You will recognize same of these classic threats from the previous chapters covering nonexperimental and quasi – experimental design. All quasi – experimental designs contain one or more classic threat threat to internal validity – in fact, that is the reason they are not true experimental ; they are potentially confounded. As you are learning the classic threats, see if you can recognize which threats are inherent to each of the quasi – experimental designs from chapter 4 : longitudinal, cross – sectional, ex pos facto, and pretest/posttest designs.
Remember that a confound is and extraneous variable that varies systematically along experimental conditions. An extraneous variable that pop up randomly across conditions is not aconfound – it simply increases the amount of variance in the scores that we call error, making it harder to detect a significant treatment effect. Designs using different subjects in each treatment group can be confounded if an extraneous variable affects some experimental groups but not others with regularity. (these are called between – subjects designs, and we will proceed to them in cha[ter 8.) Other designs in which subjects are measured multiple times can be confounded if an extraneous variable is present only in certain experimental conditions but not in others. (these are called within – subjects design, and we will deal with them in chapter 10.)
Whenever you design or evaluate experimental research, you must consider each of these eighth classic threats to internal validity. In chapter 8, we will begin learning about the various types of experimental designs. Controlling for extianeous variables the might confound experimental is one of the most important elements of the research design. If one or more of these threats are present, the experimental will lack internal validity, and you can never be sure that the effects produced on the dependent variable were really cause by the experimental manipulation.
History
The first threat is called history, and it refers to the history of the experimental. Ask yourself whether any outside event or occurrence, rather than the independent variable, could have caused the experimental effects. (It need not be a great historical event, although it could be.) History is most often a problem when a whole group of individuals is tested together in the same experimental condition. Some outside event that occurred prior to their group testing session could influence responses of the entire group, and effects produced by the event could be mistaken of effects of IV. History effects can be problematic.
Suppose you were testing two different weight – loss programs in which subjects were exposed to your treatments during daily group mectings. You assessed the benefits of the program by measuring how much weight each person lost at the end of a seven – day program. After weighing subjects in both groups, you discover that individuals who received treatment B lost an average of 2 pounds that week; whereas the average for treatment A was 4 pounds. You want to make sure that the “history” of both groups before weighing was the same. Imagine the confound in your experimental if individual in group B were weighed right after lunch, but the subject in group A were weighed just before lunch. You would not know if the differences you observed were caused by your diet program or by the fact one group had just eaten and other had not.
Maturation
The second classic threat, maturation, refers to any internal (physical or psychological) changes in subjects that might have affected score on the dependent measure (not just “maturing” in the way we typically think of it, although it could be.) The kinds of kinds of internal changes that we usually worry about in experimentation are things like boredom and fatigue that can occur during a single testing session, Boredom and fatigue are more likely in within – subjects designs that require lengthy testing sessions. You will learn that there are techniques for
You will recognize same of these classic threats from the previous chapters covering nonexperimental and quasi – experimental design. All quasi – experimental designs contain one or more classic threat threat to internal validity – in fact, that is the reason they are not true experimental ; they are potentially confounded. As you are learning the classic threats, see if you can recognize which threats are inherent to each of the quasi – experimental designs from chapter 4 : longitudinal, cross – sectional, ex pos facto, and pretest/posttest designs.
Remember that a confound is and extraneous variable that varies systematically along experimental conditions. An extraneous variable that pop up randomly across conditions is not aconfound – it simply increases the amount of variance in the scores that we call error, making it harder to detect a significant treatment effect. Designs using different subjects in each treatment group can be confounded if an extraneous variable affects some experimental groups but not others with regularity. (these are called between – subjects designs, and we will proceed to them in cha[ter 8.) Other designs in which subjects are measured multiple times can be confounded if an extraneous variable is present only in certain experimental conditions but not in others. (these are called within – subjects design, and we will deal with them in chapter 10.)
Whenever you design or evaluate experimental research, you must consider each of these eighth classic threats to internal validity. In chapter 8, we will begin learning about the various types of experimental designs. Controlling for extianeous variables the might confound experimental is one of the most important elements of the research design. If one or more of these threats are present, the experimental will lack internal validity, and you can never be sure that the effects produced on the dependent variable were really cause by the experimental manipulation.
History
The first threat is called history, and it refers to the history of the experimental. Ask yourself whether any outside event or occurrence, rather than the independent variable, could have caused the experimental effects. (It need not be a great historical event, although it could be.) History is most often a problem when a whole group of individuals is tested together in the same experimental condition. Some outside event that occurred prior to their group testing session could influence responses of the entire group, and effects produced by the event could be mistaken of effects of IV. History effects can be problematic.
Suppose you were testing two different weight – loss programs in which subjects were exposed to your treatments during daily group mectings. You assessed the benefits of the program by measuring how much weight each person lost at the end of a seven – day program. After weighing subjects in both groups, you discover that individuals who received treatment B lost an average of 2 pounds that week; whereas the average for treatment A was 4 pounds. You want to make sure that the “history” of both groups before weighing was the same. Imagine the confound in your experimental if individual in group B were weighed right after lunch, but the subject in group A were weighed just before lunch. You would not know if the differences you observed were caused by your diet program or by the fact one group had just eaten and other had not.
Maturation
The second classic threat, maturation, refers to any internal (physical or psychological) changes in subjects that might have affected score on the dependent measure (not just “maturing” in the way we typically think of it, although it could be.) The kinds of kinds of internal changes that we usually worry about in experimentation are things like boredom and fatigue that can occur during a single testing session, Boredom and fatigue are more likely in within – subjects designs that require lengthy testing sessions. You will learn that there are techniques for
การแปล กรุณารอสักครู่..
