Glossary
data: relatively uninterpreted information received through human senses
descriptive statistics: methods for numerical summary of set of sample data
inferential statistics: methods for assessing the probability of chance occurence of certain data differences or relationships
design: overall structure and strategy of a piece of research
sample: group selected from population for study or experiment
population: all possible members of a group from which a sample is taken
Types of Hypothesis
hypothesis: precise prediction of relationships between data to be measured; usually made to support more general theoretical explanation
alternative hypothesis: precise statement of relationship between data to be measured. the hypothesis tested in a research project. contrasted to the null hypothesis
experimental hypothesis: hypothesis tested in a particular experiment
research hypothesis: hypothesis tested in a particular piece of research
null hypothesis - prediction that the data do not vary in the way which will support the theory under investigation. very often the prediction that differences or correlations are 0
directional hypothesis (one-sided/one-tailed): hypothesis in which the direction of difference or relationship is predicted before testing. example - significantly more people will order chicken than beef
non directional hypothesis (two-sided/two-tailed): hypothesis in which the direction of difference or relationship is not predicted before testing. example - there will be a significant difference in the number of people who order chicken and beef
Types of Variables
confounding variable: variable which is uncontrolled annd obscures any effect sought, usually in a systemic manner
dependent variable: variable which is assumed to be directly affected by changes in the independent variable
independent variable: manipulated variable in an experiment which is assumed to have a direct effect on the dependent variable
extraneous variable: anything other than the IV which could affect the dependent variable. it may or may not have been allowed for and/or controlled.
random variable: a variable that creates unpredictable errors in measurement
participant/subject variables: variables which differ between groups of people and which may need to be controlled in order to demonstrate an effect of the DV
Types of Groups
control group: group which is used as a baseline measure against which the performance of the intervention group is assessed
experiment/treatment group: group who recieves values of the IV in ann experiment or quasi-experiment
placebo group: group who dont recieve treatment but everything else the experimental group recieve and who are sometimes lef to believe their treatment will have an effect
Types of Sampling
cluster sampling: sample selected from a specific area as beinng representative of a population
opportunity sampling: sample selected because they are easily available for testing
systematic sampling: sample selected by taking every nth case
quota sampling: sample selected so that specified group[s will appear in numbers proportional to their size in the target population
random sampling: sample selected in which every member of the target population has an equal chance of being selected
self-selecting sampling: sample selected for study on the basis of their own action in arriving at the sampling point
snowball sampling: sample selected for study by asking key figures for people they think will be important or useful to include
stratified sampling: samples are selected so that specified groups will appear in numbers proportional to their size in the target population, within each subgroup cases are selected on a random basis
Reliability and Validity
reliability: extent to which findings or measures can be repeated with similar results
replication: repetition of a study to check its validity
standardised procedure: way of testing of acquiring measures from participants which is repeated in exactly the same way each time for all common parts of the method
validity: extent to which instruments measure what it is intended to measure
ecological validity: extent to which investigation can be generalised to other places and conditions
external validity: extent to which results of the research can be generalised acorss people, places, times and other measures of the variables
internal validity: extent to which the effect found in a study can be taken to be real and caused by the identified independent variable
Philosophy
experimental realism: effect of attentionn grabbinf, interesting experiment in compensating for artificiality or demand characteristics
mundane realism: effect of research design which resembles everyday life but is not necessarily engaging to participants
positivism: methodological belief thst description of the world’s phenomena is reducvible to observable facts and the mathematical relationships between them
qualitative approach: methodological stance which holds that innformation about human events and experience, if reduced to numerical form, loses most of its important meaning and valiue for research and understanding
qualitative data: information which is not in, or reducible to, numerical form
quantitative data: information gathered which is in, or reduced to, numerical form
Study Design
correlational study: study of the extent to which one variable is related to another, often referring to non-manipulated variables measured outside the laboratory
demand characteristics: features in a study which help the participant work out what is expected of them
double blind: procedure in the experiment where neither participants not data gatherer know which treatment participants have received
evaluation apprehension: participants concerns about being tested which may affect the results
field experiment: experiment carried out in a natural setting outside the laboratory
laboratory experiment: carried out in control conditions inn the experimenters own habitat
natural experiment: experiment which exploits the occurrence of a naturally occurring independent variable
quasi-experiment: experiment in which experimenter does not have control over random allocation of participants to conditions, nor, in some cases, over the independent variable
