APRD 2004 Exam 2

Variable

something that varies

Constant

something that is fixed/does not change

Quantitative measurement

the use of numbers to describe a property of an object or an event. Describing an aspect of the "real world" using numbers

Variables can be measured at four levels

Nominal, Ordinal, Interval, Ratio

Nominal

numbers serve as tags or labels. Numbers are not placed on a meaningful scale.

Dichotomous/ binary nominal variable

a nominal variable with only two levels

Ordinal

possible values are meaningfully ordered. The do not establish the numeric difference between data points. They indicate only that one data point is ranked higher or lower than another. Distance between numbers is not set or consistent. (Check EX W6)

Interval

measured along a sale in which each position is equidistant from the other scale points. In other words, measurement intervals are equally spaced.

Ratio

ratio variables are interval variables with a natural zero point.

measurement error

when the data we collect does not represent reality. Always present to some degree.

random measurement error

measurement errors that are small, non-systematic (there is no pattern), and do not threaten the overall validity of our data. EX: a small- number of participants misread a survey question.

systematic measurement errors

measurement errors that are a big deal. An error in measurement where the tool does not accurately measure the concept and is perceived incorrectly by most or all of the participants

reliability

consistency in our measurement. Refers to the level of clarity in the tool. Refers to a measurement approach's ability to yield consistent results.

Validity

the ability of the potential of our data collection tool to capture and measure the construct or the phenomenon that we are interested in measuring. Are our questions/tests/other measures reflecting the real meaning of the concept under consideration? Refers to a measurement approach's ability to measure what it is supposed tp.

population

the entire group of people that are the focus of the study.

Why do we sample?

it is often impossible or counterproductive to collect data from all members of population.

types of sampling approaches

probability & non-probability samples

Probability samples

every element of the population has a known (though not necessarily equal) chance of being selected for inclusion. Every element has a non-zero chance of being included in the sample. Allow us to make inferences about a population.

non-probability samples

not all elements of a population have an opportunity to be included in the sample. Do not allow us to make inferences about a population.

Inference

a conclusion or opinion that is formed because of known facts or evidence

Simple random sampling

all members of a population have an equal chance of being selected for the sample. members of a population are selected at random for inclusion in the sample

stratified random sampling

a population divided into subgroups (strata). A random sample is subsequently drawn from each subgroup(strata)

disproportionate random sampling

like proportional random sampling except for the fact that sample proportions are not equivalent to population proportions.

convenience sampling

sample is drawn from those that are available or easy to collect data from

snowball sampling

generate a convenience sample of respondents. Ask sampled respondents to recommend others who might be interested in providing data.

purposive sampling

researchers purposefully select from a group of people of theoretical interest

quota sampling

generation of a sample that has attributes proportional to a given population

are convenience samples ok?

yes and no. They restrict the ability to make population-level inferences. At the same time, they are cost efficient. Not all convenience samples are created equally.

surveys

allow for data collection form a large number of people, allow for assessment of self-reported traits, attributes, and behaviors. When properly developed are a reliable means of information gathering.

Self-report data

self-report data is data provided by a study respondent without interference on the part of the researcher. Respondents tell the researcher what they think, how they feel, and how they behave

single selection measures

require participants to make a single selection from a list of options.

multiple selection measures

allow participants to make more than one selection from a list of options.

Ranking measures

require participants to rank a body of elements in terms of preference.

rating measures

require participants to rate (on a numeric scale) thoughts, feelings, or beliefs (etc.) relative to a researcher-provided prompt.

likert-type measures

require participants to indicate level of agreement relative to a researcher provided prompt.

semantic differential measures

require participants to indicate feelings/beliefs relative to a researcher provided prompt. Uses a bipolar format where only the poles are labeled.

An experiment does 3 things:

demonstrates whether something is true, examines the validity of a hypothesis/theory, and attempts to discover new information.

correlation

Variables "A" and "B" are related to one another.

time order

changes to variable A result in changes to variable B

Non-spuriousness

The relationship between variable A and variable B must not be explained by a third variable.

spuriousness

a relationship between variables that seems real, but is in fact explained by the presence of another variable.

causation

causation states that A causes B

Association

states that A and B are associated with one another.

