3. Data Recording, Planning and Analysis Flashcards
Raw Data
Raw data is the data psychologists collect from each participant
Levels of Data: Nominal Data
When the data is split into categories as to how often they occur. This is the most basic type of data gathered.
Levels of Data: Ordinal Data
This type of data is the individual’s data, psychologists can then rank each person and put them in order of how well they did.
Levels of Data: Interval Data
Has equal intervals, this type of data ranks the participants.
Levels of Data: Quantitative Data
Quantitative data is numerical data.
S: Allows comparisons between participants or groups. Easily summarized and can use descriptive statistics.
W: Lacks ecological validity as it cannot reflect how we respond in everyday life, and has a limited amount of detail.
Levels of Data: Qualitative Data
Qualitative data consists of decorative words regarding how participants are feeling.
S: Provides rich, detailed information, thus making it more valid.
W: Can often be hard to summarize and quantify data.
Levels of Data: Primary Data
When information i collected directly from the participants.
S: Psychologists know that controls were in place, thus making the data more reliable.
W: Primary data can be difficult or expensive to collect and some data may have already been collected.
Levels of Data: Secondary Data
Data that has already been collected but is accessible for the psychologist.
S: sometimes data that cannot be collected firsthand or is too expensive to collect is accessible.
W: May be affected by extraneous variables and vital information may be excluded.
Descriptive Statistics: Measures of Central Tendency, Mean
Calculate the mean by adding all the values and dividing the total by the number of values. Can be used on both discrete and continuous data.
S: All the data is used to calculate an average.
W: Very large or small numbers can distort the result (outliers)
Descriptive Statistics: Measures of Central Tendency, Median
Calculated by finding the middle number when all the values are put in order.
S: Very big or small values cannot affect the result (outliers).
W: Takes a long time to calculate for a very large set of data, and doesn’t represent the whole data set.
Descriptive Statistics: Measures of Central Tendency, Mode
The data that occurs most frequently.
S: The only average we use when the data is not numerical.
W: There may be more than one mode, or no mode at all if none of the data is the same. It is also less likely to accurately represent the data.
Descriptive Statistics: Measures of Dispersion, Range
The highest value in the data set take away the lowest value in the data set.
Descriptive Statistics: Measures of Dispersion, Variance
A measure of how much values in a data set differ from the mean.
Descriptive Statistics: Measures of Dispersion, Standard Deviation
The standard deviation is how much the data is spread from the mean. The square root of variance is the SD.
Descriptive Statistics: Ratio
The measure of two or more variables and the rate at which they change together. Calculated by dividing the bigger number by the smaller one.
Descriptive Statistics: Frequency Tables (tally charts)
A simple way of presenting data is o show a tally of the behaviour using a frequency table (records how frequently the behaviour occurs)
Descriptive Statistics: Line Graphs
A line graph is most useful to show the profession/regression of behaviour over a period of time.
*Always label axis and include a title
Descriptive Statistics: Pie Charts
Pie charts are helpful to show behaviour statistics and percentages as a proportion of a total.
* Ensure to include a title and key for each section of the chart.
Descriptive Statistics: Bar Graph
A useful and meaningful way to represent data as it is simple.
*Label axis and include a title
Descriptive Statistics: Histograms
Used only for continuous data.
*Include labels and a title
Descriptive Statistics: Scatter Diagram
Used to show correlation.
*Include title and line of best fit and label axis
Inferential Statistics: Normal Distribution
Data about any behaviour from a representative sample of the target population will tend to fall into a curve of normal distribution. Most scores will be around the midpoint and lines will steep up on either side of this. (hill shape)
Inferential Statistics: Skewed Distribution
If a distribution use was representing data from a skewed distribution curse, the mean would be different from the mode. Positive skewness curves to the left, and negative skewness to the right.
Inferential Statistics: Probability and Significance Levels
Probability is the likelihood something is happening. The usual level of significance, which is the level at which psychologists will reject the null hypothesis at, is 95%.
The shorthand way psychologists write this research is P<= 0.05 where p= the probability of the result being due to chance.
Inferential Statistics: Non Parametric inferential tests
- Chi squared = nominal, independent design
- Binomial = nominal, repeated design
- Mann-Whitney U = ordinal, independent design
- Wilcoxon = ordinal, repeated design
- Spearman’s Rho = Ordinal, test for correlation
Inferential Statistics: Type 1 Error
When the alternative hypothesis is accepted and the null is rejected, incorrectly.
Inferential Statistics: Type 2 Error
When you accept the null hypothesis and the alternative is rejected, incorrectly.
Methodological Issues: Representativeness
If the sample is a similar makeup to the target population it can be called representative.
Methodological Issues: Generalisability
When you have a representative sample you can predict behaviour from the wider target population. Whether or not its applicable.
Methodological Issues: Internal Reliability
Refers to the consistency of results of a test across items within the test.
Methodological Issues: External Reliability
Refers to the extent to which a test score varies from one time to another.
Methodological Issues: Inter-rater Reliability
In order to avoid observer bear, inter-rater reliability needs to be established. This means that there are more than one observer observing the same behaviour, if they again the same results it is inter-rater reliable.
Methodological Issues: Split-Half Reliability
A way to test internal reliability. It would test half the questions and gain a score from it, and then test the other half. If they showed the same score then they are reliable.
Methodological Issues: Test-Retest Reliability
Tests external validity of a questionnaire or piece of research. A high level of standardisation makes a piece of research replicable.
Methodological Issues: Internal Validity
How the research is measuring the DV and the effect of the IV on it, a lower internal validity would be caused by extraneous variables.
Methodological Issues : External Validity
How much the research can be generalised to other settings.
Methodological Issues: Face Validity
How good the rest or research looks to be at testing what it is meant to be testing.
Methodological Issues: Construct Validity
Where a test or study measures the actual behaviour it sets out to measure.
Methodological Issues: Concurrent Validity
Where a test or piece of research gives the same results as another test to study which claims to measure the same behaviour.
Methodological Issues: Criterion Validity
Refers to how much one measure predicts the validity of another measure.
Methodological Issues: Population Validity
How accurately the test or study measures behaviour in the general population.
Methodological Issues: Ecological Validity
How like real like a study is considered to be.
Methodological Issues: Demand Characteristics
Where participants interpret the aims of the experiment/research and change their behaviour to fit it.
Methodological Issues: Social Desirability
Where participants try to present an image that they feel will present them in a good light as a good member of society, but isn’t necessarily accurate or true to life.
Methodological Issues: Researcher/Observer effect
Effects on participants (and their behaviours and responses) which are brought about by the researcher ir observer’s presence.
Methodological Issues: Ethical Considerations
In the UK, the British Psychological Society (BPS) produced the Code of Ethics and Conduct which are used for all research taken place by psychologists. You have to look at four areas before being allowed to conduct your research: Respect, Competence, Responsibility and Integrity.
Methodological Issues: Ethics, Respect
Areas of Respect:
- Giving informed consent
- Allowing right to withdraw
- Ensuring confidentiality
Methodological Issues: Ethics, Competence
Must have a certain level of knowledge in that area in order to know that the research conducted is safe and useful.
Methodological Issues: Ethics, Responsibility
Areas of Responsibility:
- Protection of participants
- Inclusion of debrief
Methodological Issues: Ethics, Integrity
To have integrity, a study cannot have any form of deception.
