Flashcards
Research and Program Evaluation
Cards in this deck (51)
- 1Parsimonythe researcher interprets the results of a study in the simplest and least complex manner--sometimes called Occam's Razor
- 2Parametera value used to represent a characteristic of the population
- 3Statisticis a value calculated form a data sample
- 4Probabilitythe quantitative descripton of the likelihood of a particular event occurring
- 5Correlationthe relationship between variables
- 6Positive Correlationthe relationship between variables
- 7Negative Correlationthe variables change in opposite directions
- 8Zero Correlationthe variables have no relationship
- 9SPSSStatistical Package for the Social Sciences
- a software package for use with statistical analysis - 10Solomon Four-Group Designis a study of whether or not a pretest affects the subjects of a study by sensitizing or influencing them before the start of the research. Such effects must be considered in the comparison of the results with the control group
- 11Multiple Regressionis a procedure in which the researcher
uses a correlation coefficient to learn about the relationship between multiple independent variables and a dependent variable. The procedure can be used to determine the "best predicator" of a particular event or outcome. An example would be which personality trait best predicts social adjustment. - 12Factor Analysisexamines the relationships among a group of variables for the purpose of determining the simplest explanation for those relationships, which is usually the smallest number of factors involved.
- 13The Biserial Correlation Coefficientis a measure of the relationship between one variable with multiple values and another that is dichotomous.
- 14Cross-Sectionala study of the characteristics of multiple groups. The study of one group's characteristics over a period of time is a longitudinal study
- 15Degrees of Freedomhow many observations the researchers may make after he or she has made the minimum number needed for the study.
- 16Double Blind Techniqueneither the subject(s) nor the researchers know at least one variable. For example, in a drug test neither would know which subjects had received the actual drug or which had received the placebo.
- 17Homoscedasticitystatistical variances are assumed to be equal
- 18HeteroscedasticityUnequal variance of the data
- 19Semantic Differentiala method of measuring a subject's reactions to words or concepts through the use of bipolar scale using contrasting adjectives
- 20Analysis of Covariance ANCOVAis used in studies where the dependent variables are controlled. Possible techniques of controlling the variables are to use non-random samples or to statistically adjust variables that affect the dependent variable
- 21One-Way analysis of Varianceis a test for differences when the study involves three or more independent groups or levels
- 22Factorial Analysis of Variance ANOVAis used in the study of two or more variables. The 2X2 design, in which there are two independent variables each with two distinct values, is the most common, although multi-level designs can be used.
- 23Multivariate Analysis of Variance MANOVAis used in studies involving several dependent variables and at least two independent variables
- 24t-testthe t-test compares the mean of two independent data sets to determine if there is a significant statistical difference between them. The test makes reference to an established Table of t values. I can establish the existence or non-existence of relationships between data sets before a full standard deviation value is determined. This test can be especially useful for small sample groups.
- 25Independent Variablesis the factor in an experiment or study that is changed by the researcher.
- 26Dependent Variablesis what is being measured by the study and is changed in response to the changes in the independent variable
- 27Type I Error(s)or Alpha Error, the null hypothesis is rejected as false or unproven when it is actually true.
Rejected but True - 28Type II Error(s)or Beta Error, is when the researcher does not reject the null hypothesis when is is actually false.
Accept but False
Type II errors are often the result of a sample size that is too small - 29Chi Squareis used to determine if there is a significant difference in the distribution of two data sets. This test is used to determine whether the data fits a known type of distribution or whether different attributes or factors in a single data set are related or independent of each other. Another use for the test is to determine if two data sets or populations are homogeneous when compared to each other.
- 30Bivariate tabular analysisor cross break, is a method or graphically illustrating the relationship or non-relationship of two variables by the use an X/Y graph. Traditionally the independent variable is shown by the vertical axis and the dependent variance by the horizontal axis.
- 31Null Hypothesisassumes no difference or no association between variables
- 32Directional Hypothesispredicts how the independent variable will affect the dependent variable.
- 33Non-Directional Hypothesispredicts an effect, but does not state how the dependent variable will be affected
- 34Significance Levelindicated the probability of making a type I error in a hypothesis test. It is usually set as low as possible with a level of .05 or 5% being a commonly used level
- 35Meta-Analysisanswering a research question through the comparison of results for multiple studies
- 36Sample Sizethe number of samples included in a study
- 37Table of random numbersa list of random numbers (usually computer-generated) that can be assigned to potential study sample and used to randomly select those who will participate in the study.
- 38Likert Scalea rating scale on which study participants agree or disagree with statements that measure attitudes or opinions
- 39Scatterplota graphic using horizontal lines to illustrate the relationship between two variables
- 40Quantitive Researchis systematic and usually uses scientific methods.
Non-Experiemental types are: descriptive, comparative, correlational, and Ex Post Facto - 41Qualitative Researchis an in-depth investigation of the subject, which often is a group of people, but may be an individual. The group may be a family, a community, or even an ethnic or cultural minority. The motivation for the research usually involves learning the reasons why the subject practices certain behaviors.
- 42Research Methods Include:observation, case studies, and participant observation
- 43Random Samplesis choosing the subjects for a study entirely by chance. Such sampling reduces the likelihood of bias
- 44Stratified Samplingdivides the population into subgroups according to some criteria--such as age, ethnicity, gender, or SES and then selects subjects from each subgroup. This increases the validity of generalizing the findings to the entire population.
- 45Proportional Stratifiedsampling is the selection of the number of subjects from each subgroup that corresponds to the percentage of the population that fits into the subgroup
- 46Clustersampling divides the population into subdivisions or clusters, and then selects a random sample from the clusters
- 47Purposeful Samplingis the selection of subjects for in-depth study. Generalizations are not derived from these studies. Studies based on convenience or volunteer samples do not produce normal score distributions but may provide useful information
- 48Nominal Measurementare categorical variables such as gender, race, marital status, and religion. In a study nominal measurements are used to describe the demographics of the subjects.
- 49Ordinal Measurementsare used to describe variables that can be arranged in some type of order or ranking. Opinion and Attitude scales are examples.
- 50Interval Measurementsis used to describe variables with similar or equal distances between the ranks. Generations and crime rates are examples
- 51Ratio level of measurementapplies to variables that have equal intervals and zero reference point. This type of measurement is little used in the social sciences since attitudes are usually not measured at the zero level