Independent variable

are varied by researchers. the factor that the researcher changes.

dependent variable

are presumed to be affected by independent variables. The variable that is affected by the change made to the independent variable. Variable that is measured by the researcher

random assignment

all true experiments require this. Research subjects are randomly placed in experimental groups.

quasi-experiment

when an "experiment" does not use random assignment. Cannot provide evidence of causation.

random sampling

random selection of individuals from a larger population.

randomized controlled trial

most straightforward type of experiment that has 2 groups: a control group, and a treatment/experimental group.

control group

a group of participants who do everything the experiment group members do, but are not given any test, drug, invention, or manipulation.

treatment/experimental group

the group of participants who undergo a form of experimentation, such as training, taking a test or drug, or another type of intervention.

strengths of experimental research

1. only method that can definitively show causality. 2. can be replicated (repeated by another researcher)

weaknesses of experimental research

1. often, the study context is artificial 2. cross-sectional designs don't speak to long-term effects 3. in some research scenarios, experiments can raise ethical questions.

pretest

measurements taken before delivery of the experimental (or manipulated) stimuli

posttest

measurements taken after delivery of the experimental (or manipulated) stimuli

Solomon four-group design

common experimental design that contains more than two groups. It uses two

experimental groups and two control groups. From these

groups, only one experimental and one control group take

the pretest. The other two groups are subjected only to the

intervention or the placebo and do not take the pretest. The

differences in results can clearly show whether the pretest is

influencing the results of the

study

factorial

design

Factorial experiments involve

the manipulation of more than

one independent

variable. Often, these types of designs are used when there is an

expectation that two factors might work in concert with one

another to produce an effect

A/B

testing

at its most basic, is a way to compare two

versions of something to figure out which performs

better. A/B testing has numerous applications:

Marketing/marketing communications,Web design, User experience, Human factors

Types of questions appropriate for A/B testing:

Does changing the location of a design element

increase website clicks?

z

Does changing our website font increase time on page?

z

Does changing the color of

a design

element increase

clicks?

z

Does adding an interactive element decrease bounce

rate?

Content analysis

Quantitative content analysis is the systematic and

replicable examination of symbols of

communication. The systematic review of

media materials (e.g., television shows, magazine

advertisements, movies, journalistic articles, etc.) for

patterns

Content analyses are:

objective, systematic, focused on manifest content

manifest content

Content that is observable (not

inferred or assumed)

Measurable scoring units:

figuring out your basic or standard unit of

measurement. If you are doing a content analysis of the

comics page of a newspaper, you could take the comic

strip frame as your basic unit. If you are doing a content

analysis of magazine articles, you could take number of

words

Constitutive definitions

The definitions you find in dictionaries are known

as

constitutive definitions

—

they define words in terms of

other words and concepts. They are quite general and

abstrac

Operational definitions

An operational definition tells how you will measure

something and forces you to explain how you understand

or interpret a concept. Thus, if you are dealing with

media violence, you will have to describe what kind of

actions or behaviors constitute violence

codebook

the guide that coders use to code

media content

Think of it like a set of instructions

intercoder reliability

if two or more independent evaluators can

use an operational definition in a consistent manner, we can

have some confidence that our code book/coding rules are

reliable. We need the coders to agree 70% or more of the coded cases to claim intercoder reliability

Advantages of content analysis research:

Unobtrusive, Relatively inexpensive, Deals

with current events and topics of present-day interest, Uses material that is relatively easy to obtain and work with,Yields

data that can be

quantified

Disadvantages to content analysis research:

finding a representative sample can be difficult, obtaining reliability in coding can be difficult, defining terms operationally can be difficult.

content analysis research process

select a topic, identify scoring units, create a sampling plan/sample, create operational definitions, assess inter-coder reliability, code entire sample

Big data

Big data

refers to data that would typically be too

expensive to store, manage, and analyze using

traditional (relational and/or monolithic) database

systems

Big data is defined by the 3V

Volume: big data is large

Velocity: big data occurs at an unprecedented speed

Variety: big data comes in multiple formats/takes on multiple forms. (study how netflix uses big data, just in case)

Hypothesis

A formal (and testable) prediction of what will occur. Use when we have an informed guess as to what is likely to occur

research question

A question around which research activities are organized. Use when we are exploring a new area of study

variables of primary interest

variables you are explicitly interested in learning more about

control and descriptive variables

variables whose collection gives the researcher the ability to describe what the sample looks like in terms of demographics and to address spuriousness.