Chapter 12- Nonexperimental and Quasi-Experimental Designs Flashcards
correlational and nonexperimental strategies
Using these strategies, MAJOR threats to internal validity and no real attempts to control the threats. In conducting studies using these strategies, we can say that variables are related, but we certainly can’t make any kind of causal statements. We can’t say that one variable causes another.
scatter plot
A graph that shows the data from a correlational study. The two scores for each individual appear as a single point in the graph with the vertical position of the point corresponding to one score and the horizontal position corresponding to the other.
nonequivalent group designs
The researcher studies multiple groups of individuals, but these groups are generally intact- the researcher cannot randomly allocate people to groups. Because random assignment is missing, we cannot call the design a true experiment. Example: employees in workplaces or children within schools in classrooms- already intact groups, no control… could lead to assignment bias
experimental strategy
as noted above, we implement strong measures to eliminate threats to internal validity and exert high levels of control. If done well, we can make causal statements using experiments!
assignment bias
reflects the idea that the process we use to create groups to study and compare results in groups that are different on some important variable or characteristic. The groups may not be equivalent in terms of important characteristics relevant to the study and so individual differences are a large potential threat to internal validity. Quasi-experiments make some attempts to minimize the threat to internal validity of assignment bias whereas nonexperiments don’t
quasi-experimental strategy
POTENTIAL threats to internal validity (ex: perhaps we couldn’t randomly allocate participants to groups, which makes this unable to be an experiment), but attempts are made to manage/control the threats that exist.
within-subjects design
where we measure people over time. There are a large number of potential internal validity threats when we measure people over time (maturation, progressive error, history, etc). If we are unable to (or don’t) randomly allocate people to an order of an experiment, then we may be in this situation of having either a nonexperiement or a quasi-experiment. How the researcher deals with the internal validity threats determines which category it falls into.
correlation
A statistical value that measures and describes the direction and degree of relationship between two variables. The sign (+/−) indicates the direction of the relationship. The numerical value (0.0 to 1.0) indicates the strength or consistency of the relationship. The type (Pearson or Spearman) indicates the form of the relationship. Also known as correlation coefficient.
pearson correlation
A correlation used to evaluate linear (straightline) relationships.
SPEARMAN CORRELATION
A correlation used with ordinal data or to evaluate monotonic relationships.
MONOTONIC RELATIONSHIP
A consistently one-directional relationship between two variables. As one variable increases, the other variable also tends to increase or tends to decrease. The relationship can be either linear or curvilinear
coefficient of determination
The squared value of a correlation that measures the percentage of variability in one variable, which is determined or predicted by its relationship with the other variable.
statistical significance of a correlation
In a correlational study, the correlation in the sample is large enough that it is very unlikely to have been produced by random variation, but rather represents a real relationship in the population.
PREDICTOR VARIABLE
In a correlational study, a researcher often is interested in the relationship between two variables to use knowledge about one variable to help predict or explain the second variable. In this situation, the first variable is called the predictor variable.
CRITERION VARIABLE
In a correlational study, a researcher often is interested in the relationship between two variables to use knowledge about one variable to help predict or explain the second variable. In this situation, the second variable (being explained or predicted) is called the criterion variable.
