1.3 Data recording, analysis and presentation Flashcards
What are the assumptions of parametric tests?
-Populations drawn should be normally distributed
-Variances of populations should be approximately equal
-Should have at least interval or radio data
-Should be no extreme scores
When are non-parametric tests used?
-When assumptions of the parametric tests cannot be fulfilled
-When distributions are non-normal
What are the 5 types of non-parametric tests?
-Mann-Whitney U Test
-Wilcoxon Signed Ranks Test
-Chi-Square
-Binomial Sign Test
-Spearman’s Rho
What is the observed value?
The number produced after the various steps and calculations for a statistical test have been carried out
What is the critical value?
A value taken from a statistical test table, which must be reached in order for results be significant
What are the conditions of use for the Mann Whitney U Test?
-DV produced ordinal or interval data
-Independent Measures design
-Test for a difference
What are the steps for the Mann Whitney test?
- Rank data as whole cohorts of both groups together
- If there are any tied ranks add ranks they would have occupied and divide by how many that will share the rank
- Calculate R1 and R2 (the total of ranks of each group)
- Put R1 or R2 (whichever is smaller) into the formula to get the observed value
- Find the critical value. The observed value should be less than the critical value for significant results
- Write the significance statement
What are the conditions of use for the Wilcoxon signed rank test?
-DV produced ordinal or interval data
-Repeated measured design
-Test for a difference
What are the steps for the Wilcoxon signed ranks test?
- Column A minus column B to calculate the difference
- Remove any with 0 as that means no differences
- Ignore the minus sign and rank the numbers from lowest to highest
- Look at the difference column and count how many positive figures and how many negative figures there are
- Using the least frequent sign, add the rank of those with that sign, this gives the observed value
- Find the critical value. For significant results, the observed value should be less than the critical value
- Write the significance statement
What are the conditions of use for the Chi-square test?
-DV produced nominal type of data
-Independent measures design
-Difference/relationship
What are the steps for the chi-squared test?
- Calculate the totals of each row and column and the grand total from the contingency table
- Calculate the expected frequency
- Apply the formula to each cell (the O is the original value in the cell) and add these all together
- Calculate the degrees of freedom (no of rows minus 1) x (no of columns minus 1)
- Identify the critical value. For significant results, the observed value should be greater than the critical value)
- Write a significance statement
What are the conditions of use of the Binomial sign test?
-DV produces nominal data
-Repeated measures design
-Testing for a difference
What are the steps for the Binomial sign test?
- If the participant’s score for each condition is the same, they are removed
- Minus the second score from the first score and code the positive differences as + and the negative differences as -
- Count the total number of positives and negatives and identify which sign is the least frequent
- The number of the least frequent sign is the observed value
- Identify the critical value and if the observed value is less than the critical, the results are significant
- Write the significance statement
What are the conditions of use for the Spearman’s Rho test?
-Interval/ ordinal data
-Correlation
-test for relationship
What are the steps for the Spearman’s Rho test?
- Rank each co variable separately low to high (lowest = 1st and tied ranks where applicable)
- Difference rank 1 minus rank 2
- Difference squared
- Sum of difference squared column
- Put this into formula
- If r is positive then it is a positive correlation
- Find the critical value and if the OV is greater than the CV, the results are significant
- Write significance statement.
What degree of significance is used in psychology?
5%
What are the levels of measurement?
Ordinal, nominal, and interval
What is nominal data?
Data as totals of named categories
What is ordinal data?
Data as points along a scale, such that the points fall in order but there are not necessarily equal gaps between those points.
What is interval data?
Data as points on a scale that has equal gaps between the points and is widely recognised as a scale of measurement
What are the strengths of nominal data?
-Easy to generate from closed questions, so large amounts of data can be collected quickly, increasing reliability
-Quick to find the mode to assess central tendency
What are the weaknesses of nominal data?
- Without a linear scale, participants may be unable to express degrees of response
-Points are not on a linear scale so medians and means cannot be used to assess central tendency
-Can only use mode as a measure of spread
What are the strengths of ordinal data?
-More informative than nominal data; indicate relative values on a linear scale
-Easy to generate from likert and rating scales
-Points are on a linear scale so a median can be used as well as a mode to assess central tendency
What are the weaknesses of ordinal data?
-As the gaps between the points are only relative, comparisons between participants may be invalid as they may interpret the scale differently
-Gaps between the points are not equal so a mean cannot be used to asses central tendency
What are the strengths of interval data?
