Psychology Research Methods

?
  • Created by: Madeline
  • Created on: 06-06-14 12:17
Type One Error
Null wrongly rejected and alternative accepted as real because significance level too leniant.
1 of 54
Type Two Error
Null wrongly accepted and alternative rejected because significance level too stringent.
2 of 54
Why is 5% the accepted significance level?
Strikes a balance between making Type 1 and Type 2 errors.
3 of 54
What two things do correlation statistics express about the relationship between two variables?
The strength of the relationship and the direction.
4 of 54
A poitive correlation?
As one variable increases, the other variable increases.
5 of 54
A negative correlation?
As one variable increases, the other variable decreases.
6 of 54
A zero correlation?
There is no relationship between the two variables, either positive or negative.
7 of 54
What does the generalisability of results refer to?
Whether the results from a sample can be applied to the target population from which the sample is taken.
8 of 54
How to reduce bias in samples?
Avoid small samples and select a sampling technique that best represents the target population.
9 of 54
Opportunity sample?
Participants are selected on the basis of their availability.
10 of 54
Limitations of an opportunity sample?
It is unlikely to be representative due to researcher/selector bias and participants may be a narrow group of very similar people making the sample unlikely to be representative, thus difficult to generalise.
11 of 54
Strengths of an opportunity sample?
Quick and easy as the sample already exists.
12 of 54
Random sample?
each member of the population has an equal chance of being selected to take part.
13 of 54
Limitations of a random sample?
Its time consuming and could still be unrepresentative.
14 of 54
Strengths of a random sample?
There is no experimentor bias in who takes part as every person has an equal chance.
15 of 54
Stratified sample?
The researcher identifies subgroups in order to create a sample that represents the target population. They then choose randomly from each proportion.
16 of 54
Limitations of a stratified sample?
Very time consuming and complicated. Its sometimes difficult to identify the sub-groups. By chance, all the key characteristics of the target population may not be idenitified or represented in the sample.
17 of 54
Strengths of a stratified sample?
The sample should be representative of the population.
18 of 54
Systematic sample?
Every nth person is selected by the researcher.
19 of 54
Limitations of a systematic sample?
The sample may not be representative. It may not be quite as unbiased as random sampling as the researcher has some control over the selection process.
20 of 54
Strengths of a systematic sample?
It's simple to do and the researcher is guerenteed the population will be evenly sampled.
21 of 54
What should be considered when determining the type of statistical test to use?
The purpose of the investigation, the type of design used and the level of measurment.
22 of 54
What do we mean by the purpose of the investigation?
Whether it is an experiment or correlation.
23 of 54
What do we mean by the level of measurment?
Whether it is nominal, ordinal or interval data.
24 of 54
What is nominal data?
Counting frequency data. Use if the stem says 'tally' or 'catagory'.
25 of 54
What is ordinal data?
Ranking data into order, with each value being greater/larger/better than another. More informative than nominal data. Not standardised. Use if the stem says 'rank'.
26 of 54
What is interval data?
The most accurate form of measurment as it uses equal measument intervals. Any form of standardised measurment, e.g. time, weight, temperature and distance, is interval.
27 of 54
What is ratio data?
If an interval scale starts at zero it is called a ration scale, e.g. time. if unsure simply say 'data should be TREATED at least as interval'. But be able to justify why.
28 of 54
Which measure of central tendancy is sppropriate to use for nominal data?
Mode.
29 of 54
Which measure of central tendancy is sppropriate to use for ordinal data?
Median.
30 of 54
Which measure of central tendancy is sppropriate to use for interval data?
Mean.
31 of 54
What does 'Silly' stand for?
Sign test.
32 of 54
What does 'Wonka' stand for?
Wilcoxon signed ranks test.
33 of 54
What does 'Can't' stand for?
Chi-squared test.
34 of 54
What does 'Make' stand for?
Mann-Whitney test.
35 of 54
What does 'Caster' stand for?
Chi-squared test.
36 of 54
What does 'Sugar' stand for?
Spearmans Rank Correlation Coefficient.
37 of 54
What does 'Really' stand for?
Related t-test.
38 of 54
What does 'Is' stand for?
Independant t-test.
39 of 54
What does 'Pitiful' stand for?
Pearson's product moment correlation.
40 of 54
What's the standard line used once you've decided upon your statistical test?
The observed value of...is less than the critical value of....in a ... tale test, therefore we can accept the ...hypothesis and reject the ... hypothesis.
41 of 54
Mnemonic for parametric tests?
Really Cool Product.
42 of 54
Why are parametric tests described as 'powerful'?
They are better able to detect a significant effect as they are calculated using actual scores rater than ranked scored (interval rather than nominal or ordinal). However, this is a problem id the data is inconsistant or erratic.
43 of 54
The 3 strengths of paramentric tests?
Powerful, Comprehensive, Robust.
44 of 54
Why are parametric tests described as 'comprrehensive'?
They use all of the infomration available.
45 of 54
Why are the parametric tests described as 'robust'?
It is possible to do a parametric test on data which do not fit the assumptions precisely because they are able to cope with data that do not fully comply with the assumptions (the only essential criteria is that the data be interval).
46 of 54
The first criteria for parametric testing?
Data muct be at least interval.
47 of 54
The third criteria for parametric testing?
The variances should be homogenous - the spread of scores in each condition should be similar.
48 of 54
The second criteria for parametric testing?
The distribution of scores in each condition should be drawn from a population which would be expected to show a normal distribution (bell shaped) for whatever variable is being measured.
49 of 54
First feature of non-paramentric tests?
Simpler and quicker to compute than parametric tests.
50 of 54
Second feature of non-parametric tests?
Capable of being used more widely.
51 of 54
Third feature of non-parametric tests?
Not as powerful as parametric tests - they only use the rank order of the scores.
52 of 54
Fourth feature of paramentric tests?
They are in need of a higher number of scores/participants in order to match the power of parametric tests.
53 of 54
Sign Test
Converts raw data into catagories. There is no information about the size of the differences only the direction of the differences.
54 of 54

Other cards in this set

Card 2

Front

Type Two Error

Back

Null wrongly accepted and alternative rejected because significance level too stringent.

Card 3

Front

Why is 5% the accepted significance level?

Back

Preview of the front of card 3

Card 4

Front

What two things do correlation statistics express about the relationship between two variables?

Back

Preview of the front of card 4

Card 5

Front

A poitive correlation?

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all Research methods and techniques resources »