8 Data collection and measurement

physical or biological measures

Measured using specialized equipment/techniques E.G temperature blood glucose x-rays wound cultures

Structured observations

Researcher makes general notes about observed behaviors/phenomena

Structured observations

Observing people’s behaviours using a plan organized system to record what happens E.G category systems, checklist, rating scales

Hawthorne effect

People act different when they know they are being observed

Self reports : Interviews

Verbal interaction between researcher and participant, structured, semi, structured, or structured, interviewer records data

Self reports : Surveys and questionnaires

Printed/written/electronic form, structured, usually complete completed by participant

Reading scales

Rate a variable on a numerical scale

Likert scales

Measure opinions or attitudes on a scale E.G strongly agree, strongly disagree

Visual analogue scales

Plot response on a 100mm line with descriptors on each end

Social desirability bias

When participants have the tendency to misrepresent their opinions in the direction of answers, consistent with prevailing social norms

Extreme responses bias

When participant select extreme response options E.G strongly agree, never, regardless of what the item is asking

Acquiescence bias

The tendency for participants to agree with research statements without it being a true reflection of their own position or the question itself

Records and available data

Data that has been gathered previously for another purpose (secondary data)

Directness of measurement

Degree to which object, element, characteristic can be measured

Direct measurement

Measuring something directly using a measurement tool or technique E.G temperature, weight, oxygen saturation

Indirect measurement

Measuring a concept that cannot be measured as directly often using a tool or questionnaire E.G depression, self-care, coping, pain

Measurement error

Difference between what exists in reality and what is measured by research instrument

Random error

Difference between the obtained measure and the actual true value pattern

Systematic error

Measurement, error that occurs consistently in the same direction, not random. Everyone has the same error.

Reliability

How consistently does the measurement technique measure the concept of interest, range from 0 to 1.00, good 0.80, acceptable 0.70, any lower not reliable

Test-retest reliability

Degree to which retesting with the same instrument will yield the same result, assumes the factor being tested, remains the same at both testing times

Interrater reliability

Degree to which measurements remain consistent between observers, tool used for collecting data should be specific enough so that two observers would get the same results

Internal consistency

Degree to which all items in an instrument consistently measure the concept, usually measured with Cronbachs alpha (0-1, 0.80 good)

measurement validity

Degree to which an instrument measures the construct it is meant to measure

Face validity

Degree to which the instrument looks like, it is measuring the target construct, not strong evidence of validity, but may influence subjects attitudes toward the data measurement process

Construct validity

Degree to which a tool measures the abstract concept it is meant to measure, complicated to determine

Content validity

Degree to which the content of an instrument adequately captures the relevant aspects of the construct being measured, usually assessed by expert panel, maybe measured as content validity index (CVI)

Criterion related validity

Degree of relationship between the measurement and actual performance/behavior

Concurrent validity

Degree of correlation of two measures of the same concept at the same time

Predictive validity

Degree of correlation between the measure of the concept, and some future measure of the concept, usually assessed by expert panel

nominal

Categories with no natural order, you can identify which category something belongs to, but you cannot rank the categories from lowest to highest

Ordinal

Has a meaningful order, but the distance between categories is not established as equal (e.g never, sometimes, often, always) you know which response is higher or lower in the ranking, but you don’t know whether the difference between each categories equal

Interval

Has equal distances between values, but zero does not mean the complete absence of the measured quantity (e.g IQ scores), no absolute zero

Ratio

Have equal intervals and a true absolute zero (e.g height, weight, age)