# PSYB4 Research Methods

HideShow resource information
• Created by: dw23748
• Created on: 21-05-16 15:59
What is the conventional level of significance?
5% is the conventional and accepted minimal level of significance.
1 of 54
If the results of a statistical test show 5% significance then what does that mean?
There's only a 5% chance that the results are due to chance and therefore the hypothesis should be accepted.
2 of 54
How do you write down that the probability of the results occurring by chance is 5%?
p is less then 0.05
3 of 54
What do statistical tests test?
Statistical tests test hypotheses.
4 of 54
What does an alternative hypothesis predict?
An alternative hypothesis predicts that an effect will occur and that the results will be significant.
5 of 54
What's another name for an alternative hypothesis and when is this used?
Experimental hypothesis- used in the case of the study being an experiment.
6 of 54
What is a one-tailed hypothesis and what is another name for it?
A one-tailed hypothesis is also known as a directional hypothesis and it is a hypothesis which predicts which direction the results will take. For example, there will be a positive correlation between stress and the number of days taken off work.
7 of 54
What is a two-tailed hypothesis and what is another name for it? Give an example.
Also known as a non-directional hypothesis and this is a hypothesis that doesn't specify the direction of the results. For example, there will be a significant difference/relationship...
8 of 54
When to use a one-tailed hypothesis?
Evidence from previous research predicts the direction of the results.
9 of 54
What is a type 1 error and when/why does it occur?
Error of optimists. Accept the alternate hypothesis but should have rejected it. Too lenient level of significance (10%)
10 of 54
What is a type 2 error and when/why does it occur?
Error of pessimists. Reject alternate hypothesis but should have accepted it. Too strict level of significance (1%)/ use of two-tailed test.
11 of 54
What is meant by generalisation?
Refers to whether the results from the sample can be applied to the target population.
12 of 54
What are four sampling techniques?
Opportunity, random, stratified and systematic.
13 of 54
What do parametric tests do?
They make certain assumptions about the parameters of the population (e.g. mean or standard deviation) from which the sample is drawn.
14 of 54
Parametric tests are...? (Refer to 3 things).
Powerful, comprehensive (use all the data available) and robust.
15 of 54
What's the difference between parametric and non-parametric tests?
Non-parametric tests do not require the same assumptions about the population from which the sample is drawn.
16 of 54
Non-parametric tests are..? (4 points)
Quicker and easier to compute, capable of being used more widely, not as powerful (only use rank order of the scores), in need of higher number of scores/participants to match power of parametric test.
17 of 54
What are the three levels of measurement in order from lowest?
Nominal, Ordinal, Interval
18 of 54
What is meant by nominal data?
Data is simply categorised and can't be counted (frequency data), e.g. 10 smokers and 25 non-smokers. The categories are discrete- smoker or non-smoker.
19 of 54
What is meant by ordinal data?
Data is ordered but intervals between each value are unequal, e.g. the rank ordering finalists in race. The level of measurement should be used when data is based on estimates and not a universal scale.
20 of 54
What is meant by interval data?
Data is ordered and the intervals are equal, e.g. temperature. This type of data should be used if the data is based on a universal scale.
21 of 54
What is the criteria for a parametric test?
Interval data, sample data that are drawn from a normal distribution, samples must have similar variances (homogeneity of variance).
22 of 54
What three questions need to be asked when choosing a statistical test?
Is it a difference or relationship, what type of design (related/unrelated), what level of data was used?
23 of 54
Chi-square test
Nominal data, independent design, test of difference.
24 of 54
Sign test
Nominal data, repeated measures or matched pairs, test of difference.
25 of 54
Mann-Whitney test
Ordinal data, independent design, test of difference.
26 of 54
Wilcoxon signed ranks test
Ordinal data, repeated measures or matched pairs, test of difference.
27 of 54
Spearman's rank
Ordinal data, related design, test for relationship.
28 of 54
Unrelated t-test
Interval data, independent groups design, test for difference.
29 of 54
Related t-test
Interval data, repeated measures or matched pairs, test for difference.
30 of 54
Pearson's rank
Interval data, related design, test for relationship.
31 of 54
What are two strengths of the experimental method?
Can establish cause and effect between the IV and the DV, most scientific out of all methods- controlled procedures and objective measures mean that replication is possible and reliability can be checked.
32 of 54
What are two limitations of the experimental method?
