Assignment Bias | | Whether an instrument actually measures what it sets out to measure |
Active deception - Modalities of measurement | | Deliberately mislead the participant with information e.g. false information or feedback (unethical) |
Alternative hypothesis | | The same assignment (testing) may be given to two groups, however they do not consider the confounding variables that may skew their results e.g. testing two schools and not considering SES, ethnicity, culture etc. |
Assumptions when running an analysis-Independence: | | As the magnitude of the correlation increases, our ability to predict one variable based on knowledge of the other variable increases. Our ability to predict one variable based on knowledge of the other variable increases. |
Assumptions when running an analysis-Normality: | | If a measure is not reliable it is never valid |
Assumptions when running an analysis-Linearity: | | Can you establish that an instrument measures what it claims to measure through comparison to objective criteria. |
Assumptions when running an analysis-Homoscedasticity: | | more than two categories |
Availability heuristics | | Making a mental shortcut based on recent events or things that are common to us. |
Behavioural Modalities of measurement | | A response from a cross population or group at one point in time e.g. 10-19, 20-40 etc. of different participants (rather than waiting for 10-19 group to age) |
Between subjects design (BSD) | | two categories |
Binary variable | | A design of experiments. Has two or more groups of participants/subjects each being tested by a different testing factor simultaneously. Each participant participates in one and only one group. Not participating in all treatments – opposite to within subject design. |
Bivariate correlation | | Names two distinc types of things eg. Male/female; dead/alive; pregnant or not; yes or no. (just two categores – you must be one or the other). |
BSD | | Extent to which a measure is well founded and corresponds to the real world. It measures what it was intended to measure |
1. Binary | | Does it measure / do what we want it to |
2. Nominal | | each participants and should participate only once and should not influence the participation of others. |
3. Ordinal | | there should e a linear relationship between the variables. If relationship between the variable sis not linear, it will not b adequately captured and summaries by Pearson’s r. |
Categorical variable | | Actions e.g. Number of aggressive acts five minutes after watching a violent cartoon |
Coefficiency of determination | | Made up of categories e.g. humans, cats etc. |
Concurrent Validity | | What you are testing |
Constant | | Used to measure the linear association between two continuous variables |
Construct Validity | | When data are recorded simultaneously using the new instrument and existing criteria |
Content validity | | Have independent judges found and agreed that all areas of interest have been assessed |
Convergent Validity | | Are they logical. If I read the question would I understand the topic. Would it make sense (is it in my face easy to understand) |
Criterion | | Does my IQ measure correlate with other academic performance |
Cross section research | | Identify because all participants do different treatments |
Criterion validity | | (H1) e.g. if you imagine eating chocolate you will eat less of it – directional hypothesis – opposite to H0 |
Criterion validity | | Does it measure the intended construct |
Face validity | | The error variance is assumed to be the same at all points along the linear relationship. That is, the variability in once variable should be similar across all values of the other variable. All are similarly in line. |
Reliability and Validity | | Each variable should be normally distributed. Bell curve |
Valid | | same as nominal but categories have a logical order e.g. uni grades F, P, C, H & HD |
Validity | | A set of values that is the same for different individuals in your data set |