Experimental Methods and Measures Flashcards
Content Validity
Four types of test validities are used to examine how well a research design tests what it was supposed to test. Content validity refers to the extent to which a design assesses the full scope of content that researchers intend to measure.
Construct Validity
Construct validity tells us how well a test measures what it is supposed to test.
Criterion Validity
Criterion validity refers to the extent to which a test correlates with an established “criterion” or existing measure.
Predictive Validity
Predictive validity refers to the extent to which a test predicts a later score on some other test.
An experiment is carried out to measure the relative effects that pregnancy has on the intelligence of pregnant mothers. Measurements of IQ are taken before, during, and after pregnancy. Later, critics of the study argue that testing IQ only captures some aspects of intelligence while missing others. These critics would argue that the study lacks which kind of validity?
Content Validity
An evaluation is designed to predict whether individuals will develop diabetes. This evaluation is administered repeatedly over time with highly consistent results. However, it is later found that the test is extremely inaccurate in predicting diabetes onset and has several design flaws. This evaluation has:
Low validity but high reliability
True or false: An experiment with high validity also had relatively high accuracy.
This statement is false. If a scientist measures every data point incorrectly this would produce “inaccurate” data. However, if by chance the error of each data point was distributed to produce a resultant mean that accurately represented the data set, the experiment would be considered “valid”. In other words, each individual measurement can have low accuracy, but this can still sometimes produce a “valid” data set.
Acquiescence bias
Acquiescence bias is the tendency to agree with a statement, especially one that the respondent does not fully understand or is uncertain about.
Quantitative methods always produce numbers
True
Objective Measures
Unbiased and faced based
Subjective Measures
Subject to opinion
True or false: Quantitative research designs would be best for investigating a scientific topic in a field where few prior studies have been conducted.
This statement is false. In fields where not much knowledge has been collected, qualitative research, which is more exploratory, may provide more open-ended results. These open ended results can be used to come up with future research questions to be answered with quantitative designs.
True or false: Subjective research designs are non-numerical by their intrinsic nature.
This statement is false. Subjective research is research where data collection is colored by participants’ intrinsic biases. For instance, questions such as, “How are you feeling today?” However, experimenters can have participants answer these questions by using numerical scales (e.g., “How are you feeling today – on this scale from 1 to 10?”).
Extent to which a study’s results are both genuine and gernalizable
validity
Internal Validity
The ability to draw causal conclusions, can we reliably say that a change in X explains the change in Y. Causal! The fewer confounding variables the better