- Identify problem
- Develop Hypothesis
- Devise a study
- Analyse & Evaluate results
- Modify & Repeat
- Develop a Theory
Scientific test must have:
- Empirical evidence
- Hypothesis testing
Arguments against Scientific approach
- Psychologists interact with people:not with physical phenomena, so we can't really study people like a scientists can study electricity
- Behaviours in labs are unusual: behaviours cannot be isolated from a social context or it leads to demand characteristics and unusual behaviour.
- Need for control: 1 variable and one facet of behaviour studied apart from behaviour, how can we seeprate memory from past experiences?
- Notion of Objectivity is a MYTH: Past experiences/beliefs/ideas make objectivity impossible.
New Paradigm Research
Science progresses through long periods of "normal science" till there is a revolution and a new PARADIGM emerges.
a shared set of assumptions about the subject matter and method of study.
NEW PARADIGM RESEARCH
relies on methods that ANALYSE DISCURSIVE MATERIAL e.g. interviews/diaries/blogs. It is focused on Qualitive data.
Goal of research is to be published. New knowledge is published in Journals. Before it can be published it is sent off to PEER REVIEWERS. They act as a system of Quality control, ensuring the work is of the highest quality.
There are some problems with the peer review process:
- CONSISTENCY with previous knowledge is valued
- VALUES of the peer reviewers effect research
- BIAS in review
- FILE DRAWER PHENOMENON
Peer review is important because..
- A way of making a judgement about the validity, quality & importance of the research
- They assess whether methods and designs used are appropriate.
- Can asses whether Fraudulent, Flawed or fit for public.
RELIABILITY: MEASURING SOMETHING AGAIN AND AGAIN AND GETTING THE SAME RESULTS.
- Taking many measurements and averaging the score
- Using pilot studies
- Correlating data from multiple raters
- Checking data transposed from one form to another, carefully.
- Test-retest method.
VALIDITY: WHETHER A TEST / APPARATUS MEASURES WHAT IT CLAIMS TO MEASURE.
IMPROVING INTERNAL VALIDITY:
- Operationalizing variables
- SINGLE-BLIND Technique
- DOUBLE-BLIND Technique
- Face Validity
- Concurrent Validity
- Predictive Validity
- Population Validity
- Ecological Validity
best sample as it represents everyone
available people from the target population
people chosen in a systematic way
Can Do Can't Do With Participants
PROTECTION from harm
Number of PARTICULAR outcomes
Probability = ___________________________________________________
Number of POSSIBLE outcomes
O = Definitely not 1 = Definitely will.
Null Hypothesis: There Will NOT be a relationship between variables.
Experimental hypothesis: There WILL BE a relationship between variables.
LEVEL OF SIGNIFICANCE: the level at which a null hypothesis is accepted or rejected.
0.05 or 5%
One-tailed hypothesis: Directional. States direction of difference or result
Two-tailed hypothesis: Non-Directional.States simply will be difference or result
Type 1 ERROR:
Rejecting NULL and accepting experimental when THERE IS NO RELATIONSHIP.
Type 2 ERROR:
Accepting Null and rejecting Experimental WHEN THERE IS A RELATIONSHIP.
Difference between highest and lowest set of scores
Measure of dispersion. The SPREAD of data around Central Value.
Distribution of a set of data - bars joint
Summary of a set of data - bars separate
Relationship between two variables.
When info is ordered.
When info is categorised/ named
When there are equal intervals on the measurement scale
When there are equal intervals and a True point zero.
SOC CIN WOR MOI
SPEARMAN's RHO CHI-SQUARED WILCOXON MANN-WHITNEY
Ordinal data Independant design Ordinal Data Ordinal data
Correlation Nominal Data Repeated Measures Independent groups
observed value gReateR than critical observed value loWer than critical
Refers to a coding system where categories are used to analyse data before the experiment.
Refers to an identification of key terms through examining data, to be reported in their results.
Independent measures design
refers to where participants are assigned to two different conditions/groups.
Repeated measures design
refers to where participants within the study repeat the study with two different conditions.
refers to a sample when participants are matched for the two conditions.