Com 309 Exam 3
A statistical concept that involves subtracting a number, usually 1 or 2, from a sample (N or n), a group (K), a row, a column (C), or other subject designation, such as: N - 1, N - 2, K - 1, R - 1, and so on.
Degrees of freedom
The fear of making a statistical error lead to the development of this concept
Degrees of freedom
_________ use two primary methods to make data more manageable: data distributions and summary statistics
Descriptive statistics
The more time students spend on studying, the higher grades students will get on their exams" is an example of a __________ o Directional hypothesis o Non-directional hypothesis
Directional hypothesis
Two-tail test
Does not predict a direction
The survey question, "Do you still use marijuana?" is an example of a __________ question
Double-bind
Prestigious bias
Due to respondents' feelings of inadequacy or lack or knowledge about a particular topic, they often provide "prestigious" answers rather than admit to not knowing something
Thurstone Scale
Equal appearing interval scales, the techniques used to develop it are typically used to measure attitude toward a given concept or construct · Collects a large # of statements (at least 100) · Judges rate statements along an 11-category scale in which each category expressed a different degree of favorableness toward the concept o Items are ranked according to mean and median ratings · Then used to construct a questionnaire of 20-30 items where statements are worded so a person can agree or disagree with them
T or F: Correlation implies causation
FALSE; correlation does not imply causation
T or F: A correlation of -.83 is a weaker relationship than +.23
False
T or F: a distribution cannot have more than one mode
False
T or F: terms that cannot be quantified can be used in a hypothesis
False
T or F: Computation of power value helps control the Type I error
False; It helps control the type II error As power increases, there is no direct effect on Type I error: power acts independently of Type I error
T or F: A RQ is a prediction made about the effect an IV has on a DV
False; RQs do not make predictions, hypotheses do
T of F: there is not a way to depict Type I and Type II errors without considering the null hypothesis
False; Since most studies do not state the null hypothesis because it is assumed, there is a way to depict Type I and Type II errors without considering the null hypothesis
T or F: Telephone surveys are less expensive than mail and more expensive than face-to-face interviews
False; Telephone surveys are more expensive than mail and less expensive than face-to-face interviews
T or F: The mean is not sensitive to outliers and extreme outliers
False; Unlike the mode & media, mean takes all the values in the distribution into account which makes it sensitive to outliers and extreme outliers
T or F: Significance can be interpreted independently of the context in which it occurs
False; it cannot
T or F: In private sector research, it is necessary to tell the respondent how long the survey will take, e.g., "this survey will only take a few minutes."
False; it gives the respondent opportunity to say they do not have a few minutes
T or F: Newspaper and magazine ublishers prefer to measure circulation over reach
False; they prefer to measure reach
The sum of all scores (ΣX) divided by the total # of scores (N): X̄ = ΣX N
Formula for calculating the mean
The _______ scale is common in science, sociology, public opinion research and anthropology
Guttman
Audience rating =
Households/Population
1. Read questions exactly as worded and in the same order. Skip only when instructed. No exceptions 2. Never suggest, try to explain, question, or imply a question or reply. 3. If question is not understood can read again. If still not understood, record "no answer" 4. Report answers and comments exactly as given 5. Act interested, alert, and appreciative of the respondent's cooperation but never comment on their replies or express approval, disapproval, or even surprise. 6. Follow all instructions carefully 7. Thank each respondent 8. Discuss any communication problems immediately with the researcher in charge
How to minimize interviewer bias
"Students' anxiety level on giving public speech is associated with his/her grade in public speaking class," is an example of a ___________ o Research Question o Hypotheses
Hypotheses
"The more time students spend on studying, the higher grades students will get on their exams," is an example of a ___________ o Research Question o Hypotheses
Hypotheses
_________ is a formal statement predicting an outcome of the relationship between two variables
Hypotheses
· To determine the statistical significance of a research study, the researcher must set a probability level, or significance level, against which the null hypothesis is tested · If the results indicate a probability lower than this level, the researcher can reject the null hypothesis · If the outcome has a high probability, the researcher must support (or, more precisely, fail to reject) the null hypothesis · Since the null hypothesis is not generally stated, acceptance and rejection apply to the research hypothesis, not the null hypothesis
Hypothesis testing (accepting or rejecting the null hypothesis)
Qualitative Research
Involves several methods of data collection: focus groups, field observations, in-depth interviews, and case studies
Quantitative Research
Involves several methods of data collection: telephone, internet, and mail surveys · Questions are static or standardized
__________ surveys are usually difficult to conduct as most are thrown away, response rate is usually very low—under 40%
__________ are a quick and easy way to collect personal data but use convenience sampling so results are not generalizable to the population
Mall interviews
To determine the ______, the scores are first ordered from smallest to largest
Median
"The more time students spend on drinking, the less hours of sleep students will get," is an example of a __________ o Positive hypothesis o Negative hypothesis
Negative hypothesis
"A student's anxiety level on giving public speech is associated with his/her grade in public speaking class," is an example of a __________ o Directional hypothesis o Non-directional hypothesis
Non-directional hypothesis
______ sampling involves non-random selection based on convenience or other criteria and is not intended to represent the population
Non-probability
In hypothesis testing, or significance testing, the researcher either rejects or accepts the ___________
Null hypothesis
__________ is a statement that statistical differences or relationships among phenomena have no difference (comparison) or no association (relationships).
