click below
click below
Normal Size Small Size show me how
PSY290 exam 2
| Question | Answer |
|---|---|
| Alpha level | Usually .05 Reduced to .01 for drug studies. Is the probability of a Type I error. |
| Correlation | Allows the experimenter to determine simultaneously the degree and direction of a relationship with a single statistic. |
| Correlation coefficient | Used to determine the strength and direction of a linear relationship between 2 variables. Goes from -1 to 1. |
| Descriptive Statistics | Simplifying groups of people, methods of organizing and summarizing data. |
| Inferential statistics | Set of procedures of taking sample data and analyzing the sample error to get a more accurate estimate of the parameters. |
| Graphs: histogram, bar graph, polygon | Histogram: Bars touching, continuous data Bar Graph: Bars spaced out, discrete data Polygon: Line graph, continuous data |
| Null hypothesis/Null results | Predicts the results if the IV has no effect on the DV |
| One-tailed vs. Two-tailed tests | One tailed: means you're predicting a direction, less conservative, and if you predict the wrong direction, you can't do anything. Two tailed:Non directional, no prediction of direction |
| Type I & Type II errors | I: Shows evidence of a treatment effect when there isn't one prob=alpha II failure to reject the null when it is in fact false |
| Population | Total set of potential observations from which a sample can be drawn |
| Sample | Observations selected from a population |
| Random selection | Used to ensure that everyone in the population has an equal chance of being selected for your population |
| Random assignment (randomization) | The point is to create equal groups prior to any manipulation. |
| Representative sample | You need one in order to get meaningful results that apply to the population of interest. |
| Variability | How much the scores differ from the mean |
| Baseline | A measurement used as the basis for comparison, usually when no treatment is given. It is a way to create equal groups? |
| Confound | simultaneous variation of a second variable w/ an IV of interest so that any effect on the DV can't be attributed to the DV. Incorrect correlation. |
| Critical experiments | A key experiment that purports to distinguish among competing theories |
| demand characteristics | those keys available to subjects in an experiment that may enable them to determine the purpose of the experiment or what is expected by the experimenter. Can be reduced by running experiment in a natural setting. |
| Descriptive techniques | Case study survey naturalistic observation |
| between subjects design | test design in which each subject is tested under only one level of each independent variable. You would run an independent T test. No potential for carry-over effects. more conservative (advntg) but you will need more subjects (disadvntg) |
| Within subjects design | An experimental design in which each subject is tested under more than one level of the independent variable. Disadvantages: Carry-over effects are a potential problem and so you must counterbalance. Advantages: reduced individual differences, fewer subje |
| Mixed design | an experimental design containing both within and between subject independent variables. Is inherent in a quasi-experiment. |
| Double-blind | an experimental technique in which neither the subject nor the experimenter knows which subjects are in which treatment conditions. Used to get rid of experimenter effects. |
| Experimental error | any variation in the dependent variable that is not caused by the independent variable |
| experimental hypothesis | the research hypothesis that specifies the effects of the independent variables |
| experimentation advantages and disadvantages | Advantages: You can CONTROL extraneous variables and ISOLATE the variable(s) of interest. BEcause of this we can DETERMINE CAUSALITY Disadvantages: Artificial environment (people don't act naturally), demand characteristics (subjects know), experimenter |
| experimenter effects/bias | clue provided by the experimenter as to the way the research is conducted, and it goes on a spectrum from unintentional to faking data. To get rid of Experimenter effects, you must run a double blind study and you don't have a choice in drug research. |
| explanatory techniques | ? |
| Factors | independent variables and subject variables come into play in factorial design. ex: Boys and girls each tested in A and B. |
| Factorial notation | ? |
| Generalization | formation of broad propositions derived from individual facts |
| interaction | an experimental result that occurs when the level of one IV are differentially affected by the levels of other IVs |
| levels of variables/factors | values of IVs?? |
| Matching | Way to create equivalent groups by deciding important qualities to match on, but that is where you can mess up--deciding the qualities to match on. After pairs are made, random assignment is used to separate pairs. |
| Operational definition | a definition of a concept in terms of the operations that must be performed to demonstrate the concept |
| Difference between subjects and participants | PArticipants: humans Subjects: critters |
| predictive techniques | Correlation |
| Quasi experminent | an experiment in which the IV occurs naturally and there is no direct manipulation by the experimenter |
| replication | the repetition of an earlier experiment to duplicate and perhaps extend its findings |
| single blind | either the researcher or the subject doesn't know |
| sources of variability | experimental error, sample error, or a treatment effect |
| strong inference | Platt's view that scientific progress comes about through a series of of tests of alternative theoretical outcomes. |
| theory | a set of related statements that explains a variety of occurrences |
| what-if experiments | experiments performed to see what might happen rather than to test a specific hypothesis |