experimenter expectancy: tendency for experimenter’s knowledge of what is being tested to influence the outcome of research
experimenter reliability: the intent to which the results produced by 2 or more experimenters are related
ex post facto research: research where pre-existinng and non manipulated variables among people are measured for difference or correlation
observational study: research which simply measures characteristics of how people are or behave but doesn’t intervene
pleasing the experimenter: tendency of participants to act in accordance with what they think the experimenter would like
single blind: procedure in an experiment where participants do not know which treatment they have received
social desirability: tendency of participants in research to want to look good and provide socially acceptable answers
counterbalancing: half participants do conditions in particular order and the other half take in the opposite order
independent samples: two or more separate groups take the various conditions of the independent variable
matched pairs: each participant in onne group/condition is paired on specific variables with a participant in another group/condition
repeated measures: each participant takes part in all conditions on the independent variable
Observational Research Methods
content analysis: analysis of content of media sources/descriotions/verbal reports through coding, categorisation and rating
diary method: participant makes regular record of relevant events
disclosure: letting people know that they are the object of observation
event sampling: observation and recording of specific events defined for the study
formative approach: approach to observation in which the focus of the observation may change as the study progresses and early data are analysed
knowledge elicitation: gathering data which is assumed to form the observed person’s knowledge and understanding of a specific system. often using verbal prompts
controlled observation: observation in which many variables are kept constant
indirect observation: observation without intervention in observed people’s own environment
non-participant observation: observation in which observer does not take part or play a role in the group observed
participant observation: observation in which observer takes part or plays a role in the group observed
structured observation: observation which uses an explicitly defined framework for data recording
observational design: study which is solely observational and does not include and experimentation
observational technique: study using observation in some way and which may or may not be an experiment
observer bias: effect causing unwanted variations in data recorded which are produced because of characteristics of the observer
inter-rater reliability: extent to which observers agree in their rating or coding
point sampling: observation of one person long enough to record one category of behaviour before moving on to next individual to be observed
rating: assessment of behaviour observed by choosing a point along a scale
roleplay: study in which participants act out parts
simulation: study in which participants recreate and play through, to some extent, a complete social settig
time sampling: observation of individuals for set lengths of time
verbal protocol: recording of participants talk when they have been asked to talk or think aloud
Interviews
case history: record of person’s important life events gathered and analysed in a case study
case study: in-depth study of one individual or group, usually qualitative in nature
clinical method: interview method using structure of questions to be asked but permitting tailorinng of later questions to the individuals responses
disguise: dimension of design which is the extent to which interviewees are kept ignorant of the aims of the questioning
focus group: group with common interest ho meet to discuss an issue in a collective interview in order for researchers to assess opinio
open-ended question: interview item to which interviewees can respond in any way they please and at any length
panel: stratified group who are consulted in order for opinion to be assessed
Questionnaires
Likert scale: scale on which respondant can choose from a dimension of responses, usually from strongly against/disagree to stonngly for/agree
semantic differential: scale measuring meaning of an object for the respondent by having them describe it using a point between the extremes of several bipolar adjectives
Thurstone Scale: scale in which raters assess the relative strength of each item and respondents agreeing with that item receive the average rated value for it
diagnostic item: item not obviously or directly connected to the attitude object, yet which correlates well with overall scores and therefor has discriminatory power
discriminatory power: extent to which items, or the test as a whole, separate people along the scoring dimension
factor analysis: statistical technique, using patterns of test or subtest correlations, which provides support for theoretical constructs by locating clusters
psychometric tests: yesys which attempt to quantify psychological variables, skills, abilities, character etc
reliability: consistency and stability of a test
chronbach’s alpha: a generalused Kuder-Richardson type test item discrimination/reliability for a response scale with several points
external reliability: stability of a test - its tendency to produce the same results when repeated
internal reliability: consistency of a test - extent to which items tend ti be measuring the same thing and not in opposition to one another
item analysis: checking each item in a scale by comparing its relationship with the total scores on the scale
split half: comparing scores on two parts (formed by a random and equal division of the items in a test)
response acquiescence: tendency for people to agree with test items as a habitual response
standardisation: adjusting test until scores on it form a normal distribution
validity: extent to which a test measures what was intended
concurrent validity: extent to which test results conform with those on some other measures taken at the same time