Null hypothesis
The 3 central concepts of measurement
Numeral, assignment, and rules
Nominal (4 Levels of Measurement)
Numerals or other symbols are used to classify people, objects, or characteristics (e.g., rocks) · The numerals are simply labels that stand for respective categories · Basic property is equivalence (all categories are equal) · All categories are exhaustive and mutually exclusive
In Pearson's R, what is the lowest r value possible?
0.00
1. Should be compatible with current knowledge in the area · If available literature strongly suggests on point of view, researchers who develop hypotheses who oppose this knowledge without basis only slow development of the area 2. Should be logically consistent · If a hypotheses suggests A=B and B=C, then A must also equal C 3. Should be stated concisely 4. Should be testable · Fine line between what is testable or not · If a hypotheses turns out to be untestable, correct it and move on
4 essential characteristics of a useful hypothesis
1. Aid in summarization 2. "Getting at what is going on" 3. Extracting "information" from the data 4. Aid in communication
4 purposes of statistic
1. State the hypotheses (null hypothesis & research/alternative hypothesis) 2. Set the criterion/critical value for a decision (p ≤ .05): How much error are we willing to accept to reject the null? 3. Compute sample statistics (e.g., t for t-test, F for ANOVA, r for correlation, etc.) 4. Compare the computed statistical value with the critical value to reach a conclusion
4 steps of hypothesis testing
1. Reader profiles 2. Item selection studies 3. Uses and gratifications studies 4. Journalist-reader comparisons
4 types of newspaper readership
Ratio (4 Levels of Measurement)
A scale has all the properties of an interval scale plus one more: true zero point (e.g., time and distance) · Ratio judgements can be made · Researchers who use this level or the interval level, can use parametric statistics
Likert Scale
A scale where a number of statements are developed with respect to a topic and respondents can strongly agree -- strongly disagree AKA summated rating approach; This is the most commonly used scale in mass media research
Guttman Scale
A scalogram analysis based on the idea that items can be arranged along a continuum in such a way that a person who agrees with an item or finds it acceptable, will find acceptable all other items expressing a less extreme position
Sample
A subset of the population of interest
This test is used when researchers may want to investigate several levels of an independent variable or possibly several independent variables, and possibly several different groups o t-test o degrees of freedom o contingent table analysis o ANOVA o chi-square test
ANOVA
__________ explains the source of variance in a set of scores on one or more independent variables o t-test o degrees of freedom o contingent table analysis o ANOVA o chi-square test
ANOVA
CAPI
An alternative version of personal interview is a self-administered interview where respondents answer on a personal computer (see computer assisted personal interviewing CAPI pg 215)
Correlation
An empirical relationship between two variables such that changes in one are associated with another
Degrees of freedom
An intentional and predetermined reduction in sample size to provide a conservation data adjustment to compensate for research error
Central Limit Theorem
As the size ( n ) of a simple random sample increases, the shape of the sampling distribution of ( x̄ ) tends toward being normally distributed. · If sample is drawn at random and is large enough - has an approximate normal distribution o Bell curve
___________ interviewing eliminates many problems with questionnaires such as skip patterns and rotation of questions
CAITI (computer-aided telephone interviewing)
In ___________ interviewing, the interview is conducted in person and either the interviewer or respondent enters the information over the computer
CAPI (computer-assisted personal interviewing)
Cost / (Circulation/1,000) =
CPM, $ per thousand
Criterion for __________: · Correlation · Time order -Cause precedes the effect in time -The "cause" must be before the effect
Causality
________ helps answer the question: "What is a typical score?" - what constitutes as a "typical score" depends on the level of measurement and how the data will be used · what constitutes as a "typical score" depends on the level of measurement and how the data will be used
Central tendency
Examples of __________ questions are: · Likert-type: strongly agree- strongly disagree · Semantic differential: rating between two opposites (bipolar scale) · Graphically oriented measures: pain scales :) - :(
Close-ended
__________ is constructed by adding the number of scores in one interval to the number of scores in the intervals above it
Cumulative frequency
Positive relationship (Scatterplots)
both variables increase together
Unstructured interview
broad questions allow interviewers freedom to determine what further questions to ask to obtain the required information
3 assumptions that must be made for ________ to be a valid description of the relationship 1. the data represent interval or ratio measurements 2. the relationship between X and Y is linear, not curvilinear 3. the distributions of the X and Y variables are symmetrical and comparable
Pearson's R
_________ is used to measure the degree of relationship between two variables
Pearson's R
· An estimate of the strength of the relationship, as indicated by the number · A statement about the direction of the relationship, as shown by the + or - sign Varies between +1.00 and -1.00: 1 is the level of relationship and + or - is the direction 1 indicates perfect correlation, either negative or positive
Pearson's R
__________ use trained interviewers and invite a respondent to a field service location or research office (sometimes conducted at the person's work or home)
Personal interviews
"The more time students spend on studying, the higher grades students will get on their exams," is an example of a __________ o Positive hypothesis o Negative hypothesis
Positive hypothesis
One-tail test
Predicts that the result will fail in only one direction - either positive or negative
______ sampling uses random selection and is intended to represent the population
Probability
________ research is more flexible and in-depth, and allows the researcher to view behavior in a natural setting
Qualitative
_________ research requires that the variables be measured
Quantitative
Close-ended questions
Questions that can be answered in short or single word responses · Respondents select an answer from a list provided by the researcher
Mean
Represents the average of a set of scores (X̄)
"Is students' anxiety about giving public speech associated with their grades in public speaking course?" is an example of a ___________ o Research Question o Hypotheses
Research Question
"Will students get higher grades in the course after they spend more time on studying?" is an example of a ___________ o Research Question o Hypotheses
Research Question
________ are used frequently in problem or policy-oriented studies where the researcher is not interested in testing the statistical significance of the findings
Research questions
_______ is a fundamental component of virtually all statistical formulas
Sample size
A census is the only way to survey every person in the population; ________ collects data among a portion of the population instead.
