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;
85 Cards in this Set
- Front
- Back
What kind of statistics does chapter 14 focus on? |
inferential statistics |
|
define inferential statistics |
statistics used in the process of making inference from a sample to a population. |
|
differences between techniques being tested may be due to what? |
real difference or sampling error |
|
define sampling error |
difference among group means because the samples are not 100 percent representative of a population; the extent to which sample values (statistics) deviate from those that would be obtained from the entire population (parameter). |
|
is it reasonable to conclude the differences are real and the treatments are not equally effective if the means (rope-skipping example) are large? |
yes, but if the differences among means are small, it could be sampling error. |
|
what are the steps in the hypothesis-testing procedure? |
1. state hypothesis 2. select the probability level 3. consult the statistical table 4. conduct the statistical test 5. accept or reject the null hypothesis |
|
define statistical hypothesis? |
hypothesis tested with a statistical test. (null hypothesis). Ho |
|
statistical or null hypothesis states what? |
there is no difference (in treatments, measurements, etc.) Statistical hypothesis does not have to agree with the research hypothesis. |
|
define research hypothesis |
a tentative explanation or prediction of the eventual outcome of a research problem; normally this is the outcome expected by the investigator. |
|
define alpha level |
probability level selected that warrants rejection of the null hypothesis. |
|
the difference between the population statement in the null hypothesis and what is found in the samples is due to what? |
null hypothesis being rejected or sampling error. |
|
what does the statistical test during step 4 of the hypothesis test procedure determine? |
provides the probability of the sample finding occurring if the null hypothesis is true. |
|
if probability is small.... |
...the null hypothesis is rejected and the difference is considered to be real. Otherwise, the null hypothesis is accepted and the difference is considered to be due to sampling. |
|
what is the purpose of step 2 of the hypothesis testing procedure? |
to select the probability level that warrants rejection of the null hypothesis. This probability level is called the alpha level and is usually .05 or .01 |
|
in step three what happens if the p value is less than or equal to the alpha level? |
the null hypothesis is rejected. |
|
how is the p value determined? |
consult statistical table to fine the value that that the statistical test of the sample would need to equal or exceed in order to reject the null hypothesis at the chosen alpha level. |
|
what is purpose of step 4 in the hypothesis-testing procedure? |
statistical test is calculated either by hand or computer. |
|
what is the purpose of step 5 in the hypothesis-testing procedure? |
this is the final step, where the decision is made either to accept or to reject the null hypothesis. |
|
in step 5 what happens if the value of the statistical test is greater than the table value or if the p value for the value of the statistical test is less than the alpha level? |
null hypothesis is rejected |
|
does accepting the null hypothesis prove it is true? |
no...only that it is plausible. |
|
what does it mean when literature says a statistical test was significant? |
indicates the value of the statistical test warranted rejection of the null hypothesis and there is a real difference between sample not due to sampling error. |
|
what does it mean when literature says a statistical test nonsignificant? |
indicates that the null hypothesis was retained, and any differences were attributed to sampling error. |
|
t test and analysis of variance (ANOVA) generally applied in what types of research? |
experimental |
|
chi-square test test generally applied in what types of research? |
descriptive |
|
define one-group t test |
a statistical test used with one sample. |
|
define degrees of freedom |
a value associated with a statistical test which is used when finding a table value of the statistical test. DOF = (n-1) |
|
define critical value |
the tabled value of the statistical test needed to reject the null hypothesis. |
|
define critical region |
the region at which all values of the statistical test are at or beyond the critical value. |
|
define two independent groups t test. |
a statistical test used with two independent samples. Used when either two samples are drawn from the same population and each administered a different treatment or when the sample is drawn from each of two populations.
|
|
two dependent groups t test |
a statistical test used with two dependent groups or columns of scores. |
|
define repeated measures |
a two dependent groups t-test design where participants are tested before and after a treatment. |
|
what is the purpose of a repeated measures design? |
to determine if an experimental treatment is effective. (test, treatment, re-test). |
|
what is an alternative to the two independent variable design? |
matched pair design. (individuals closely matched in the distributed to two different groups). |
|
what is a problem with match paired method? |
loss of participants. When one member of a pair is lost from the experiment, the other member was be removed from the study. |
|
why are two alternate hypotheses normally chosen? |
because the null hypothesis and the two alternated hypothesis cover all possibilities. |
|
the number of alternate hypothesis determines what? |
whether the statistical test is one-tailed or two-tailed. The smaller the table value, the easier it is to reject the null hypothesis. |
|
what does a smaller alpha level mean (.01 compared to .05) |
makes it more difficult to reject the null hypothesis. The smaller the alpha level, the larger the difference must be between what is hypothesized at the population level and what was found at the sample level. |
|
define type 1 error |
rejection of a true null hypothesis |
|
define type 2 error |
acceptance or non-rejection of a false null hypothesis. probability of making a type 2 error is symbolized by beta. |
|
what effects the probability of type 2 error. |
1. larger alpha level = smaller probability 2. larger sample size = smaller probability 3. more false the null hypothesis = less probability |
|
define statistical power |
probability of not making a type 2 error (1-B) |
|
what factors give more statistical power? |
1. one-tailed test over two-tailed test 2. 2 dep group over 2 indep group design
everything that decreases probability of type 2 error increases statistical power. |
|
define analysis of variance (ANOVA) |
a statistical test with two or more groups |
|
when is an ANOVA used? |
when a research problem or question involves more than two treatment groups. ANOVA is used to analyze the data. Can also be used when there are just two treatment groups in place of a t test.
