used? (practise on whiteboard)
A
- When it is looking for a difference
- When the data is ordinal
- When the design is related
10
Q
When should a Mann-Whitney U test be used? (practise on whiteboard)
A
- When looking for a difference
- When data is ordinal
- When the design is unrelated
11
Q
When should Spearman’s rho be used?
A
- When looking for a correlation
- When the data is ordinal
12
Q
When should the related t-test be used?
A
- When looking for a difference
- When data is interval
- When the design is related
13
Q
When should the unrelated t-test be used?
A
- When looking for a difference
- When data is interval
- When the design is unrelated
14
Q
When should Pearson’s r be used?
A
- When looking for a correlation
- When data is interval
15
Q
What phrase can help remember inferential tests?
A
Simon (Cowell (2x)) Wants More Singers Receiving Unanimous Praise.
16
Q
What is meant by nominal data?
A
Data represented in the form of categories.
It is discrete as only one item can appear in one of the categories
17
Q
What is meant by ordinal data?
A
- Data ordered in some way.
- It is usually measured with rating scales
- It DOES NOT have equal intervals between each unit (e.g: someone rating 8 does not enjoy something twice as much as another who rated 4).
18
Q
Criticise ordinal data
A
- Ordinal data lacks precision as it is based on subjective opinion as opposed to objective measures.
- As a result, ordinal data is not used and it is the ranks - not the scores - used in calculations. (1st, 2nd, 3rd).
19
Q
What is meant by interval data?
A
- Data based on numerical scales
- Units have an equal, precisely defined size
20
Q
What type of data is considered parametric?
A
Interval data
21
Q
What is another name for a one tailed hypothesis?
A
Directional hypothesis
22
Q
What is another name for a two tailed hypothesis?
A
Non-directional hypothesis
23
Q
What is a significance level?
A
- The point at which a researcher can claim to have discovered a large enough difference or correlation.
- So they can reject a null hypothesis and accept an alternative hypothesis
24
Q
What is the usual level of signficance in psychological research?
A
0.05. (5%).
25
If P ≤ 0.05 what does this mean?
- This means that the probability that the result occurred when there is NO effect in the population is equal or less than 5%
- This means that if a researcher claims to have found a significant difference/correlation, there is still up to a 5% chance it isn't true.
26
What is the calculated value?
The result from the test which is compared to the critical value
27
What is the critical value?
A number from a table that informs whether we should reject the null hypothesis and accept the alternative hypothesis or not.
28
What 3 things are needed when using a table of critical values?
- Whether the hypothesis was one or two tailed
- Number of participants in the study (N value): this can work out the degrees of freedom
- The level of significance (P value) - usually 0.05 unless informed otherwise.
29
When would a more stringent level of significance (p value) be used - such as 0.01?
When the study may have *human cost* - such as drug trials - or 'one-off' studies that could not be repeated in the future.
- In all research, if there is a *large* difference between calculated and critical values - in the preferred direction - more stringent levels would be used as the *lower the P value, the more statistically significant the result*
30
What is a Type I error?
- When the null hypothesis is rejected and the alternative hypothesis is accepted (false positive)
31
What is a type II error?
- When the null hypothesis is accepted when the alternative hypothesis should've been accepted (false negative).
32
When are we more likely to make a Type I error?
- When the significance level is too lenient (high) - such as 0.1 (10%) rather than 5%.
33
When are we more likely to make a Type II error?
- When the significance level is too stringent (low) - such as 0.01 (1%) - this is as potentially significant values may have been missed?
34
Why do psychologists favour 0.05 (5%) significance level?
It best balances the risk of making a Type I or Type II error
35
When is a graph considered to have a *normal distribution*
When the mean, median, and mode all occupy the same midpoint of the curve.
36
What is important in graphs with curves?
- The 'tails' never touch the horizontal x-axis (never reach 0), as more extreme scores are always theoretically possible.
37
What are the features of a positive skew?
- Where most of the distribution is concentrated towards the left of the graph
- There is a resulting long tail on the right: in a positive direction
38
What are the features of a negative skew?
- Most of the distribution is concentrated towards the right of the graph
- There is a resulting long tail to the left: in a negative direction
39
What are the position of measures of central tendency on a *positive skewed graph*?
- The mode is always on the highest peak
- The median is always between the mean and mode
- The mean is dragged towards the right 'tail' (due to higher scorers pulling mean to the right)
40
What are the position of measures of central tendency on a *negative skewed graph*?
- The mode is always on the highest peak
- The median is always between the mean and mode
- The mean is dragged towards the left 'tail' (due to lower scorers being in the minority).