construct validity: extent to which test results support a network of research hypothesis based on the assumed characteristics of a theoretical psychological variable
content validity: extent to which test covers the whole relevant topic area
criterion validity: extent to which test scores can be used to make a specific prediction on another measure
face validity: extent to which the validity of a test is self-evident
known groups validity: test of criterion validity involving groups between whom scores on the test should differentiate
predictive validity: extent to which test scores can predict scores on another measure in the future
Comparison studies
cohort: large sample of people, often children, identified for longitudinal or cross-sectional study
cohort effect: confounding in cross-sectional study when two different age groups have had quite different experiences
cross-generational problem: confounding occurring when one longitudinally studied group is compared with another who have generally had quite different social experiences
cultural relativity: view that a person’s behaviour and characteristics can only be understood through that person’s own cultural environment
ethnocentrism: bias of viewing another’s culture from one’s own cultural perspective
indigenous psychology: psychological methodology developed by and within one culture, not imported from another
cross-cultural study: comparative study of two or more different societies or social subgroups
cross sectional study: comparative study of several groups captured for measurement at a single time point
cross-sectional, short-term longitudinal study: comparative study of several age groups followed through over a relatively short period
longitudinal study: comparative study of one group over a relatively long period
time-lag study: comparative study in which a sample of a specific age is selected each time the study is run (across relatively low intervals)
Measurement
categorical variable: variable is not measurable on a linear scale and which has only discrete values
frequency data: data presented as numbers of cases in specific categories
measured variable: variable is at least ordered
psychometrist: person who develops psychological measures and attempts to standardise the scales up to an interval level of measurement
qualitative difference: difference between cases in kind and not numerically measurable, though being different can be counted
quantitative difference: difference between cases measurable by number
continuous scale: scale on which it is always possible to subdivide units of measurement
discrete scale: scale containing only separated values of the variable measurement
interval level of measurement: level at which each unit measures an equal amount
nominal level of measurement: level at which numbers, if used, are mere labels identifying discrete categories of a categorical variable into which cases are sorted
ordinal level of measurement: level at which cases are arranged in rank positions
plastic interval level of measurement: scale which appears to be interval but on which equal numbers do not measure equal amounts
ratio level of measurement: level at which each unit measures an equal amount and proportions on the scale are meaningful; a real 0 exists.
Descriptive Statistics
absolute value: value of a number, ignoring its sign; a number treated as positive even if it is originally negative
average: common language term of central tendancy
bar chart: chart in which one axis represents a categorical or at least discrete variable
box plot: chart showing central spread of data and position of relative extremes
central tendency: formal term for any measure of the typical or middle value in a group
class interval: categories into which a continuous data scale is divided in order to summarise frequencies
deciles: interval of 10% on a continuous scale
deviation value/score: amount by which a particular score is different from the mean of its set
dispersion: technical term for any measure of the variation or spread of a sample of data or a population
bi-modal distribution: distribution with two prominent peaks
cumulative frequency distributions: distribution showing total numbers above or below each class interval
frequency: distribution showing how often certain values occurred
negative skew: distribution which is not symmetrical about the vertical centre and which contains a lot more lower than higher values, relative to the mode
positive skew: distribution which is not symmetrical about the vertical centre and which contains a lot more higher than lower values, relative to the mode
normal distribution: continuous distribution, bell-shaped, symmetrical about its mid-point and the result of a variable affected by many random influences
exploratory data analysis: close examination of data by a variety of means, including visual display, before submitting them to significance testing
frequency polygon: chart showing only the peaks of class intervals
histogram: chart containing whole of continuous data set divided into proportional intervals
mean: a measure of central tendency = mean of all absolute deviations
mean deviation: measure of dispersion = mean of all absolute deviations
median: mid-point of data set
medial position/location: place where median is to be found in the ordered data set
mode: most frequent value
parameter: statistical measure of population
percentile: point on continuous distribution which cuts off a certain percentage of cases
quartile: point on continuous distribution which cuts off one of the quarters
range: top minus bottom plus 1
sampling error: difference between sample mean and the true population mean
semi-interquartile range: distance between first and third quartile in a continuous distribution
standard deviation: the square root of the sum of all squared deviations, divided by N.
z-score/standard score: measure of individual deviation: number of standard deviations a particular score is from its sample mean
variance: square of standard deviation
variation ratio: proportion of non-modal values to all values
Probability and Significance
critical value: value with which a statistic, calculated from sample data, can be compared in order to decide whether a null hypothesis should be rejected; related to level of probability chosen.