total variability in a set of scores is divided into two or more components or sums of squares (SS) |
|
define sums of squares |
variability values in ANOVA |
|
define square (MS) value |
values used in ANOVA to calculate an F statistic. |
|
what is an F statistic |
the ratio of two mean square values, which is used to test the null hypothesis. F test is conducted at step 4 of five-step hypothesis testing procedure. |
|
define one-way ANOVA |
a statistical analysis used with two or more independent groups. An extension of the two independent groups design. |
|
define repeated measures ANOVA |
a statistical analysis where each participant is measured on two or more occasions. If only two measures for each participant, this is an alternate test to t test for dependent groups design. |
|
what is the best way to conduct a repeated measures ANOVA with SPSS |
with the reliability program |
|
define random blocks ANOVA |
a statistical analysis where participants similar in terms of a variable are placed together in a block and then randomly assigned to treatment groups. extension of matched-pair t test when there are three or more groups. |
|
what is Maucly's Test of Sphericity? |
the results of the Mauchly test determines whether a univariate test or multivariate test should be used for the repeated-measures ANOVA. If the approximate chi-square value is not significant, then the univariate test should be used. If chi-square value is significant multivariate test should be used. |
|
a "pair" in matched pair t test is to _________ in a random blocks ANOVA |
block. The matched pairs in a random blocks ANOVA are place together in a block. |
|
true or false: the analysis for the random blocks design is the same as the analysis for the repeated measures ANOVA design. |
true |
|
define two-way ANOVA |
an ANOVA design with rows and columns. Also called two-dimensional ANOVA
have multiple scores within one cell compared to only one score in a cell with one-way ANOVA |
|
examples of two-way ANOVA |
1. random blocks design 2. factorial design 3. factorial design with rows as levels of one treatment, columns are levels of second treatment and there are multiple participants in each cell. |
|
define factorial design |
a two-way ANOVA design with rows representing a classification or treatment and columns representing a treatment. |
|
simple effect test |
test to compare the column means for each row or the row means for each column in a two-way ANOVA design. |
|
two-way ANOVA is also called what? |
factorial analysis |
|
what are the assumptions underlying statistical tests? |
1. scores are at least interval. 2. random sampling of participants 3. for t test for two independent groups, one-way ANOVA and factorial ANOVA it is assumed that groups are independent of each other 4. assumed that scores are normally distributed at the population level (30 participants is normally the minimum) 5. in studies using multiple samples, populations represented are assumed to be equally variable |
|
what is effect size (ES) |
it is a way to decide whether the statistically significant difference is a practically significant difference. Cohen ES less than .20 is small, around .50 is medium, and greater than .80 is large |
|
how do you calculate ES? |
mean score A - mean score B/SD for one group or pooled SD for combined groups. |
|
define two-group comparisons |
statistical techniques to compare groups two at a time following a significant F in ANOVA. Also called multiple comparisons and posteriori comparisons. Used to determine if two groups differ significantly. |
|
list a few two-group comparison techniques |
1. Scheffe 2. Tukey 3. Duncan 4. Bonferroni |
|
define per-comparison error rate |
probability of making a type 1 error in a single two-group comparison |
|
define experiment-wise error rate |
probability of making a type 1 error somewhere in all the two-group comparisons conducted. |
|
what is the ANCOVA? |
analysis of covariance
alternative to most ANOVA designs. statistically adjusts the difference among the group means on the criterion variable to allow for the fac that groups differ in mean score on some other variables and then applies ANOVA to the adjusted criterion.
reduces error variance (reflected in the denominator of the F test). |
|
define parametric statistic |
statistic that requires interval data and a normal distribution. |
|
define nonparamtric statistic |
statistic that has no requirement of interval data and normal distribution |
|
define parametric test |
statistical tests that assume interval data and normal distribution scores. |
|
define nonparamtric tests |
statsitical tests that do not assume interval data and normal distribution scores. |
|
define one-way (goodness of fit) chi-square tests. |
a nonparametric statistical test using frequencies arranged in columns. defines whether or not expected and observe frequencies are in agreement. |
|
define expected frequency |
the frequency hypothesized or expected for an answer in a chi-square test. |
|
observed frequency |
the frequency in the sample for an answer in a chi-square test |
|
true or false: the less agreement there is between expected and observed outcome the bigger the chi-square value. |
true (if there was perfect agreement, the value would be 0) |
|
define two-way (contingency table) chi-square tests. |
nonparametric statistical test using frequencies arranged in rows and columns. |
|
define correlation |
a mathematical technique for determining the relationship between two sets of scores. When two variables are correlated, it becomes possible to make predictions. |
|
define regression |
that statistical model used to predict performance on on variable from another. |
|
define prediction |
estimation of a person's score on one measure based on the individuals' score on one or more other measures. |
|
simple prediction |
prediction of an individual's score on a measure based on the individual's score on another measure. |
|
y-intercept |
the point where the graphed prediction line crosses the y-axis; the value of the predicted score when the predictor score is zero. |
|
define standard error prediction |
the standard deviation for the errors of prediction; an index of the accuracy of a prediction equation. |
|
define cross validation |
method for checking the accuracy of a prediction equation on a second group of individuals similar to the first group. |
|
define multiple prediction |
prediction of an individuals score on a measure based on the individuals scores on several other measures. |
|
define polynomomial regression |
a regression model which does not assume a linear relationship; a curvilenear correlation coefficient is computed. |