Sampling
__________ are used to eliminate unwanted respondents; that is, to include only respondents who answer questions in a specific manner. Often have respondents skip one or more questions
Screener questions/Filter questions
X = any score in a series of scores X̄ = the mean (read "X bar"; M is also commonly used to denote the mean) Σ = summation (Greek capital sigma letter) N = the total number of scores in a distribution
Symbols for calculating the mean
_________ surveys offer more control and higher response rates than most mail surveys, but they are limited in the types of questions that can be asked
Telephone
· Use trained interviewers who ask questions and record responses, usually on a computer · Respondents generally do not see the actual questionnaire · Fill a middle ground/ compromise between mail and personal interviews (similar to personal interviews) · Interviewers are extremely important
Telephone surveys
What is the alternative (null hypothesis) to the research hypothesis: The level of attention paid to radio commercials is positively related to the amount of recall of the commercial
The level of attention paid to radio commercials is not related to the amount of recall of the commercial
The ________ scale uses an interval measurement scale but can be time consuming and labor intensive
Thurstone
Spurious relationship
a coincidental statistical correlation between two variables, shown to be caused by another variable
Mailing list
a collection of names and addresses of respondents who belong to a narrowly defined group
T or F: The probability of committing a Type I error is equal to the established level of significance and is under direct control of the researcher
True
T or F: The terms one-tail testing and two-tail testing refer to the type of prediction made in a research study
True
T or F: The unit of analysis in an item-selection study is a specific news article
True
T or F: When conducting research there is a chance to get data that support the alternative hypothesis simply by luck or random accident, not because the alternative hypothesis is true
True
T or F: When testing for statistical significance, there could be data that does not support the alternative hypothesis even though the hypothesis is true
True
T or F: if a distribution has more than one mode, it becomes a non-effective way of analyzing data
True
T or F: All terms in used in a hypotheses must have an operational definition
True example: "There is a significant difference between the *recall* of TV commercials for subjects exposed to *low-frequency* broadcasts and that for subjects exposed to *high-frequency* broadcasts"
T or F: Survey research finds correlation, while experiment research finds causality
True because the IV cannot be manipulated in surveys but can be in experiments
T or F: A statistically significant result is not necessarily a meaningful result
True;
T or F: An ANOVA can be used to investigate several independent variables (factors)
True; A one-way ANOVA investigates one independent variable, a two-way ANOVA investigates two independent variables, and so on...
T or F: It is must better to form an untestable hypothesis than to form none at all
True; An untestable hypothesis may eventually become a respectable research question
T or F: There is no difference in value between a study that finds statistically significant results and a study that does not
True; Both provide valuable info
T or F; A high power level may help interpret the research results
True; If an experiment barely reaches the significance level but has high power, researcher can place more faith in the result
T or F: Researchers rarely state the null hypothesis
True; Since every hypothesis has its's logical alternative, stating the null is redundant
Scattergram/Scatterplot
a graphic technique for portraying the relationship between two or more variables
Journalist-reader comparisons
a group of journalists is questioned about a certain topic, and their answers are compared to those of their readers to see whether there is any correspondence between the two groups
Partial Correlation
a method researchers use when they believe that a confounding or spurious variable may affect the relationship between the independent and the dependent variables
Purpose sampling
a non-probability sampling technique where people are selected based on the purpose of the research · can select certain people for special reasons
Snowball sampling
a non-probability sampling technique where selected individuals contact other similar individuals · useful when the desired sample's characteristic is rare or difficult to identify (ex. victims of sexual violence)
quota sampling
a non-probability sampling technique where the # of subjects in specific categories are pre-planned, or individuals are selected to match demographic proportion · e.g., if the population is split 49% male and 51% female, you would pick a samples of the same %
Convenience sampling
a non-probability sampling technique where you select whichever individuals are easiest to reach · not generalizable
A correlation coefficient of -1.00 indicates: o absolutely no relationship o a perfect positive correlation o a perfect negative correlation
a perfect negative correlation
A correlation coefficient of +1.00 indicates o absolutely no relationship o a perfect positive correlation o a perfect negative correlation
a perfect positive correlation; both X & Y are completely covariant (they vary together)
Simple random sampling
a probability sampling technique in which every member of the population has an equal probability of being selected for the sample · e.g., draw names from a hat (must be truly random)
Systematic random
a probability sampling technique in which the first participant is randomly chosen, then others are chosen at intervals · calculate a skip interval (every Xth case) and move on to the next selected one
Stratified random sampling
a probability sampling technique in which the researcher identifies particular demographic categories of interest and then randomly selects individuals within each category.