directional hypothesis: prediciton which states inn which direction differences will occur
eyeball test: informal test of data made simply by inspection and mental calculation plus experience of values
non-directional hypothesis: prediction which does not state in which direction differences will occur
one-tailed test: test made if the research hypothesis is directional
two-tailed test: test made if the research hypothesis is non-directional
probability: a numerical measure of pure chance
empirical probability: a measure of probability based on existing data and comparing number of target events which have occurred with total number of relevant events
logical probability: a measure of probability calculated from analytical formulae and first principles
subjective probability: a measure of probability made on the basis of human internal, and often emotional, assessment
probability distribution: a histogram or table showing the probabilities associated with a complete range of possible events
significance levels: levels of probability at which it is agreed that the null hypothesis will be rejected
10% significance level: too high for rejection of the null but might merit further investigation
5% significance level: conventional significance level
1% significance level: preferred for greater confidence than the conventional one and which should be set where research is controversial or unique
Type 1 error: mistake made when rejecting the null hypothesis when it is true
Type 2 error: mistake made when retaining the null hypothesis when it is false
Simple Tests of Difference
degrees of freedom: number of cells in frequency table which are free to vary if row and column totals are known. also used in other tests where it defines the number of individual values free to vary when a group total is known
expected frequencies: frequencies theoretically expected in table if no relationship exists between variables
observed frequencies: frequencies actually obtained and submitted to a significance test using Chi-square
test statistic: the figure calculated at the end of a statistical test which is then compared with cirical value tables
chi-squared test: test of association between two variables, using unrelated data at a nominal level
goodness of fit test: frequency level test used to decide whether a given distribution is close enough to a theoretical pattern
binomial sign test: nominal level test of differences between two sets of related data
Mann-Whitney test: ordinal level test for differences between two sets of unrelated data using U
Wilcoxon rank sum: ordinal level test for differences between two sets of unrelated data using T
Wilcoxon signed ranks test: ordinal level test for differences between two related sets using T
ties: feature of data when scores are given identical rank values
homogeneity of variance: assumption to be satisfied by data before proceeding with a parametic test. occurs when the two variances are not significantly different
parametric test: relatively powerful significance test for data at interval level or above. the tests make estimations of population characteristics and the data tested must therefore satisfy certain assumptions
power of test: likelihood of a test detecting a significant difference when the null hypothesis is false
power efficiency: comparison of the power of two different tests of significance
related t-test: parametric difference test for related data
unrelated t-test: parametric difference test for unrelated data
robustness: tendency of test to give satisfactory probability estimates even when data assumptions for the test vary somewhat from the ideal
sampling distribution: hypothetical population distribution which can be estimated from sample statistics
standard error: standard deviation of a hypothetical sampling distribution
Correlation
correlation: relationship between two variables
coefficient correlation: numerical value of relationship between two variables
curvilinear correlation: relationship between two variables which gives a low value for r because the relationship does not fit a straight line but a good curve
negative correlation: relationship where as values of one variable increases, related values of the other tend to decrease
positive correlation: relationship where as values of one variable increases, related values of the other also increase
significance: measure of whether a correlation was likely to have occurred by chance or not
strength: measure of the degree of matching measured by a correlation
coefficient of determination: percentage of variability in one variable predictable from another using the correlation coefficient between them
guessing error: extent to which predictions of values of one variable are likely to be incorrect using values of another variable
multiple regression: technique in which the value of one variable is estimated using the correlationn with several other variables
pearson’s product-moment coefficient: parametric measure of correlation
phi co-efficient: measure of correlation between true dichotomous variables
point biserial correlation: measure of correlation where one variable is dichotomous
scattergram: diagram showing placement of paired values on a two-dimensional grid
spearmans rho: non-parametric ordinal level measure of correlation, correlation of ranks
variance estimate: estimate of variability in one variable using variance of a correlated variable
ANOVA
a prior comparisons: differences between means or combinations of means where were predicted from theory before the data were collected
analysis of variance (ANOVA): statistical technique which compares variances within and between samples in order to estimate the significance between sets of means
between groups sums of squares: sum of squares of deviations of group means from the grand mean; used to calculate the variance component related to the effect (distance between group means)
