Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
71 Cards in this Set
- Front
- Back
|
Characteristics of Experiments
|
Most rigorous scientific method that allows to test theories and the relationships the theories specify
Testing hypothesized relationships between variables The only method that allows to establish causality High degree of control over conditions By nature, artificial environment (lab conditions) |
|
|
Elements in Experiment
|
Three Pairs
Independent and dependent variables Pretesting and post-testing Experimental and control groups These elements allow to establish causation |
|
|
(Experiment)
Causation |
Time order: the independent variable precedes the dependent one in time
Direct association: change in independent variable results in change in dependent variable Non-spuriousness: there is no other intervening variable explaining away or changing the relationship Experiment is the only method that allows to create all three conditions through control; longitudinal (panel) studies can test for the time effect and associations, but not for non-spuriousness. |
|
|
(Experiment)
Ways to test relationships |
Laboratory experiment – “clean,” but artificial, fully controlled
Simulation – closer to reality, but still tightly controlled by changing various conditions Field (natural) experiment – done in real conditions, allows for other conditions to be taken into account |
|
|
(Experiment)
Establishing Control |
Goal: control all conditions carefully to ensure that the resulting change could only be attributed to independent variable and nothing else
|
|
|
(Experiment)
Control is established through: |
Equal start-up conditions (established through pre-test and random assignment of participants)
Manipulation of conditions (exposure to relevant message, to different message, and no message) Comparisons of all conditions |
|
|
(Experiment)
Validity Problems in Experiments |
Internal invalidity
External invalidity |
|
|
(Experiment)
Internal invalidity |
the possibility that the conclusions drawn from experimental results may not accurately reflect what went on in the experiment itself
|
|
|
(Experiment)
External invalidity |
the possibility that conclusions drawn from the experimental results may not be generalizable to the “real” world
|
|
|
(Experiment)
Sources of Internal Invalidity |
Historical events may occur during the course of the experiment that effect the results.
Maturation of the subjects. Testing and retesting can influence behavior (subjects can figure out the purpose and give expected answers). Instrumentation (if different tests are used, their sensitivity and results may differ) Statistical regression of subjects starting out in extreme positions in the group as a whole (outlier problem). Selection biases Experimental mortality Demoralized control group subjects Interaction (multiplied effect) of the above sources of invalidity. |
|
|
(Experiments)
Selection biases |
comparisons are meaningless if the groups are not comparable
|
|
|
(Experiments)
Experimental mortality |
subjects drop out of the study before it's completed.
|
|
|
(Experiments)
Pre-experimental Designs |
One-shot case study
One-group pretest-posttest design Static-group comparison |
|
|
(Experiments)
One-shot case study |
single group of subjects is measured on a variable following experimental stimulus.
X____O |
|
|
(Experiments)
One-group pretest-posttest design |
adds a pre-test for the group, but lacks a control group.
O_____X_____O |
|
|
(Experiments)
Static-group comparison |
includes experimental and control group, but no pre-test.
X______O O |
|
|
(Experiments)
Selecting the Subjects |
Most experiments involve undergraduate students – problem with representation, but generalizations are still possible if the following conditions are met:
Probability sampling Random assignment to experimental conditions (groups) Matching – pairs of subjects are matched on the basis of their similarities on one or more variables, and one member of the pair is assigned to the experimental group and the other – to the control group |
|
|
(Experiments)
True Experimental Designs |
Include random selection and random assignment of subjects to groups, allow to decrease external validity
Classical experiment Solomon 4-group design Post-test only control group |
|
|
(Experiments)
Solomon 4-group design |
Four groups of subjects, assigned randomly:
Groups 1 and 2 are the control and experimental group. Group 3 does not have the pre-test. Group 4 is only posttested. O_____X______O O____________O X______O O |
|
|
(Experiments)
Post-test only control group design |
Includes Groups 3 and 4 of the Solomon design.
X_______O O With proper randomization, only these two groups are needed to control problems of internal invalidity and the interaction between testing and stimulus. |
|
|
(Experiments)
Experimental Method Strengths |
Isolation of the experimental variable over time (causality).
Experiments can be replicated several times using different groups of subjects. Logical rigor that is much more difficult to achieve in other social science methods. |
|
|
(Experiments)
Weaknesses (generalizability): |
Artificiality of laboratory setting.