multistage sampling
a probability sampling technique involving at least two stages: a random sample of clusters followed by a random sample of people within the selected clusters
Measurement
a process in which a researcher assigns numerals to objects, evens, or properties according to certain rules
Double-bind question
a question that contains a hidden premise and regardless of how the respondent answers, an affirmative response to the hidden premise is implied
Leading question
a question that suggests a certain response (either literally or by implication) or contains a hidden premise
Multiple choice question
a question with a set number of answer choices
Mail survey
a research method in which respondents answer a questionnaire sent through the mail
Non-probability sampling
a sampling technique in which there is no way to calculate the likelihood that a specific element of the population being studied will be chosen
Double-barreled question
a single survey question that asks two questions but allows only one answer
Research hypothesis
a statement expressing differences or relationships among phenomena
Pearson's R
a statistic that measures the direction and strength of the linear relation between two variables that have been measured on an interval or ratio scale · Varies between -1.00 and +1.00
Hypothesis testing
a statistical method that uses sample data to evaluate the validity of a hypothesis about a population parameter
t-Test
a statistical test used to evaluate the size and significance of the difference between two means
Uses and gratification study
a study of the motives for media usage and the rewards that are sought
Reader-nonreader study
a study that helps newspapers identify nonprint readers and attract them to their online editions · can be conducted via personal, telephone, online, or mail interviews with minor modifications
Error example: Car might have a steering problem, so you take the car to the mechanic... · Assume you know nothing wrong with your car's steering · Mechanic can tell you (1) there is a problem with steering or (2) there is no problem with steering · If the mechanic says there is no problem, it is the correct decision · If the mechanic misreads some test results and says there is a problem, this is an error · In this situation you would have falsely rejected a true null hypothesis ("There is nothing wrong with the steering") and committed a ____________
Type I error
To reduce the probability of a ________ the researcher can simply set the level of significance closer to zero o Type I error o Type II error
Type I error
Setting a significance level at .0001 will virtually eliminate Type ____ error but will dramatically increase the odds of Type ____ error
Type I, Type II
A low power level may indicate a _______
Type II error
Error example: Car might have a steering problem, so you take the car to the mechanic · There really is something wrong with the steering · Mechanic can tell you (1) there is a problem with steering or (2) there is no problem with steering · If the mechanic says there is a problem, it is the correct decision · If the mechanic overlooks some test results and says there is no problem, that is an error · In this situation you failed to reject the false null hypothesis and committed a __________
Type II error
This error is more difficult to conceptualize as the researcher does not have direct control; instead, it is controlled indirectly by the design of the research o Type I error o Type II error
Type II error
The level of Type ____ error is inversely proportional to the level of Type ____
Type II, Type I
Forced-choice question
a survey question format in which respondents give their opinion by picking the best of two or more options (used in media studies, gather info about lifestyles and they are always listed in pairs)
Frequency distribution
a table of scores ordered according to magnitude and frequency of occurrence
Hypothesis
a testable prediction based on theory or previous observation
Linear regressions
a way of using the association between two variables as a method of prediction
Assignment (Measurement concepts)
The destination of the numerals or numbers to certain objects or events
Error in sampling
The difference between the opinions and behaviors of the sample and the opinions & behaviors of the population
Which error has more consequences, type I or type II?
Type II; Suppose the researcher is testing the efficacy of a new drug - Which result is more harmful: making a Type I error and claiming the drug works when it does not, or making a Type II error and overlooking a drug that might actually work? If your car's steering is really broken but your mechanic fails to identify it (Type II error) you might have an accident, but if your steering is working properly and the mechanic mistakenly says it's broken (Type I error) you will pay for unnecessarily repairs, but not as bad as having an accident
· Histogram · Frequency polygon · Frequency curve
Types of graphs
1. Readership 2. Circulation 3. Management 4. Website utilization
Types of newspaper research
Semantic Differential Scale
Used to measure the meaning an item has for an individual AKA a bipolar rating scale
Ordinal (4 Levels of Measurement)
Usually ranked along dimension, such as from smallest to largest (e.g., socioeconomic status) · Posses equivalence (all categories are equal) · Numerals have some mathematical meaning (ranks) · Any given category can be higher or lower than another category
Interval (4 Levels of Measurement)
When a scale has all the properties of an ordinal scale and the intervals between adjacent points on the scale are of equal value (e.g., temperature) · No true zero point; a researcher cannot make statements of a proportion nature
Numeral (Measurement concepts)
When given quantitative meaning, becomes a number used in math & stat computations
When making two different measurements the same entity or person, researchers commonly designate one measurement as the ______ variable and the other as the ______ variable
X, Y
A correlation coefficient of 0.00 indicates: o absolutely no relationship o a perfect positive correlation o a perfect negative correlation
absolutely no relationship
The readership of a particular item in the print version is usually measured by means of _______, when the interviewer shows a copy of the paper to the respondent to find out which stories the respondent remembers
aided recall
Variables measured at the nominal level are always _______
always discrete
CPM (cost per thousand)
amount it will cost to reach 1,000 potential customers with a particular advertising type and time slot (article/commercial)
Multiple regression
an extension of linear regression, used to analyze the relationship between two or more independent variables and a single dependent (criterion) variable
Contingency table analysis (AKA crosstabulation or crosstabs)
an extension of the goodness-of-fit test except two or more variables can be tested simultaneously
Open-ended question
an interview question that encourages an answer phrased in the respondent's own words
In an ____________ survey, two or more variables are usually examined to investigate research questions or test research hypotheses. These results allow researchers to examine interrelationships among variables and develop explanatory inferences.