bonferroni t tests: procedure for testing several planned comparisons between groups of means
error rate per comparison: given the significance level set, the likelihood of an error in each tests made on the data
family wise error rate: the probability of having made at least one type error in all the tests made on one set of data
error sum of squares: sum of all the squares of deviations of each score from its group mean, for all scores in a set of data where there are two or more groups; used to calculate an estimate of the ‘unexplained variance‘ with which to compare the ‘explained variance of group means around the grand mean
F test/variance ratio test: comparison of two variances by dividing one by the other
grand mean: mean of all scores in a data set, irrespective of groups
linear coefficients: values to be entered into an equation for calculating linear contrasts
linear contrasts: procedure for testing beterrn individual pairs of means or combinations of means when planned comparisons have been made
mean sum of squares: sum of squares divided by degrees of freedom
Newman-Keuls test: procedure for testing all possible pairs of means in a data set for significance, so as long as number of groups is relatively low
pairwise comparisons: comparison of just two means from a set of means
planned comparisons: tests which was intended to make because of theoretical predictions, before data were collected
post hoc comparisons: tests between means, or groups of means, only decided upon after inspection of data
Tukey’s test of significance: procedure for testing all possible pairs of means from a data set where there are relatively large number of groups
unrelated design: an ANOVA design using only unrelated samples in all factors (independent variables)
factor in ANOVA: one of the IVs in a design with more than one IV
factorial design: a research design that involves more than one IV
interaction effect: effect of one factor which is significant but which depends upon only certain levels of other factors
levels in ANOVA: the different conditions of an IV
main effect: effect of one factor which is significant across all its levels taken together, irrespective of any other factors
mixed model: an ANOVA design using at least one repeated measure IV and at least one unrelated IV
simple effect: one level of a factor only has a significant effect across levels of the other factors
between subjects variation: variation associated with the differences between participants’ overall totals in repeat measures design
between conditions variation: variation calculated in a repeat measure design, which comes from how scores between conditions vary when the variation between participants overall totals have been removed
ANOVA residual: a term for the remaining variation in a repeat measures design when variation between subjects and between conditions has been removed
ANCOVA: statistical procedure used to investigate differences between two means which may be adjusted to allow for the fact that the two groups differ on a variable which correlates with the DV (covariate)
co-variate: a variable which correlates with the DV and on which two groups, who are being investigated for difference, differ. the biasing effect of the confounding variable can be adjusted for in ANCOVA.
criterion variable: variable which is being predicted in regression procedures
MANOVA: statistical procedure for testing the effects of one or more IVs on more than one DV
multiple regression: statistical procedure in which the correlations of several predictor variables with a criterion variable are summed to give a better prediction of that variable
multiple regression co-efficient: a value indicating the strength of prediction of the combined set of predictor variables being used in multiple regression
predictor variable: variable being used to predict a criterion variable in regression procedures
regression: procedure of predicting a criterion variable (Y) from a predictor variable (X) using the line of best fit around which correlated pairs of X and Y scores are arranged
regression coefficient: value indicating the extent to which each predictor variable predicts scores on the criterion variable in multiple regression procedures
regression residual: difference between an actual score and what it would be as predicted by a predictor variable using a regression procedure (Y-Y-hat)
suppressor variable: variable whose common variance can be partialled out of the variance of a predictor variable so that the latter can more accurately predict values of a criterion variable in multiple regression procedures
Analysing qualitative data
early analysis: formation of hypothesis and theoretical ideas during the acquisition of data
grounded theory: theory grounded in specific observational data = patterns emerge from the data set and are not imposed on it before it is gathered
negative case analysis: analysis of reasons why a single case does not fit patterns identified so far
triangulation: comparison of at least two views od the same thing
transcription: written recording of directly recorded speech, as exactly as possible, but depending upon approach; usually often includes pauses, intonation etc
Ethics
debriefing: informing participants about the full nature and rationale of the study they’ve experienced and removing any harm to self-image or self-esteem
deception: leading participant to believe that something other than the true IV is involved or, at least, not giving full information to the participant about the IV, DV or overall procedure
intervention: research which makes some alteration to people’s lives beyond the specific research setting, in some cases because there is an intention to remove disadvantage or make improvements of some kind in people’s overall condition
invasion of privacy: effect of research which intrudes on people’s personal lives
involuntary participation: taking part in research without agreement or knowledge of the study