Determination of the quantity of a characteristic that is present; involves assigning of numbers or some other classification.
Measures that involve sorting of subjects into categories based on their characteristics.
Clearly stated meaning of an abstract idea or concept used by a researcher in a study.
An explanation of the procedures that must be performed to accurately represent the concepts.
Data collected directly from the subject for the purpose of the research study. Examples include surveys, questionnaires, observations, or physiologic studies.
Data collected for other purposes and used in the research study. Examples include patient medical records, employee or patient satisfaction surveys, organizational business reports, or gov databases.
Data that can be named and placed into categories but cannot be ranked or measured on a scale.
Categorical data that can be put in rank order. The scales contain intervals between entries that vary, limiting statistial analyses and comparisons across the scales or between subjects.
Data measured on a scale that has consistent intervals between measurement units and allows for broad selection of mathematical operations and analytic options.
Data measured on interal scales that have a true zero.
The difference betwen the actual attribute (true score) and the amount of attribute that was represented by the measure (observed score)
A nonreproducible error that arises from a variety of factors in measurement. These errors do not affectaverage scores in a data set but do afect the variation that exists around the average.
A bias in measurement that is consistent but not accurate and that underestimates, overestimates, or misses data in a way that is not random.
The use of procedures to minimize measurement error with physical instruments by objectively verifying that the instrument is measuring a characteristic accurately.
The degree of reproducibility or the generatiln of consistent values every time an instrument is used.
The extent to which an instrument is consistent within itself as measured with the alpha coefficient statistic.
The extent to which an instrument is consistent across raters, as measured with a percentage agreement or a kappa statistic.
The ability of an instrument to consistently measure what it is supposed to measure.
A subjective judgement about whether a measurement makes sense by assessing that itmes of the instrument are the attributes being measured (face validity) or by verifying items with a panel of experts.
An outline for determining content validity that includes the analysis of basic content and the assessment objectives.
An indication that a measurement captures the abstract concept that is the basis of the study. A common method of construct validation is factor analysis.
A correlation of the research instrument to some external manifestation of the characteristic.
A measurement of criterion-related validity that is present when an instrument reflects actual performance.
A measurement of criterion-related validity that is indicated when an instrument can predict future performance.
A measurement of criterion-related validity that is demonstrated by the instruments capacity to differentiate those who have a characteristic from those who do not.
A measure of discriminate validity in the biomedical sciences that indicates an instrument has the capacity to detect disease if it is present.
A measure of discriminate validity in the biomedical sciences that indicates an instrument has the capacity to differentiate when the disease is not present.
A measure that indicates change in the subject's condition when an intervention is effective.
A qualitative data measure focused on the stability of the information across individuals or over time.
A qualitative data measure focused on ensuring that the results represent the underlying meaning of the data.
To represent the underlying characteristics, numbers used in research measurements must be:
Clearly linked to the research question
Appropriate to represent the variable of interest
The first step of the measurement strategy is to give careful thought to the concepts represented in the research question.
The research question will describe the phenomena or characteristics of interest in a study. These must be translated into observational attributes before they can be measured. Eventually these concepts may be represetned as a physical attribute. These are called attribute variables.
A conceptual definition describes the concept that is the foundation of the variable by using other concepts.
Operational definitions ensure that the researcher is measuring attributes reliably.
Data are considered the most reliable because the data are collected by the researcher for a single specific purpose but are time consuming to collect and the quality depends on many factors such as subjects memory or problems communicating.
Data are often easier and quicker to collect. Retrieved from data sets that have already been collected, usually for another reason.
The _____ is the best and most sensitive source of objective data regarding patient conditions.
This type of data includes gender, marital status, ethnicity, diagnosis, and other variables that can be measured on a scale. Easy to collect and summarize yet they are the least sophisticated type of measure.
A pain scale is an example of ____ data. There is a high and low end of the scale---or good and bad, big and little---so the categories can be ranked.Each subject response can be placed at only one place on the scale.
The primary distinction of ordinal data is that the entries on the scale cannot be directly compared across the scale or between subjects because intervals between entries may not be the same.
_____ is a threat to the internal validity of a research study, so it means that the overall study results are more credible.
Random error by and large does not affect the average scores in a data set, but it does affect the amount of variation that exists around the average.
Systematic error has a more serious effect on the results of a research study because measures with systematic error may appear to be accurate.
The primary way to minimize measurement error with physical instruments (blood glucose monitor) is through ____
Using multiple instruments reduces measurement error because reliability increases when agreement is observed across multiple ways of measuing the same concept.
Measured with the alpha coefficient statistic.
The alpha coefficient should have a value of 0.7 or greater. Cronbachs alpha represents the extent to which the variability on individual items represents the variability in the overall instrument.
Stability among individuals is measured by an item , _____, which should have a positive sign and an absolute value close to 0.5.
This statistic focuses on the degree of agreement between raters and generates a p value, reflecting the statistical significance of the agreement.
An instrument will only be as strong as its reliability
Construct validity may be the most important type of validity test to ensure that results will represent reality. A common method of construct validation is factor analysis which groups itens within the instrument according to their shared variability.
Most validity tests use the correlation coefficient to represent the degree of relationship between the instrument and the reference.
Sensitivity helps the researcher avoid false negatives and specificity helps the researcher avoid false positives.
Random error is expected in a research study but systematic error will bias the results.
The level of measurement drives the type of statistical analysis that can be conducted to answer the research question.
The reliability and validity of an instrument are the most important characteristics and they should be documented in the research report. Using an existing instrument is desirable for its efficiency and the capacity to provide a comparison with existing studies.