Reliability and Validity

Question 1

What is meant by reliability in measurements?

Answer 1

Reliability describes the stability of measurements. If encapsulates the degree to which the measurements change when the dependent variable is subject to the same stimuli. Should the measurements change wildly, we say that the reliability is low since the effect of the stimuli is unpredictable. Low levels of reliability will impair the research efforts as unreliable measurements makes hypothesis testing much more challenging. We can accept some unreliability by compensating with more data

Question 2

What is meant by validity in measurements?

Answer 2

Validity is the correctness of the measurements. If we drive at a constant speed but the measurements vary significantly, we have invalid measurements. High levels of validity is a necessity since no research can be based on inaccurate data.

Question 3

What are the different types of reliability?

Answer 3

1. Retest reliability assert whether similar results are achieved when recently conducted test are carried out once more.
2. Inter-rater reliability assert whether similar results are obtained when other researchers carry out the same tests.
3. Parallel forms (or Equivalent forms) reliability measures the correlation between two equivalent versions of a test. An exam should have high parallel reliability with the exam in the same course from the year prior.
4. Internal consistency measures whether different items (questions/procedures) within a test do in fact measure the same thing. Should the replies to two different questions contradict one another then the test might prove unreliable..

Question 4

What are the two kinds of validity?

Answer 4

1. Internal validity encapsulates the extent to which the design of the research study and data collected allows the researchers to draw accurate conclusions concerning cause and effect. The internal validity of a research study hinges on whether the researchers have taken the necessary steps to eliminate all other explanations.
2. External validity describes whether the results provided by the study can be generalized. This is highly important as most studies are conducted in highly controlled environments which seldom reflect the real world. Studies with high internal but low external validity are of little value. Methods frequently utilized for ensuring high external validity is to conduct your studies in a real-world setting, i.e., not a lab, gather a representative sample, and replicate the study in different settings.

Question 5

What are the main methods of testing validity?

Answer 5

1. Face validity is the degree to which a research strategy seems to accurately measure a given characteristic.
2. Content validity is whether an assessment instrument or procedure adequately reflects the full breadth of the characteristic being assessed.
3. Criterion validity asserts the extent to which the results from an assessment strategy correlates with similar assessment strategies for testing the same phenomena. E.g, how well does your strategy align with the most popular assessment strategy in the field.
4. Construct validity evaluates whether the measurement tool used really measures what we are looking to measure.