Research methods

  • Created by: Ellie
  • Created on: 08-06-14 16:05

Types of data

Nominal data: data is in categories e.g. how many black people and how many white people are there in the army?

Ordinal data: data placed into a scale or put in rankings e.g. 1st 2nd 3rd, low/medium/high. Often subjective

Interval/ratio data: data with standardised mesurements - equal amounts in between e.g. minutes, centimetres, pounds etc data is objective

1 of 10

Inferential statistics

Inferential tests: allow psychologists to draw conclusions from their findings. They calculate the probability that the results have arisen by chance. If the findings are unlikely to be a fluke then they are called significant. The tests include Chi squared (X2), Spearman's Rho (R), Wilcoxon (T), and Mann Whitney (U).

Observed value: the value calculated by the inferential test

Critical value: the value found in the table of critical values

Degree of freedom: (number of rows -1) x (number of columns -1)

How do you know the results are significant or not?

Spearman's Rho and Chi squared = the observed value has to be greater than the critical value

Wilcoxon and Mann Whitney = the observed value has to be less than or equal to the critical value

2 of 10

Inferential statistics continued

P = probability the results are down to chance. Researchers choose the maximum probability that there results are down to chance that they are still willing to accept as being significant. Usually it is set at p<_ 0.05 meaning they are allowing for 5% chance their results are a fluke. 

Null hypothesis: predicts that any difference/correlation found in the results is due to chance.

If our results are within our acceptable probability then we view them as significant and we can accept the experimental/alternative hypothesis and reject the null hypothesis.

If they are not then we reject the experimental/alternative hypothesis and we accept the null hypothesis. 

Type 1 error: the significance level is too high (too bIg), too lenient e.g. p<_0.5 (accepting results with a 50% probability of being down to chance). The null hypothesis may be falsely rejected and the experimental/alternative hypothesis may be falesy accepted, the researchers suggesting that a certain phenomena exists when it was actually down to chance. 

Type 2 error: the significance level is too low (too smaII), too strict e.g. p<_0.01 (only accepting results if there is only 1% probability they are down to chance). They may falsely accept the null hypothesis and falsely reject the experimental/alternative hypothesis when the phenomena actually exists and was not due to chance. 

3 of 10

Peer review

Definition of peer review: the assessement of scientific work by other experts in the same field

It has three major functions:

1. Helps researchers keep in touch with new scientific developments; knowledge grows through sharing information.

2. If they are working in the same field researchers may be able to improve upon or disprove someone else's theory. 

3. It helps prevent poor quality or fraudulent results entering the public domain. This helps sustain the future credibility of results. It is also important as universities and other organisations are often assessed for future government funding based on the quality of their published research.

4 of 10

Issues with peer review

1. Anonymity: is used so that reviewers can be honest and objective but it may often have the opposite effect. People can use anonymity to settle rivalries as psychologists often compete for the same grants and jobs. That and social relationships may affect objectivity

2. Unable to find a reviewer: it is not always possible to find an appropriate expert. This means poor quality results may be published because the reviewer didn't really understand it enough to see the flaws.

3. Positive bias: peer review tends to favour the publication of positive results.

4. Values in science: peer review often results in a preference for research that supports existing theory over research that contradicts former beliefs. Peer review is often criticised for checking the acceptability and not the validity. 

5 of 10

Headings for designing a study

1. Hypothesis: see whether the question asks for a hypothesis and if so whether is says if it should be directional or non-directional

2. Design: experiment type + details e.g. covert/overt, naturalistic/controlled, structured/non-structured, open ended questions? example questions

3. Sampling: who are your participants? How many? What sampling technique + justify

4. Procedure: describe details 

5. Ethical issues: how will you address all relevant issues

6 of 10

Reporting psychological investigations

1. Abstract: summary of results including aims, hypothesis, method, results and conclusions.

2. Introduction: starts with a review of previous research done in the area so the reader is convinced for the reasons for this study. Then states aims, hypothesis and predictions. 

3. Method: a detailed description - enough detail for replication. 

4. Results: what was found + inferential and descriptive statistics. In the case of qualitative data themes and catagories are described with examples. 

5. Discussion: interpret results and consider implications for future research and real life applications. Summary of results, consider methodology and give suggestions for improvements.

6. References: full details of any journals, articles, or books mentioned. 

7 of 10

Writing consent forms and debriefs

Consent forms

  • Detailed information on what the study is about and what participants will have to do.
  • Reference to ethical issues e.g. confidentiality, debrief, right to withdrawal
  • Please sign _____________


  • Remind them of their right to withdraw (+contact details)
  • Inform them of the true aims/hypothesis of study if deception was used
  • What the use of the data will be e.g. where will it be published
  • If you have anymore questions please contact
  • Thank you
8 of 10

Sampling techniques

  • Random: e.g. assign all names in your sampling frame a number and then pull them out of a hat. It's quick and easy but it is unlikely to be representative of target population.
  • Systematic: e.g. every tenth person on a register. Easy but likely unrepresentative
  • Stratified: sub groups in target population identified e.g. gender, race, age and a predetermined number is selected from each sub group in the sampling frame according to their representation in the target population. Representative but time consuming and difficult.
  • Opportunity: participants selected from naturally occuring groups based on availability at the time. Quick and easy but likely unrepresentative and there may be a bias towards helpful people.
  • Volunteer: e.g. putting up a poster. It's convenient and participants are less likely to drop out however likely to be bias as certain types of people are less likely to volunteer e.g. shy people, it is unlikely to be representative and you may get demand characteristics. 
  • Snowball: start with one person in the target population and ask them to lead you to others and so forth. Useful for target populations who are difficult to find e.g. vegans however may lead to biases e.g. vegans may only know other vegans of the same class/race/age/religion. Also may reach a dead end.
9 of 10

Features of science

1. Empiricism

2. Objectivity

3. Control

4. Replicability

5. Theory construction

10 of 10


No comments have yet been made

Similar Psychology resources:

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