MSW 550 Quiz 2
Validity
"Measure is accurate" - The ability of a test to measure what it is intended to measure
Reliability
"Measure is consistent" - yields same result each time - not same as being accurate
Sources of SW research questions:
- Info needed to solve social welfare problems - Info needed for practice, policy or planning decisions - Curiosity and personal interests
Research questions need to be
- Narrow and specific - Answerable by observable evidence -Relevant to social welfare policy/practice
Literature reviews help researchers to:
- Select a research question - Identify obstacles with topic, and how others overcame those - Decide which type of research best builds on current knowledge
Three choices for operationally defining variables in social work:
- Self-report - Direct observation - Examination of available records
Levels of Measurement: Interval
- attributes on a continuum with equal distances between points - NO absolute zero Ex. range of IQ between 100 and 110 is same as 120 and 130. Nobody has zero IQ.
Levels of Measurement: Ratio
- attributes on a continuum with equal distances between points - has true/absolute zero Ex: time, distance, time spent watching TV, number of words per story
Conceptualization in qualitative inquiry
- frequently not operationalized in advance - observations help us understand new phenomenon or build deeper understanding of known one
To ensure your question is sensitive to needs, lifestyles, and concerns of study population:
- review current literature - consult leaders in the community - consult relevant service agencies
Variables
A concept being investigated, characterized by different attributes or categories (ex. Gender, SES, ethnicity, age)
Moderating variable can also be
An alternative explanation for relationship observed between IV and DV
inter-rater reliability
Consistency of scoring by two or more raters using the same item, scale, or instrument
systematic error
Consistent, directional error in measurement due to: - investigator bias (esp. Leading questions) - subject "acquiescence response set" (all yes or all no answers, regardless of truth) -subject "social desirability bias" (answers in ways that present self/group in positive light)
IV and DV relationship changes at different levels of a variable
Curvilinear relationship
random error
Error that is due to chance and is not standardized. Decreases precision.
spurious relationship
False association between two variables that is actually due to the effect of some third variable
test-retest reliability
Found by comparing a test taker's scores on the same test taken on separate occasions
Operationalization
How variables are defined and measured
Criteria to select an existing scale:
Length Difficulty Reliability Validity
criteria-related validity (3 types)
Measures against a known. - Concurrent: degree to which a measure correlates with a known measure of the same concept (ACT scores and GPA) - Predictive: degree to which a measure accurately predicts an outcome compared to known indicator (GPA predicts success better than ACT) - Known groups: degree to which measure differentiates between groups known to differ (KKK vs. social worker scores on racism measure)
face validity
Measures whether a test looks like it tests what it is supposed to test. - does NOT mean it *accurately* tests for the concept
mediating variable
Mechanism by which IV affects DV
concept
Mental image that symbolizes an idea, object, event, behavior or person
Levels of Measurement: Nominal
NAMES, LABELS, CATEGORIES Cannot be arranged in an order (e.g., eye colors)
As the IV increases or decreases, the DV changes in the OPPOSITE direction (ex: IV increases, DV decreases).
Negative relationship
Levels of Measurement
Nominal, Ordinal, Interval, Ratio (NOIR)
Qualitative Validity
Obtaining impressions, patterns or trends of the experience or topic under investigation. Good because no quantitative measure is 100% accurate
Error in existing records reviews
Over/under reporting in original records/notes (ex. Because practitioner wants to look good on paper)
As IV increases or decreases, the DV changes in the SAME direction (ex. IV increases, DV increases)
Positive relationship
Conceptualization
Process of identifying, selecting and providing theoretical rationale for a chosen topic of study
Levels of Measurement: Ordinal
Ranks objects by degree (eg. non smoker, 1/2 pack/day, 1 pack/week, etc) - Distance between categories is not equal - no absolute zero
Considerations for feasibility of a topic:
Scope, time required, cost, ethical considerations, cooperation of others
Hypothesis
Testable statement about how changes in one variable are expected to change another variable
Independent Variable (IV)
The "cause" - variable that is manipulated
Dependent Variable (DV)
The "effect" - variable being measured, explained, caused
construct validity
The degree to which the content of a test is representative of the domain it's supposed to cover. Convergent: two depression tests yield same result Divergent: depression and anxiety tests yield different results
Control Variable (CV)
Variables held constant during an experiment
Error in self-report measures
Vulnerable to social desirability bias and outright deception
Error in observation
Vulnerable to: - Data values recorded incorrectly due to observer bias - Subject modifying their behavior according to social desirability bias
moderating variable
influences the direction or strength of the relationship between IV and DV
content validity
the *extent* to which a test samples the concept of interest Ex: high validity = depression test asks about behavioral, emotional and physical health
Triangulation
the use of multiple measures to determine validity of data. Ex: 3 different depression instruments yield same results
inter-item reliability
assesses the degree of consistency among the items on a given scale/measure