Social processes that occur in a lab might not occur in a more natural social setting. |
|
|
Descriptive Statistics
|
simply describe how the data are distributed within the sample
Frequencies Cross-tabulations Correlations |
|
|
(Descriptive Statistics)
Frequencies |
How often a certain data point (observation) occurs in the dataset
|
|
|
(Descriptive Statistics)
Cross-tabulations |
how observations break down by two or more variables
|
|
|
(Descriptive Statistics)
Correlations |
to what extent variables are associated or overlap with each other
|
|
|
(Descriptive Statistics)
Inferential statistics |
provide evidence that what is observed in the sample is not due to chance but has statistical significance
|
|
|
(Descriptive Statistics)
Univariate analyses |
Frequencies for categorical (non-parametric) data
Frequencies for continuous (parametric) data - measures of central tendency: Show how the scores are distributed around the mean and report the values for Mean – average score of all observations Median – midpoint, or the 50th percentile of the data set Mode – the most frequently occurring value Variance – amount of dispersion around the mean (sum of squared deviations from the mean divided by the number of observations) Standard deviation – ranges of expected scores’ deviations from the mean based on a normal curve (square root of the variance) |
|
|
(Descriptive Statistics)
Bivariate analyses |
Categorical data – cross tabulations
Continuous data – correlations |
|
|
(Descriptive Statistics)
Categorical data |
cross tabulations
|
|
|
(Descriptive Statistics)
Continuous data |
correlations
|
|
|
(Descriptive Statistics)
Variance |
amount of dispersion around the mean (sum of squared deviations from the mean divided by the number of observations)
|
|
|
(Descriptive Statistics)
Standard deviation |
ranges of expected scores’ deviations from the mean based on a normal curve (square root of the variance)
|
|
|
Inferential Statistics
|
Explore relationships between variables
Allow to talk about the results with a certain level of confidence Allow to infer the results from the sample to larger populations if the sample was drawn randomly Allow to conclude that the differences observed are not due to the chance or error but really are significant |
|
|
(Inferential Statistics)
Statistical reasoning |
The difference (change) in the dependent variable can be attributed to an independent variable
Research hypotheses are formulated based on a theory to test these relationships Statistical tests are based on the assumption that there is no difference/relationship (null hypothesis), and the results either support or reject the null hypothesis, but never prove the research hypothesis Two types of research hypotheses: Non-directional Directional Decide the level of acceptable error (5% or less) Establish the level of significance (95% or higher) Draw a sample large enough to decrease errors |
|
|
(Inferential Statistics)
Non-directional hypothesis |
specify only differences in the relationship
|
|
|
(Inferential Statistics)
Directional hypothesis |
specify the direction – positive or negative – of the relationship
|
|
|
(Inferential Statistics)
Type I error |
false rejection of a true population relationship
|
|
|
(Inferential Statistics)
Type II error |
false acceptance of a relationship that does not exist in the true population
|
|
|
(Inferential Statistics)
Chi-Square |
Nonparametric Test
Usually performed on one or two variables Is based on comparisons of actual values to expected values For one variable, the SPSS command is Nonparametric tests – Chi-Square With two variables, has the same logic as cross-tabulations – Descriptive Statistics – Cross-Tabs – Statistics – select Chi-Square Determines the magnitude (not the strength or direction) of the relationship through the option “Contingency Coefficient,” which should be more than .30 and significant |
|
|
(Inferential Statistics)
t-test |
Parametric Test
Parametric (continuous) analysis – small-sample statistic Looks at the sample mean as an approximation of the true population mean, compares means of two groups for significant differences The independent variable must be dichotomous The dependent variable must be continuous The maximum number of observations is 100 Most commonly used t-test – Independent Samples, but also possible to do one-sample T-test (comparison to an earlier reference study) and paired-sample t-test (comparison of two samples collected at different times) |
|
|
(Inferential Statistics)
Analysis of Variance (ANOVA) |
Parametric test
Logical continuation of t-test, but Allows to work with large samples (over 100 observations), more than two groups in the independent variable, and more than one independent variable (multivariate ANOVA) Similarly to t-test, establishes whether the means in the independent variable’s groups are statistically different, shows the strength of relationships through F-statistic Is incapable of showing the direction of the relationship |
|
|
(Inferential Statistics)
Parametric tests |
t-test
Analysis of Variance (ANOVA) Correlation analysis |
|
|
(Inferential Statistics)
Correlation analysis |
Parametric test