analytical
TV station owners surveying the market to determine how lifestyles affect viewing habits in order to determine whether viewers' lifestyles can be used to predict the success of syndicated programming is an example of an ____________ survey
analytical
null hypothesis
asserts that the statistical differences or relationships discovered in an analysis are due to chance or random error
Analytical survey
attempts to describe and explain why situations exist
Descriptive survey
attempts to describe or document current conditions or attitudes, i.e., to explain what exists at the moment
_______ surveys: · have a more representative sample · Incentives slash compensation (pay for minutes) · Attention problems (people might be on the move)
cell phone
Predictive (4 types of validity)
checking a measurement instrument against future outcomes
Before the internet, ________ referred to the number of people who received a newspaper or magazine via subscription or bought it at a vending machine
circulation
Today, _______ it is a less useful measurement, due to media available across multiple platforms (computer, tablet, smart phone, e-reader)
circulation
Advantages of _________: · Provide greater uniformity in responses · Answers are easy to quantify
close-ended questions
Disadvantages of _________: · Researchers often fail to include some important responses · To fix this they can include "other" with a fill in the blank which is treated like an open-ended question (with content analysis & codebook)
close-ended questions
Descriptive statistics
condense data sets to allow for easier interpretation
If scattergram results show a very low correlation, the researcher should investigate the presence of a ____________
confounding variable
_________ is the numerical expressions of the degree to which two variables change in relation to each other
correlation
The ______ variable is influenced by the ______ variable.
dependent, independent
Broadcast stations and networks continually surveying their audience to determine programming tastes, changing values, and lifestyle variations that might affect programming is an example of a ___________ survey
descriptive
Variables measured at the interval or ratio level are _______
either discrete or continuous
Face (4 types of validity)
examining the measurement device to verify it measures what it appears to measure
4 types of validity
face, predictive, concurrent, construct
T or f: A two-tail test is much stronger than one-tail
false; other way around
When calculating the mean, if the data are contained in a ____________, a slightly different formula is used: X̄ = fx N x = midpoint of any given interval f = frequency of that interval
frequency distribution
· Usually consist of two perpendicular lines: the x-axis or abscissa (horizontal) and the y-axis or ordinate (vertical) · One common standard of graphic format is to list the scores along the x-axis and the frequency (or relative frequency) along the y-axis - the height of the line or bar then indicates the frequency of a scare
frequency distribution graph
· Typical table includes 2 columns · Can also be constructed using grouped intervals, each of which contain several score levels · Can also include columns for proportions or percentages -obtain the percentage, divide the frequency of the responses by N (the total # of responses in the distribution) -percentages allow comparisons to be made between different frequency distributions that are based on different values of N · Some include frequency distributions others include the cumulative frequency
frequency distribution table
Variables measured at the interval level are _______
generally discrete
Reach x Average Frequency =
gross rating points (GRP)
More natural settings (e.g., field experiments) have _______ validity and _______ validity.
high external, low internal
Frequency polygon
if a line is drawn from the midpoint of each interval at its peak along the y-axis to each adjacent midpoint/peak (displays same information as a histogram)
Discrete variable
includes a finite set of values and cannot be divided into subparts example: dates, zip codes, marital status, etc.
Preparing a persuasive introduction (should be short, realistically worded, nonthreatening, serious, neutral, and pleasant but firm) is one way to __________
increase the response rate
Researchers dismiss the null by showing that variations are not due to error but because of the _________
independent variables
We have __________ when we are able to rule out possible but incorrect explanations of results
internal validity
internal consistency
involves examining the consistence of performance among the items that compose a scale
The following readership example is what what type of study? "the interviewer preselects items for which readership data are to be gathered and asks subjects about those items only - Calls are made on the same day the issue of the paper is published and respondents are asked to bring a copy of the paper to the phone and have them identify the items they have read" o journalist-reader comparison o reader-nonreader study o item-selection study o uses and gratifications study
item-selection study
Less natural settings (e.g., lab experiments) have _______ validity and _______ validity.
