Data Test Facts Flashcards
Explain experimental research style
- one variable is changed (IV) while the others are kept constant
- Allows researcher to establish a cause-effect relationship between IV and DV
- consists of control group and an experimental group so results can be compared.
Explain correlation research design style
- measures two variables for a statistical relationship
- in correlation study, a variable can be any defined numerical value
- used for quantitative measures
Explain quasi-experiment design.
- quasi-experiments are like true experiments BUT individuals are not randomly allocated into control or experimental groups.
- this is because they can’t be, the IV is something pre-determined.
Explain case studies design
- in depth study of particular person or group
- often conducted over long period of time
- only used in ethically dubious cases or with rare phenomenon
Explain observational research design.
- researcher watching specific behaviours of interest
- it can be done with the participants knowing (overt) or without the participant knowing (covert)
Explain self-report design
- require participants to rate their own behaviours or actions using tools such as questionnaires and surveys.
Explain interviews design
- one on one or small groups can be interviewed by a researcher asking questions.
- different types of interviews and types of questions to be answered:
- open questions that require explanation; closed questions that require just one word answers.
- semi-structured interviews where questions are pre-prepared.
Explain convenience sampling
- non-probability sampling in which a sample is drawn from the part of population that is close to hand
- no criteria, people just need to be willing to participate.
- susceptible to sampling bias
Explain repeated measures design
- involves multiple of the same variable taken on the same or matched subjects either under different conditions or over two or more time periods
Explain matched participants/pairs design
- participants are paired with another participant who has a similar trait or characteristic
- they are then split and allocated to different conditions - attempt to control extraneous variables.
Explain independent groups design.
- different participants are used for each condition in design.
What is a p value
- determines how likely it is that the result occurred by chance alone.
- the smaller the p value, the more reliable and statistically significant the results are.
What is an alpha level?
- the ideal level that the result occurred by chance alone
- 1% or 5% are common
Explain a t-test and it’s purpose
- test of the null hypothesis
- 30 people or more
- assumption of normal distribution.
- compares difference between between two groups when the dependent variable is either interval or ratio data.
- unpaired t-test (independent groups) : compares two different sets of subjects
- paired t-test (within-subjects t-test) : comparisons of same set of subjects exposed to two conditions.
- PARAMETRIC
What is an alternate hypothesis?
The hypothesis which assumes that the data answers the research question.
What is a null hypothesis?
The hypothesis which disproves the research question, where the data does not answer the question.
Explain a Mann Whitney U test and its purpose.
- test of the null hypothesis
- does not require assumption of normal distribution.
- less than 30 participants
- not as statistically powerful.
- compares difference when dependent variable is ordinal or continuous.
- NON-PARAMETRIC
What is a confidence interval?
- allows us to estimate the range in which our true population mean falls, given what we know about the population from the sample we’ve observed.
- always calculated with a given level of certainty (usually 95%), along with standard error or standard deviation.
Explain Spearman’s coefficient
- determines how likely it was that the results were found by chance alone (same as p value)
- non-linear graphs used for
What happens to the null hypothesis if the U-crit is bigger than or equal to the U-stat
Reject the null hypothesis
What happens to the null hypothesis if the U-crit is smaller than the U-stat
Fail to reject the null hypothesis