Week 4 notes
Popular in Research Methods in Criminology
verified elite notetaker
Popular in CRIMINOLOGY
verified elite notetaker
This 4 page Class Notes was uploaded by Janaki Padmakumar on Saturday September 17, 2016. The Class Notes belongs to CCJ3701 at University of Florida taught by Chris Gibson in Fall 2016. Since its upload, it has received 7 views. For similar materials see Research Methods in Criminology in CRIMINOLOGY at University of Florida.
Reviews for Week 4 notes
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 09/17/16
Week 4 notes Research Methods L5 Measurement, validity and applicability Abstract concepts being measured- violence, delinquency, substance use, etc. Many researchers use different measures to examine the same concept- difficult to compare results between researchers since different measures are being used Measurement: Concept- derived from theories, ideas or prior research Conceptualization- define what the term is (topic of research) o Important for people to understand (across subjects or researchers) what a concept means o E.g. Substance abuse- what observations should be associated with it? To obtain a measurement, potential observations/indicators should be defined Barely able to speak, leaning on a building with liquor bottle College student drinking heavily at party Someone in AA drinking a beer o Some concepts of interests don't have multiple indicators (like sex- ask if male or female) o Issues like substance abuse have multiple indicators since multiple substances can be used o None of these examples can clearly be used to define substance abuse or not because they can be interpreted in different situations/scenarios - all very different observations that are perhaps related to same issue o DSM-IV definition "Repeated use of a substance to the extent that it interferes with adequate social, vocational and self-care functioning" Operationalization- clear definition of concept and measures being used (what operations are used based on concept definition) o Process of specifying operations that indicate variable values for each case o Indicators are used to measure concepts o Can use different operation types given concept of interest Asking questions Observing social interaction Taking urine or blood samples Taping physical spaces Retrieval of official data and statistics o If multiple measures of same data exist, include those to triangulate results (e.g. like self reports and official data being used in conjunction) o E.g. measuring disorder in a neighborhood- objective measurement vs. perceptions of residents Measurement- values assigned that differ between units of analysis; can be classified or grouped- determines statistical analysis to be used o Complexity of mathematical means that are used to express relationships between variable values Nominal (weakest)- attributes are only named Categorical or qualitative Values do not have mathematical interpretation Vary in kind or quality, but not amount Attributes, not values Does not represent more or less than States if groups are same or different e.g. male or female (0 or 1) Types of nominal measures: race, gender, sex, occupation, hair color, political affiliations, etc. Ordinal- attributes can be ordered Categories that can be ranked- more than or less than classifications Specifies order of cases and how they are classified Discrete categories with >/< distinctions Likert scale- strongly agree, agree, neutral, disagree, strongly disagree Not a problem, problem, big problem Educational levels o Continuous/ quantitative measurements Numerical level of measurement Interval- distance is meaningful Different from one another More than or less than How much more/less than No fixed zero point Can add and subtract, equal interval distances Ratios are not meaningful *think temperature- this can extend into the negatives, which is ok with interval, but not ratio Ratio- absolute zero (zero means zero) Fixed zero point Multiply or divide Can create ratios (how many times greater) Age, number of arrests, number of church burnings, length of residence, number or times married o Collect at the highest complexity possible so that it can be collapsed down Multiple items measures Why not use single items o Single items do not make responses that stay consistent over time- highly unreliable Can be "yes" today, switch to "no" tomorrow o Concepts have a broad scope and cannot be easily evaluated with one question Multiple item scales: o Scales have multiple items o These items are combined or summed o Each item needs to measure something with an underlying, quantitative measurement continuum o Respondents give ratings of each item o Likert scale items L6 Measurement quality Definitions and term application questions List RQ and ID independent and dependent variable Readings and lectures on study guide Applying terminology to research scenarios- variable identification, unit of analysis How do you measure something well? Psychometrics- methods by which we go about examining measurement quality Two components make up psychometrics: o Reliability: measure consistently measures a factor; yields consistent scores when the item being measured remains constant Consistency of measurement o Validity: measure what its intended to measure; match between conceptual and operational definition Accuracy of measurement Important because it generated statistical standards for measurement quality Self reporting- college alcohol surveys about drinking behaviors o Measure twice within a short interval of time o Ask subject 1 how much they drink- answer remains not at all- reliable and valid measure o Subject 2 says not at all both time (reliably measuring), but he has a drink in hand (inaccurate measure) Combining different measures allow for triangulation- increase in validity Four types of reliability (consistently measures something that is assumed not to change): Inter-rater reliability o Observation in neighborhoods checklist for physical disorder in neighborhood- abandoned buildings, grafitti, damage, bars on windows etc. o Ask surveyor to go into neighborhood and record area- return to lab and see if developed checklist can reliably measure disorder in neighborhood o Give systematic checklist to two different people and separate them- let them watch the same video o Each mark what they see in video about disorder o If checklist is reliable, person 1 and person 2 should have very similar results (high correlation means developed instrument is reliable) o If low correlation between raters, instrument is unreliable Wording in checklist, definitions, instruction confusing all can contribute to lack of correlation Test-retest reliability o One person measured twice- test and subsequent retest o E.g. IQ testing- Wechler's test for verbal intelligence- administer to a person at time 1, and administer to same person at time 2 o If measure is reliable, then very similar scores should be obtained if taken within a short time difference o Requires financial resources Parallel forms reliability o Apply randomly to two groups- should be a correlation if measuring the same thing o Less commonly used Internal consistency reliability o Most commonly used in criminology and psychology o Multiple item scales- narcissistic personality inventory- has multiple statements participants respond to; using statistics, responses to each attributes are formed into a composite score o Score represents responses to all items on that measure concept o Need to assess how multiple responses remain consistent o Coefficient should be close to 1- increase in reliability; if closer to 0 then unreliable; R>.8 to claim reliable measure
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'