low external, high internal
Advantages of ________: · Survey over a wide geographic area for a reasonable cost · Often the only way to reach people who live in hard-to-reach areas or other countries · Allow for selective sampling using specialized mailing lists · Provided anonymity; respondents are more likely to answer sensitive questions · Low cost - do not require a large staff of trained interviewers
mail surveys
Disadvantages of _______: · Must be self-explanatory · Slowest form of data collection - many researchers set a cutoff date · Researchers never know exactly who answers the questions · Low return rates
mail surveys
Statistics
mathematical methods to collect, organize, summarize, and analyze data · provide reliable results only when the data collection and research methods follow established scientific procedures
For every distribution, 3 characteristic numbers can be identified: _______, ______, and ______
mean, median, and mode
Feeling thermometer question
measures feelings based on scale 0-100; can rate an idea or object
Concurrent (4 types of validity)
measuring instrument is checked against some present criterion
A calculation is not necessary to determine the ________; most often found by inspecting the distribution
mode
The ______ is easy to determine but because it only focuses on one score, it can hide important factors
mode
_________ is used to predict the dependent variable using information derived from an analysis of the independent variables o ANOVA o Least squares technique o Multiple regression
multiple regression
Once _________ are identified, researchers attempt to describe them by means of traditional demographic variables o reader profiles o non-readers o reader methodologies
non-readers
Chi-Square Goodness of Fit and Contingency Table Analysis are examples of o Parametric statistics o Nonparametric statistics
nonparametric statistics
_________ statistics appropriate with nominal & ordinal data, ________ statistics are appropriate for interval & ratio data
nonparametric, parametric
When calculating the median, if the distribution has an ______ of scores, the median is the middle scores. If the distribution has an ______, the median is a hypothetic score halfway between the two middle scores
odd number, even number
___________ are used when researchers have knowledge of the area and are able to predict the outcome o one-tail test o two-tail test
one-tail test
In _______ surveys: · Respondent is contacted via telephone, letter, or email and asked to participate · Survey either done through email or through link online · Often consists only of data collection · Small studies can be done for free
online
Advantages of ________: · Low cost · Representative · No geographic limitations or specific time constraints · Flexible and can expose respondents to almost any type of audio or visual material
online surveys
Disadvantages of ________: · Not yet a way to ensure that the person recruited is actually the person who completes the questionnaire · No control over data-gathering procedures - Suggestion to reduce the amount of error in online data collection: · "The phenomenon of omnipresent extreme outliers" calculate z-scores pg 218
online surveys
Mutually exclusive question
only one response per question, used in multiple choice
Advantages of _________: · Allow for more freedom answering and an opportunity for in-depth responses · Researchers can follow-up and probe for more details and info on the respondents' feelings and motives behind their answers · Allow for answers researchers did not foresee when designing the questionnaire—answers that may suggest relationships with other answers or variables · Particularly useful in a pilot study test as researchers may not know what types of responses to expect from the subjects.
open-ended questions
Disadvantages of __________: · Takes a lot of time to collect and analyze responses · A content analysis with coding must be done on each question. These results are used to produce a codebook: a menu list of responses · Difficult to interpret respondent answers. Verbatims: actual comments
open-ended questions
Statistical methods are divided into two categories: _________ statistics and _______ statistics
parametric and nonparametric
t-Test and ANOVA are examples of o Parametric statistics o Nonparametric statistics
parametric statistics
_________ statistics assume normality, while_________ statistics make no assumption about normally distributed data
parametric, nonparametric
___________ enables the researcher to determine the influence of the controlled variable o simple correlation o partial correlation
partial correlation
Advantages of __________: · Flexible means of obtaining face-to-face information · Depth & greater detail · Interviewer can develop a rapport with the respondents and be able to elicit replies to sensitive questions that remain unanswered by mail or telephone · Identity of respondent can be known or controlled, whereas in mail the family might confer on the answer · Once the interview has begun its harder for the respondents to terminate the interview before all questions are asked (telephone can hang-up)
personal interviews
Disadvantages of _________: · Time intensive and costly · Potential for interviewer bias (the appearance or behaviors of the interviewer may affect the respondent's answers) · Organization for recruiting, training, and administering a field staff of interviewers is much greater than other methods · If conducted during the day, most respondents will not be employed outside the home, otherwise these will need to be scheduled on a weekend or evening
personal interviews
When the probability level, sample size, and effects size are known the researcher can consult ______ to determine the level of power in their study
power tables
________ consist of sets of curves that represent different sample sizes, levels of significance (.05 etc.), and types of tests (one- or two-tail)
power tables
In _____________ research there is no need to explain the purpose or value of the survey to respondents or tell them how long it will take
private sector
In hypothesis testing, to determine the statistical significance of a research study, the researcher must set a __________, against which the null hypothesis is tested
probability level
Least squares technique
produces a line that is the best summary between two variables (May come close to all dots, even though it may not pass through any of them)
Central tendency
provide information about the grouping of numbers in a distribution by giving a single number that characterizes the entire distribution
Reader profiles
provides a demographic summary of the readers of a particular online or print publication
Statistical power analysis (aka power analysis)
provides an estimate of the minimum number of respondents needed to provide the best chance to discover if something do or does not exist
