Cream Pack 3 Flashcards
Why is the level of measurement important?
Knowing the level of measurement helps you decide what descriptive statistics and which inferential statistical test is appropriate for your
data.
What are the 3 levels of measurement?
- nominal
- ordinal
- Interval
What is nominal level data?
Frequency counts in mutually exclusive categories.
Give an example of nominal level data
A researcher recording the frequency counts of people’s favourite colours.
What is ordinal level data?
Ordering data in a certain way. However the exact difference between each point on the scale is not known.
Give an example of ordinal level data
Ranking people’s test scores.
We know the position but not exactly how much better or worse each person is
What is interval level data?
Numbers that come from a scale of equal intervals.
Give an example of interval level data
Measuring people’s height in cm
Which is the three levels of measurement contains the most information?
Interval data
Which is the three levels of measurement contains the least information?
Nominal data
The number of words recalled by participants in each of two conditions: when they are given 30 seconds or
45 seconds to memorise them.
What level of measurement are implied in the data collected?
Ordinal or interval
The time taken in seconds for participants to read a script which is handwritten as opposed to typed.
What level of measurement are implied in the data collected?
Interval
A dental receptionist is asked to judge whether people in the waiting room are ‘calm’ ‘anxious’ ‘very anxious’ or ‘hysterical’
What level of measurement are implied in the data collected?
Nominal
Distance two participants stand apart when asked to take part in an intimate conversation, measured from photos.
What level of measurement are implied in the data collected?
Interval
What is ratio level of measurement?
A ratio scale of measurement is simply an interval scale in which it is impossible to get a negative value.
E.g. it is impossible to have a negative reaction time.
How can the levels of measurement be converted?
Interval data can be converted into ordinal or nominal data.
Ordinal data can be converted into nominal.
Every time we convert information is lost.
Nominal level data is made up of data in ‘mutually exclusive” categories. What are mutually exclusive categories?
Categories that do not overlap and only mean one thing.
Your friend argues that, since she came top in each of the three maths tests held in her class this year, she must be far better than all the other students. What might you point out to her?
It is ordinal data so the exact difference between her scores and the students below her is not known.
What is probability?
a numerical measure of chance.
It represents how likely it is that the results are due to chance.
Why do researchers use inferential statistical tests?
Researchers use inferential statistical tests to help them assess the probability that the difference or correlation they observed in the
results is statistically significant (due to chance) or not.
i.e. whether to accept or reject the null hypothesis
Name 4 inferential statistical tests
- Chi - Squared
- Wilcoxon Signed ranks
- Mann- Whitney U
- Spearman’s ho correlation coefficient
What is the traditional level of significance chosen by psychologists?
(p ≤ 0.05) 5% level
OR
95% significant
What does the (p ≤ 0.05) significant level mean?
The probability (p) of the results occurring due to chance is less than or equal to (≤) 5% (0.05).
What would happen if the probability of the results occurring due to chance is greater than the 5% level?
Researchers would accept their null hypothesis and state that any difference or correlation seen in the results was due to chance.
What is the significance level?
The probability at which the null hypothesis can be rejected
State what the observed value means
The value calculated by the researcher themselves
State what the critical value means
Pre-calculated table of values
What are the 5 statements of significance?
- the observed and critical values used and which was bigger
- the sample size
- the significance level used
- whether the one or two tailed test was used
- which hypothesis was accepted
When does a researcher choose the 10% level? p≤0.1
This is used when a researcher does
not need a strict level of significance
e.g. when beginning a new area of
research
When does a researcher choose the 1% level? p≤0.01
-This is used when a researcher
needs a very strict level because any
mistakes in the research could have
serious consequences
What is the most strict level of significance?
p≤0.01
What is the least strict level of significance?
p≤0.1
What is a Type 1 error?
Accepting the alternative hypotheses when the results are actually due to chance. (false positive)
Say there is an effect when there isn’t one.
What is a Type 2 error?
Accepting the null hypothesis when the results were not due to chance. (false negative)
You say the is not an effect when actually there is.
What type of error is a researcher more likely to make using the 10% significance level?
Type 1 error
What type of error is a researcher more likely to make using the 1% significance level?
Type 2 error
What does p>0.5 mean?
The probability of results being due to chance is more than 5%
If results were found to be not significant at p<0.05 which hypothesis would have to be accepted?
Null hypothesis
How do researchers judge the significance of their results?
Using inferential statistical tests then compare the observed value from the critical value
What effect does a one-tailed test compared to a two tailed test have on whether the results are judged to be significant?
On a statistical basis it is easier for a one tailed test rather than two tailed as you know the exact direction.
What are non-parametric tests of significance?
These are the statistical tests carried out when dealing with ordinal or nominal levels of measurement.
What are the 3 decisions needed to be made when selecting an appropriate statistical test?
1) What are you testing for? (difference or correlation?)
2) Level of measurement (nominal or ordinal?)
3) Unrelated (independent groups) or Related data? (repeated measures + matched pairs)
Table for Non-parametric test and their conditions for use
Practice justifying your choice of test.
State the test that should be chosen.
And give reasons. (Linked to the three decisions) explaining detail why.