Often lack ecological validity due to controlled setting, there can be problems with demand characteristics- participants often look for clues as to the expected behaviour.
33 of 54
What are three strengths of the questionnaire method?
Large amounts of data can be collected quickly, closed questions can be replicated quite easily, open questions can provide rich and detailed info.
34 of 54
What are three limitations of the questionnaire method?
Socially desirable responses, closed q's are susceptible to response set, rely on self-report data which may be biased.
35 of 54
What are two strengths of interviews?
Large amounts of data can be gathered, provides access to information not easily available by other methods.
36 of 54
What are four limitations of interviews?
Relies on self-report (subjective), difficult to establish cause and effect, information may be excluded, open q's may be difficult to summarise.
37 of 54
What are three strengths of correlation studies?
Show a relationship between 2 variables, enable predictions about increase/decrease in 1 variable over the other, useful where experimental manipulation would be unethical.
38 of 54
What are two limitations of correlation studies?
Can't show cause and effect, any relationship may be due to some other unknown variable.
39 of 54
What are two strengths of the observational method?
High in ecological validity, useful where experimental manipulation would be unethical.
40 of 54
What are four limitations of the observational method?
Can't infer cause and effect, prone to observer bias, difficult to replicate, if people know they're being watched they'll change behaviour or if they don't know then ethical issues arise.
41 of 54
What are two strengths of content analysis?
Few ethical issues as participants aren't directly involved, it allows the researcher to study a wide range of material.
42 of 54
What are three limitations of content analysis?
Behaviour studied out of context so could be misinterpreted, categories are decided in advanced and are based on researcher's expectations, qualitative content analysis are open to interpretation.
43 of 54
What are four strengths of the case study method?
Data is rich in detail, high validity as it relates to real life, a single case can be used to challenge a theory, may be the only way to study rare or unusual behaviour.
44 of 54
What are five limitations of the case study method?
Results can't be generalised, often involves retrospective data (unreliable), close relationship between researcher and participant may introduce bias, cause and effect difficult to establish, not based on rigorous methods;replication is not possible
45 of 54
What are four strengths of qualitative data?
Useful in generating new theories, high face validity as they focus on meaningful issues, richer and more detailed, behaviour can be studied on context.
46 of 54
What are four limitations of qualitative data?
Difficult to replicate, data analysis is difficult, difficult to establish cause and effect, subjective and low in reliability.
47 of 54
What are three strengths of quantitative data?
Easily summarised in graphs or statistics, easy to replicate, objective data so easier to analyse.
48 of 54
What are two limitations of quantitative data?
Low ecological validity, less meaningful than qualitative data.
49 of 54
What is inter-observer reliability and how should it be assessed?
Multiple researchers asses the results by using the same observers schedual and comparing the results. If there is a strong positive correlation between the results then they have high inter-observer reliability.
50 of 54
What is test/retest reliability used for and how should it be assessed?
Used for questionnaires and it is assessed by giving the same questionnaire to the participants a while later. If strong +ve correlation then high test/retest reliability.
51 of 54
What is face validity used for and how should it be assessed?
Used for questionnaires and it is assessed before giving q's to p's. Read the q's and check to see if they look like they measure what they're supposed to measure. If they do then the questionnaire has face validity.
52 of 54
What is concurrent validity?
Involves correlating results on a test with results on an already established test of the same measure. If the new test is valid, a strong positive correlation should be found.
53 of 54
What are the 7 ethical issues that you should consider whenever planning research?
Confidentiality, informed consent, privacy, deception, right to withdraw, protection from harm, debriefing.
54 of 54

## Other cards in this set

### Card 2

#### Front

If the results of a statistical test show 5% significance then what does that mean?

#### Back

There's only a 5% chance that the results are due to chance and therefore the hypothesis should be accepted.

### Card 3

#### Front

How do you write down that the probability of the results occurring by chance is 5%?

### Card 4

#### Front

What do statistical tests test?

### Card 5

#### Front

What does an alternative hypothesis predict?