________ and ________ are often used to construct reader profiles because there may be significant differences in the nature & extent of print and online newspaper and magazine reading among individuals who have the same demographic characteristics
psychographic studies and lifestyle segmentation
It is common practice in mass media research to set the probability level at .01 or .05, which means that 1 or 5 times out of 100 significant results of the study occurs because of ___________
random error or chance
Sample size (3 parameters of statistical power)
refers to the number of subjects used in an experiment
Construct (4 types of validity)
relating a measuring instrument to some overall theoretic framework to ensure the measurement is logically related to other concepts in the framework
scale
represents a composite measure of a variable; is based on more than one item
· Some research is informal to solve relatively simple problems; some research is based on theory and requires formally worded questions · All researchers must start with a tentative generalization on the relationship between two variables · This generalization may take two forms: ____________ & ___________
research questions and statistical hypotheses
Lifestyle segmentation
respondents are asked a battery of questions concerning their activities, hobbies, interests, and attitudes · results are analyzed to see which cluster together, groups are identified and labeled, e.g., newspaper industry research that segmented readers based on how they manipulated time into three groups: routinized, relaxed, and harried
The _______ scale measures attitude on opposite spectrum (pleasurable/unpleasant, honest/dishonest)
semantic differential scale
Frequency curve
similar to a frequency polygon except the points are connected by a continuous unbroken curve of lines
_________ involves measuring the relationship between two variables o simple correlation o partial correlation
simple correlation
Data distributions
simply a collection of numbers
Comparison hypothesis
specifies how persons falling into different groups of a categorical variable will compare in terms of some second variable (categorical or continuous) e.g., female vs male consumers' difference on time sent online shopping
Relationship hypothesis
specifies nature of the association between two continuous IV and DV (both variables are quantitative)
Structured interview
standardized questions are asked in a predetermined order; little freedom is given to interviewers
A _________ result simply tells the researcher that an observed result is probably not the result of chance or error. It is up to the researcher to determine if the result has any social significance
statistically significant
In __________ research, the Independent variables cannot be manipulated the way they are in lab experiments. Without this control the researcher cannot be certain whether the relationships are casual or non-casual.
survey
Advantages of _________: · Used to investigate problems in realistic settings (examine things where they happen instead of in a lab in artificial conditions) · Cost is reasonable considering the amount of information gathered · A large amount of data can be collected with relative ease from a variety of people · Surveys are not constrained by geographic boundaries; can be conducted almost anywhere · Data that are helpful to survey research already exists. Data archives, gov't documents, census materials, radio & TV, rating books, etc. With archive data, it is possible to conduct an entire survey without ever developing a questionnaire or contacting a single respondent.
survey research
Disadvantages of __________: · Independent variables cannot be manipulated the way they are in lab experiments. Without this control the researcher cannot be certain whether the relationships are casual or non-casual. · Inappropriate wording or placement of questions within a questionnaire can bias results. · Wrong respondents may be included · Some survey research is becoming difficult to conduct because response rates continue to decline
survey research
General problems with __________: · Subjects or respondents are often unable to recall information · Prestigious bias · Respondents often give elaborate answers to simple questions · Often complicated by the inability of respondents to explain their true feelings, perceptions, and beliefs
survey research
________ is the most elementary method for comparing two groups' mean scores o t-test o degrees of freedom o contingent table analysis o ANOVA o chi-square test
t-test
Type II error (beta error)
the acceptance of a null hypothesis that should be rejected
Skewness
the concentration of scores around a particular point on the x-axis
Stability
the consistency of a result or of a measure at different points in time
Effects size (3 parameters of statistical power)
the degree to which the null hypothesis is rejected
__________ are generally used when little info is available about the research area o one-tail test o two-tail test
two-tail test
probability level (3 parameters of statistical power)
under the direct control of the researcher and the level predetermines the probability of committing a Type I error
Analysis of Variance (ANOVA)
used for designs with three or more sample means
Chi-square test
used to compare the observed frequencies of a phenomenon with the frequencies that might be expected or hypothesized
Item-selection study
used to determine who reads specific parts of the print or online version
The following readership example is what what type of study? "Respondents are given a list of possible uses and gratifications and asked whether any of these are the motives behind their reading e.g., How much do you agree or disagree with each statement: "I read the newspaper because it is entertaining/kill time/keep up to date/relax/find out what other people are saying about things." o journalist-reader comparison o reader-nonreader study o item-selection study o uses and gratifications study
uses and gratifications study
Psychographic studies
usually ask readers to indicate their level of agreement or disagreement with a large number of attitudinal statements · Results are analyzed to see how they correlate or cluster together · People who show high levels of agreement with questions that cluster together can be described with labels, e.g. answers might have people labeled as "progressives" or "traditionalists:
Histogram (bar chart)
vertical bars represent frequencies
Multivariate Analysis
when multi-dependent variables are measured in a single study
Inverse (or negative) relationship (Scatterplots)
when one variable increases while the other correspondingly decreases
exploritory research
when researchers have no clear idea what they will find and have no prior info to use for predictions
Uncorrelated (Scatterplots)
when there is no tendency for a high score on one variable to be associated with a high or low score on another variable
