In: Statistics and Probability
PLEASE FILL IN BLANKS WITH CORRECT ANSWER. THANK YOU
CHAPTER 21: POSTSCRIPT: WHICH TEST?
Text Review
If you attempting a research project of your own, you will probably be selecting a statistical test from among the ones you have studied so far. The following is a review of some important ideas and a look at some questions you might ask in order to select the appropriate test.
What is the intent of your study? If you wish to summarize existing data, you will be using
(1) ______________________ statistics. If you wish to generalize beyond existing data, you will be using (2)___________________ statistics.
In the behavioral sciences, hypothesis tests are usually preferred to confidence intervals. However, even if you follow this preference, if the null hypothesis is rejected, consider estimating the possible size of the effect by constructing a (3)_________________________________________.
When deciding which test to use, it must be determined whether the observations are quantitative or qualitative. If the data are (4)_________________________, the appropriate hypothesis test should be selected from the various t or F tests or their nonparametric counterparts. When the observations are (5)________________________________, the appropriate test will be the chi-square.
Sometimes the decision as to whether the data are qualitative or quantitative is not so easy. If you are not sure, use the following guidelines to help make your decision. (1) Focus on a single observation. (2) Focus on numerical summaries. (3) Focus on keywords.
Once the nature of the data has been decided, another question to be resolved in selecting the appropriate test is the number of groups. If only one group is involved, then a (6)_________ test for a single population is appropriate. If there are two groups, a t test may also be appropriate, but you must decide whether the groups are matched or independent. If the concern of the researcher is to determine whether paired observations are significantly correlated, then the appropriate test is the t test for a population coefficient. When there are more than two groups, the F test for (7)_______________ is the correct choice. The ANOVA test can be either one-way or (8)_________________.
Finally, the nonparametric tests are to be used when the original observations are (9)_____________ or when some assumption is violated.
What is the intent of your study? If you wish to summarize existing data, you will be using
(1) DESCRIPTIVE statistics.
EXPLANATION: DESCRIPTIVE STATISTICS ARE MEAN MEDIAN MODE AND QUARTILES WHICH HELP US TO UNDERSTAND THE SHAPE OF THE DISTRIBUTION.
If you wish to generalize beyond existing data, you will be using (2) INFERENTIAL statistics.
EXPLANATION: INFERENTIAL STATISTICS DRAW CERTAIN INFERENCE ABOUT THE POPULATION.
In the behavioral sciences, hypothesis tests are usually preferred to confidence intervals. However, even if you follow this preference, if the null hypothesis is rejected, consider estimating the possible size of the effect by constructing a (3) CONFIDENCE INTERVAL
EXPLANATION: BECAUSE IT HELPS TO UNDERSTAND WHETHER THE GIVEN POPULATION VALUE LIES WITHIN IT OR NOT.
When deciding which test to use, it must be determined whether the observations are quantitative or qualitative. If the data are (4)QUALITATIVE , the appropriate hypothesis test should be selected from the various t or F tests or their nonparametric counterparts.
EXPLANATION: FOR QUALITATIVE DATA WE USE parametric test.
When the observations are (5) QUANTITATIVE , the appropriate test will be the chi-square.
Sometimes the decision as to whether the data are qualitative or quantitative is not so easy. If you are not sure, use the following guidelines to help make your decision. (1) Focus on a single observation. (2) Focus on numerical summaries. (3) Focus on keywords.
Once the nature of the data has been decided, another question to be resolved in selecting the appropriate test is the number of groups. If only one group is involved, then a (6) SINGLE SAMPLE test for a single population is appropriate. If there are two groups, a t test may also be appropriate, but you must decide whether the groups are matched or independent. If the concern of the researcher is to determine whether paired observations are significantly correlated, then the appropriate test is the t test for a population coefficient. When there are more than two groups, the F test for (7) ANALYSIS OF VARIANCE is the correct choice.
EXPLANATION: WHEN EVER WE HAVE TO TEST MORE THAN TWO GROUPS THEN ANOVA IS BEST BECAUSE IT REDUCES ERROR.
The ANOVA test can be either one-way or (8) TWO WAY
Finally, the nonparametric tests are to be used when the original observations are NOT NORMALLY DISTRIBUTED or when some assumption is violated.
EXPLANATION:WE USE NON PARAMETRIC DATA WHEN THE DATA IS NORMALLY DISTRIBUTED.