[Solved] Yegidis Et Al

 

Respond to two colleagues by:

• Comparing the threats of internal validity they chose with the ones you chose
• Suggesting ways in which one of their identified threats could be eliminated or its effects reduced when conducting a study

1- jasmine 

 

What is the difference between internal and external validity?

According to the text, the difference between Internal validity and external validity is that internal validity refers to the amount of confidence the researcher has that only exposure to the independent variable produced changes in the defendant variable and that there are no other factor that accounted for those changes and external validity refers to numbers and characteristics of the cases that were studied and the findings are believed to look beyond cases that were actually studied to their population.

Select two internal validity threats and describe each.

Two internal validity threats are history and testing effects. In regards to History, an event can occur during the research study that can have a major effect on the dependent variable which can change the outcome and the information that you are trying to obtain. The testing effects are a threat as the use of pretest to measure the dependent variable may result in a change of measurements of the variable.

After examining each category and list of designs, determine which design controls for all threats to internal validity. Explain how this design controls all threats. 

One way to control for all threats to internal validity is Random Assignment. “A highly desirable effect of random assignment is that, if performed correctly (and a reasonable large sample size is used), it will control for the effects of all cofounding variables, even those whose existence we have no reason to suspect. This is why random assignment is such an important feature of experimental designs”. (Yegidis et al., 2018)

Reference

2- jo

 

Internal validity refers to the degree to which an experimental design has controlled for confounding variables. If a study has internal validity, it means the effect on the dependent variable was actually caused by the independent variable and not by some other factor. Establishing internal validity is a pre-requisite for determining that the relationship between the variables is indeed causal and not merely associative or correlative.

External validity is the extent to which the research results apply to everybody, not just a specific group of people. This is also referred to as broad generalizability. For example, a study conducted on people in only one neighborhood would not necessarily apply to everyone who lives in that city. Researchers naturally want to be able to put everyone in a category because that creates less confusion. It might be tempting to say that a study applies to all Pacific Islanders, but this fails to account for the differences between the large number of groups who identify as Pacific Islanders. The external validity of a study is measured by degrees. In other words, you cannot definitively state that a study either has external validity or it does not. The question is rather – how much external validity does the study have?

There are numerous threats to internal validity, and I will discuss two of them here. One threat is referred to as testing effect. This can occur in studies where a pre-test and a post-test are given. The study participant will remember the first test, creating bias in how they answer the second test. Another threat is lack of sample comparability. This means that the experimental and control groups are not similar enough to state that the independent variable led to the effects on the dependent variable. If the two groups are not similar to begin with, there is no reason to assume that the independent variable caused changes seen in the experimental group.

The rigorous design of truly experimental research controls for confounding variables by the process of random assignment of participants to either experimental or control groups. Participants are selected from a very large group, with each member of the group having an equal chance of being selected for the study. This prevents sampling bias and assures that the results of the study truly are representative of the entire population. Randomly assigning participants also controls for any unseen variables.