Shows the strength and direction of the relationship Calculates the coefficient showing how much of the variance in the dependent variable is explained by the independent variable – Pearson coefficient (.70 and higher indicates a strong relationship) Is incapable of showing causality – which variable is preceding which in time, or which one is dependent and which - independent |
|
|
(Inferential Statistics)
Regression Analysis |
Shows the strength, direction, and causality (caution: survey data are more limited than experimental data because of the time element) of the relationship
Produces ANOVA results and works with multiple independent variables (multivariate regressions), Can be performed for categorical dependent variables: - dichotomous variables (binary logistic regression), - multi-category nominal variable (multinomial logit), - ordered variable (ordinal logistic regression) Linear regression is based on the following equation: Y = α + β1X1 + β2X2 + … βnXn + e, where Y is the dependent variable, X1-Xn - independent variables, α – intercept (constant), β – regression coefficient, and e – error term (unexplained variance) Interpretation: a one-unit change in X (independent variable) results into a β-unit change in Y (dependent variable), or as the independent variable changes by one unit, the dependent variable decreases (negative relationship) or increases (positive relationship) by β units. |
|
|
(Inferential Statistics)
Summary: testing relationships |
see powerpoint
|
|
|
Case Study Definition
|
As an informal method, it is an in-depth study of particular people, organizations, events, or processes with the purpose of exploring, understanding, and explaining them
|
|
|
Case Study
In public relations, is widely used in |
crisis communications
organizational studies documentation of success/failure documentation of the decision-making process reports for the stakeholders, etc |
|
|
Case Study
Types of Case Studies |
record keeping
teaching case studies research case studies |
|
|
Case Study
record keeping |
medical, psychological, social work - each patient is a case, documentation of treatment/ action
|
|
|
Case Study
teaching case studies |
do not contain a complete or accurate rendition of events, serve as frameworks for discussion
|
|
|
Case Study
research case studies |
investigate a phenomenon within its real-life context
|
|
|
Case Study
Advantages of method |
Focuses on something that already happened, provides detail only found in hindsight through examination of context
Ability to compare end results with initial objectives, evaluate impact Ability to accumulate rich data through direct observation, participatory research, and secondary research Valuable as sources of good and bad examples of public relations – lessons that can be learned |
|
|
Approaches to Case Studies
|
Linear
Process Grounded |
|
|
Case Study
Linear |
case is a unique entity unto itself, demonstrates good and bad practices through a historical perspective of time, space, and context
|
|
|
Case Study
Process |
case is a snapshot of the larger public relations process, allows insight into similar situations and solutions
|
|
|
Case Study
Grounded |
takes a formal structure (MBO or programmed approach), but also indicates process
|
|
|
Observation
|
An exploratory method, used to understand new and insufficiently studied phenomena
Borrowed from anthropology and sociology which have developed the ethnographic methods of studying people’s behavior Requires good observation skills and meticulous recording of everything observed Challenge – to analyze the data and uncover patterns in findings |
|
|
Observation
Researcher’s roles in observation |
Complete observer
Complete participant |
|
|
Observation
Complete observer |
the researcher only observes the group without any participation
|
|
|
Observation
Complete participant |
the researcher participates in all activities just as any other member of the group
|
|
|
Case Study
Grounded |
takes a formal structure (MBO or programmed approach), but also indicates process
|
|
|
Case Study
Grounded |
takes a formal structure (MBO or programmed approach), but also indicates process
|
|
|
Observation
|
An exploratory method, used to understand new and insufficiently studied phenomena
Borrowed from anthropology and sociology which have developed the ethnographic methods of studying people’s behavior Requires good observation skills and meticulous recording of everything observed Challenge – to analyze the data and uncover patterns in findings |
|
|
Observation
|
An exploratory method, used to understand new and insufficiently studied phenomena
Borrowed from anthropology and sociology which have developed the ethnographic methods of studying people’s behavior Requires good observation skills and meticulous recording of everything observed Challenge – to analyze the data and uncover patterns in findings |
|
|
Observation
Researcher’s roles in observation |
Complete observer
Complete participant |
|
|
Observation
Researcher’s roles in observation |
Complete observer
Complete participant |
|
|
Observation
Complete observer |
the researcher only observes the group without any participation
|
|
|
Observation
Complete observer |
the researcher only observes the group without any participation
|
|
|
Observation
Complete participant |
the researcher participates in all activities just as any other member of the group
|
|
|
Observation
Complete participant |
the researcher participates in all activities just as any other member of the group
|
