Statistics Flashcards
What is the mnemonic for remembering the statistical tests?
Carrots ━━> Chi squared
Should ━━> Sign test
Come ━━> Chi squared
Mashed ━━> Mann-Whitney U-test
With ━━> Wilcoxon
Swede ━━> Spearman’s rho
Under ━━> Unrelated t-test
Roast ━━> Related t-test
Potatoes ━━> Pearson’s r
When is the sign test used?
When looking for a difference between PAIRED data.
E.G.
repeated measures design
matched pairs
which generates NOMINAL data.
What are the steps of carrying out a sign test?
1 = Is the hypothesis directional or non‐directional?
2 = Work out the sign.
Record each pair of data with a + or – (depending on whether the difference is positive or negative). If there is no difference ‘0’ is recorded.
3 = Calculate the value of S
(calculated by adding up the total number of pluses and minus and selecting the smaller value)
4 = Calculate the value of N
(N is the total number of scores, minus any nil scores ‘0’)
5 = Find the critical value of S (see table below)
6 = Determine whether the results are significant or not
(typically, the p ≤ 0.05 probability level is used unless otherwise stated)
(calculated S value must be equal to or smaller than the critical S value)
7.Reporting the conclusions of the sign test.
Once you have worked out the results of the sign test, you can draw a conclusion.
What are the factors affecting the choice of a statistical test?
1 > Test of difference or test of association
2 > Experimental design - related or unrelated
3 > Levels of measurement (type of data) - nominal, ordinal, interval
How would you be able to tell the difference between a test of difference or test of association?
A test of difference will include an experiment - having a control and an experimental group.
A test of correlation will include co-variables, simply asking pps instead of making them do tasks.
Which experimental designs fall under ‘related’?
Repeated measures (pps are the same)
Matched pairs (pps are very similar)
Which experimental designs fall under ‘unrelated’?
Independent groups design (pps are not the same/similar)
What is meant by related data?
Data in which participants in each condition are related or similar in some manner.
What is meant by unrelated data?
Refers to having two separate groups of people in each condition of the study who are not related or similar in some manner.
What are the types of quantitative data and levels of measurement?
Nominal
Ordinal
Interval
What is nominal data?
CATEGORICAL data.
Each pps will only appear in 1 category - this is called discrete data.
Cannot appear in both.
E.G.
Researcher wants to know if more students doing A-level psychology went to a school or a college - data would be categorised as either ‘school’ or ‘college’.
What is ordinal data?
ORDERED data.
Used to simply rank data.
Intervals between data are not equal.
Values assigned have no meaning beyond the purpose of stating where one scored appeared in relation to others.
Usually collected from self-report measures and data is subjective.
E.G.
If people were asked to rate their preference of local restaurants on a scale of 1-10 with 10 being their favourite.
What is interval data?
Includes units of measurements (seconds, mm, degrees)
Intervals between values are equal in measurement.
More objective and scientific in nature.
E.G.
Temperature and time.
The difference between 3 and 4 degrees Celsius is the same as the difference between 35 and 36 degrees Celsius.
How would you summarise the differences between the types of quantitative data / levels of measurement?
Nominal > categorical (girl, boy)
Ordinal > ordered (ratings, Likert scale)
Interval > measurements (degrees, mm)
What is meant by primary data?
Data collected directly first-hand from the pps for the purpose of THIS study.
What is meant by secondary data?
Data which was collected second-hand from ANOTHER purpose/different study and then used in a new study.
What are the 3Ds?
D = test of Difference
D = Design
D = levels of Data
What is the Rule of R?
If the test has an R in the name, the observed value HAS to be moRe than the critical value for the results to be significant.
What are the common levels of significance used?
5% = 0.05
1% = 0.01
When would a 1% significance level be used? [2 marks]
Any study that is CONTROVERSIAL or SENSITIVE in nature.
We use 1% because we want to be 99% sure that the results are due to difference and not chance.
E.G. controversial studies include…
effect of race on IQ levels.
What could using a 1% or a 10% level of significance risk causing?
If the LOS is too strict (set a 1%) = Type II error (false negative)
If the LOS is too lenient (set at 10%) = Type I error (false positive)
What is meant by a Type I error?
FALSE POSITIVE
When the alternative hypothesis is wrongfully accepted when actually results are due to chance so the null should have been accepted.
Accepted alternative.
Meant to accept null.
What is meant by a Type II error?
FALSE NEGATIVE
When the alternative hypothesis is wrongfully rejected when actually it was a real result and so the alternative should have been accepted and null is rejected.
Accepted null
Meant to accept alternative.
What is meant by the level of significance?
The level of probability at which it is agreed that the null hypothesis will be rejected and the alternative will be accepted.
What are the 2 types of statistical tests?
Parametric tests
Non-parametric tests
What is a calculated value?
The number that is generated from a statistical test that is ALREADY given to you (unless its a Sign test).
The calculated value is compared with the critical value (that you have to work out using the table).
What is a one-tailed test?
A test with a directional hypothesis.
Tests for the possibility of an effect in one direction.
What is a two-tailed test?
A test with a non-directional hypothesis.
Tests for the possibility of an effect in two directions—positive and negative.
How would you know whether to use a one-tailed or two-tailed test?
One-tailed test > directional hypothesis
Two-tailed test > non-directional hypothesis
When can a parametric test be used?
Only when INTERVAL data is collected -
Related/Unrelated t-tests and Pearson’s r
And when there is a NORMAL distribution.
How would you calculate the degrees of freedom in a related test?
Sample size - 1
E.G.
If N=18…
df = 17
How would you calculate the degrees of freedom in an unrelated test?
Sample size - 2
E.G.
If N=18…
df = 16
Remember it is 2 because there are 2 letters before ‘related’