-More informative than nominal and ordinal data as the points are directly comparable because all the points are of equal value
-Easy to generate from closed questions
-Points are on a linear scale, with equal gaps between the points so a mean can be used as well as the mode and median to assess central tendency and variance or standard deviation can be used as measures of spread
-Scientific measurements are highly reliable and have an absolute zero baseline
What is a weakness of interval data?
In interval scales that are not scientific measures, there is no absolute baseline to the scale so scoring zero may not mean that the participant does not demonstrate that variable at all, merely that the scale doesn’t measure it
What is the difference between quantitative and qualitative data?
Quantitative data is numerical and is a way of operationalising the variables whereas qualitative data is description based
What are the strengths and weaknesses of quantitative data?
-Tends to be objective
-Tends to be highly reliable
-Can be analysed using inferential stats
BUT
-Limits participants’ responses, making the data less valid for if appropriate response options are not available
-Lacks context
What are the strengths and weaknesses of qualitative data?
-Highly valid as it is likely that the participants are able to fully express their answers
-Less likely that key or rare observations will be lost
BUT
-Tends to be subjective
-May be invalid if the researcher has been biased
-May be difficult to make generalisations from the findings
What is the difference between primary and secondary data?
Primary data is the results of a first hand investigation; collecting information directly from a sample. Whereas secondary data is information that is obtained about the results of an investigation that has already been conducted by another researcher, possibly for a different purpose. This can then be re-used in a new analysis.
What is a frequency table/tally chart?
A grid showing the possible categories of results in which a tick or tally is made each time the item is scored. These can be added to give a total in each category
What are the measures of central tendency?
Mode, median, and mean
What are the measures of dispersion?
Variance, range, and standard deviation
How do you calculate range?
Find the largest and smallest value. Subtract the smallest value from the largest THEN add 1
What is variance?
A measure of dispersion that calculates the average difference between each score in the data set and the mean. Bigger values indicate greater dispersion
How does variance relate to standard deviation?
standard deviation squared = variance
In what ways can data be presented visually?
Line graph, pie chart, bar chart, histogram, scatter diagram
What does the peak of a normally distributed curve represent?
Mode, mean, and median
What percentage of values are within the first 3 standard deviations of a normally distributed curve?
1 sd - 68%
2 sd - 95%
3 sd - 99.7%
In what direction does a positively skewed curve tend to?
The left
What is type 1 error?
If the alternative hypothesis is accepted when, in fact, the distribution of results is due to chance.
What is a type 2 error?
If the null hypothesis is accepted when the distribution of results is not due to chance.
How can you reduce the risk of making a type 1 error?
reducing significance from p<0.05 to p<0.01. But this increases the risk of making a type 2 error
What code of ethics do psychologists follow in the UK?
the British Psychological Society’s Code of Ethics and Conduct
What 4 areas of ethical considerations are there and what do they each entail?
Respect - informed consent, right to withdraw, and confidentiality
Competence
Responsibility - protection of participant and debrief
Integrity - deception
What is meant by representative?
To what extent the sample of a study is representative of (similar to proportionally) the target population
What is meant by generalisability?
The extent that the results of the study can be generalised/ applied to the target population
Internal reliability
The extent to which the results are consistent across the same measure
External reliability
The extent to which a measure varies from one use to the next
Inter-rater reliability
A method of measuring the consistency of a measure by assessing the measures of multiple different observers or “rater” to ensure similarities
Test retest reliability
A measure of reliability obtained by administering the same test twice over a period of time to a group of individuals
Split half reliability
Determined by dividing the total set of items (eg. questions) relating to a construct of interest into halves and comparing the results obtained from the two subsets of items thus created
Internal validity
The degree of confidence that the causal relationship being tested is trustworthy and not influenced by other factors or variables
External validity
The combination of ecological validity and population validity
Face validity
The extent to which a study appears to do what it is supposed to - its effectiveness
Construct validity
The extent to which a test or measure accurately assesses what it’s supposed to
Concurrent validity
The extent of the agreement between two measures or assessments taken at the same time
Criterion validity
Evaluates how accurately a test measures the outcome it was designed to measure
Population validity
Whether you can reasonably generalise the findings from a sample to a larger group of people
Ecological validity
The extent to which findings in a study have the ability to be generalised to real life scenarios and still be valid
Demand characteristics
When a participant showcases particular behaviours due to knowing they are being studied/observed, making the results unrealistic
Social desirability
Describes the tendency of participants to respond in a way that they think is viewed favourably by others/socially acceptable, as opposed to their genuine beliefs
Researcher/observer effect
A participant changing their behaviour as a result of knowing they are being observed
Researcher/observer bias
When a researcher has inherent or deliberate bias towards certain behaviours conclusions or people