Standard deviation
· Standardized unit indicating how much off the sample mean is from the population mean · Is used in measuring sampling error
Continuous variable
· can take on any value, including fractions · can be meaningfully broken into subparts
external validity
· extent to which we can generalize findings to real-world settings · study conditions closely resemble natural settings, subjects usually provide a truer picture of their normal behavior
internal validity
· the extent to which we can draw cause-and-effect inferences from a study · when we are able to rule out possible but incorrect explanations of results
Probability sampling
A sampling technique in which every element in the population being studied has a known chance of being selected for the study
Checklist question
A fixed-alternative question that allows the respondent to provide multiple answers to a single question by checking off items (often used in pilot studies, can check what applies)
Research question
A question that can be answered by an experiment or series of experiments (can be tested)
Survey
A research technique that uses a standardized questionnaire to collect information
Region of rejection
In theoretical sampling distribution (a graphed display of sampling results), the proportion of the area in which the null hypothesis is rejected · Defined by the level of significance chosen by the researcher · The null hypothesis is retained in the region between the two rejection values (or levels)
Sampling units
Individual units of a population (what you are selection for inclusion in your research)
Rules (Measurement concepts)
Specify the way that numerals or numbers are assigned
1. Select a sample -Use a prepared sampling frame, typically from a mailing list 2. Construct the questionnaire - Must be concise and specific 3. Write a cover letter - A brief note explaining the purpose 4. Assemble the package. 5. Mail the surveys 6. Monitor return rates 7. Send follow-up mailings (The first should be sent 2 wks after, a second 2 more wks after. Can be sent to entire sample or only those who fail to answer) 8. Tabulate and analyze data
Stages of mail survey
1. Select a sample - The same as selecting a sample in any other method; based on predetermined set of screening requirements 2. Construct a questionnaire - Flexible, easy to ask detailed questions, private 3. Prepare an interviewer instruction guide - One-on-one interviewer guides are not very detailed; door-to-door interviewer guides contain more info (door-to-door are rarely used in US because of cost and the hesitancy of respondents to participate) 4. Train the Interviewers 5. Collect the data - Cost and labor intensive, can take several days - several weeks 6. Make necessary callbacks - Each requires interviewer to return to interview location 7. Verify the results - As with telephone interview, a subsample of each interviewer's data is verified 8. Tabulate the data
Stages of personal interviews
Sampling
Studying a portion of the population to make judgments about the entire population
__________ is a quantitative measuring tool that is used to make generalizations about a population.
Survey Research
Cumulative frequency
The sums of the frequencies of the data values from smallest to largest
T of F: When calculating the median, if no score neatly bisects the distribution, the two middle scores must be added together and divided by 2
True
T or F a structure interview is easier to tabulate and analyze, but an unstructured interview elicits more detail but takes a great deal of time to score & analyze
True
T or F: A pilot study or pretest of a questionnaire usually solves most problems with close-ended questions
True
T or F: Any statistical analysis tests the null hypothesis and seeks to reject the null hypothesis
True
T or F: As Type I decreases, Type II increases and vice versa
True
T or F: If two variables are perfectly related (r= 1.00), knowledge of a person's score on one variable allows the researcher to determine the score on the other
True
T or F: In Pearson's R, the strength of a relationship depends solely on the number, and it must be interpreted in terms of absolute value
True
T or F: Much of current circulation research that focuses on print examines how to maintain a circulation base that is attractive to advertisers and is profitable for the publication
True
T or F: Nothing in science can be proven, only disproven
True
T or F: Outliers pull the mean away from their direction
True
T or F: RQ's and hypotheses are identical but hypotheses predict an outcome, while RQ do not
True
T or F: RQs are used frequently in problem or policy-oriented studies where the researcher is not interested in testing the statistical significance of the findings
True
T or F: RQs may be a simple question about the relationship between two or more variables or about the components of a phenomenon
True
T or F: Reliability is necessary to establish validity, but a reliable measure is not necessarily a valid one
True
T or F: The best way to discover whether a research instrument is adequately designed is to pretest it
True
T or F: The most common form of ANOVA tests between two or more group means
True
variable
the empirical counterpart of a construct or concept
Readership
the group of people who regularly read a particular newspaper, magazine, etc.
Median
the midpoint of distribution (half the scores lie above, and half lie below it.)
Causality
the notion that a change in one factor results in a corresponding change in another
Response rate
the percentage of respondents who complete the questionnaire among those who are contacted/selected
Arbitron
the professional research organization that measures radio audiences
Nonspurious
the quality of a relationship between two variables that cannot be attributed to a third variable
Type I error (alpha error)
the rejection of a null hypothesis that should be accepted
Curvilinear relationship (Scatterplots)
the relationship can be positive (or negative) up to a point and then becomes inverse (or positive)
Mode
the score or scores that occur most frequently
Normal curve
the symmetrical bell-shaped curve that describes the distribution of many physical and psychological attributes · Most scores fall near the average, and fewer and fewer scores lie near the extremes.
Reach
the total audience that reads a newspaper or magazine in print or digitally
Triangulation
the use of multiple methods to study one research question · can use both qualitative and quantitative methods
dependent variable
the value of this variable is not manipulated
criterion variable
the variable in a multiple-regression analysis that the researchers are most interested in understanding or predicting
independent variable
the variable that is systematically varied by the researcher
Pretest
to conduct a mini study with a small sample to determine whether the study approach is correct and to help refine the questions
