Research Methods

Objectivity - free of opinion/bias - based purely on emperical evidence

Replicability - check + verify the information by repeating a study

Falsifiability - ability to demonstrate that a theory is wrong 

- New Knowledge is published in scientifc journals - permanent records

- Checked by professionals in same field who peer review it 

- Only high quality work is published 

ISSUES WITH PEER REVIEW 

- Rejection of new research which does not reflect the current paradigm

- File drawer phenonemon - research which does not supoport the hypothesis = drawer - rather than published

- Reviewer bias - if the expert does not agree/ if the research doesn't come from a specific institution - may not be published

Internal - controlling variables within a study - e.g.  does the DV measure what it was designed to measure?

HOW TO IMPROVE

- use measures/scales which have been previously validated - controll all variables + run pilot study

External - generalised beyond the study itself 

HOW TO IMPROVE

- replicate with different groups of people in different setttings

Internal - consistency within a test - e.g. how well items measure the same thing

HOW TO IMPROVE

- use the split-half method to compare results from half of the test to another

External - ability to replicate the same test and find similar results each time 

HOW TO IMPROVE 

- use test-retest method with same/similar pps - high correlation = reliable test

Informed consent

Deception

Debriefing 

Right to withdraw

Confidentiallity 

Protection from physical/psychological harm 

Probability - likelihood of something happening - expressed between the numbers 0 and 1

Statistical tests used to work out probability of out results to be due to chance. 

In order to avoid accepting/rejecting our experimental hypothesis incorrectly - level of probability = 5% (0.05) = 5% results due to chance. 

If we are too strict/lenient with our data - type 1 or type 2 error can occur

Type 1 - being too lenient e.g. (10%) - reject null, accept experimental when we should reject it 

Type 2 - being too strict e.g. (1%) - accept null, reject experimental when actually correct

Nominal - data put into categories (or named)

Ordinal - data can be put in order (position in a race)

Interval - comparative numbers/scores taken (GCSE scores)

Man-Whitney U - independent groups design, difference, ordinal/interval data

Wilcoxon - repeated measures, difference, ordinal/interval data

Spearman's Rho - relationship (correlation), ordinal/interval data

Chi-Square - nominal data

Qualitative Data- collected by interviews/open-ended questions in questions 

Content analysis is a technique for systematically describing written, spoken or visual communication.

- transcribe the data into written form

- read & re-read the data

- code/organise the data into key themes/categories

- arrange into groups

- reflect on what the data tells us 

Abstract - aims, hypothesis, procedure, results/conclusions

Introduction - review of previous research, link to previous research, state aims/predictions

Method - detailed to ensure replicability, design, sample, apparatus, procedure, ethics

Results - descriptive statistics = tables/graphs, inferential statistics = Chi Square etc. 

Discussion - interpret results, relation to previous research, consnideration of methodology

Referrences - full details of journal articles

Design - lab/field/natural, indepedent/repeated/matched, IV and DV, control of any EV, dealing with ethical issues

Participants- sample size, sampling method (opportunity, random, volunteer)

Materials - questionnaire, pens, paper, stopwatch, consent form etc.

Procedure (step by step guide to allow replication) - duration, whereabouts, how many people recording data, how the data is recorded