Statistical logic

Emile

Reviewd Question

  1. Stage 1: focusing the study;Stage 2:sampling; Stage 3: setting up statistical decisions; Stage 4:necessary considerations; stage 5:Statistical decision.
  2. Research hypotheses, because the author consider hypotheses show different ways that the researcher evisions the study will turn out. In short, if we may get the so different interpretation of the results.
  3. a null hypothesis is essentially a hypothesis of no relationship in a correlational study.
    alternatve hypotheses: comparing to a null hypothesis, alternative hypotheses state the potential outcomes.
    Both types are considered in the logic of a study. In a study, there should be at least one null hypothesis and one alternative hypothesis for each pair or set of variables being studied.


  4.   random samplestratified random sample
    Advantagesa.The use of random sample eliminates biases in the researcher's choice of subjects.

    b. Each subject must have an equal chance of being delected so that any biases of the researcher are obviated.

    a.The researcher can randomly select from each of the different groups, or strata, in proportion to their occurrence in the population.

    b. This procedure also requires random selection but adds a certain amount of precision to the representativeness of the samplee and allows for the use of the identified characteristics as varaibles.

    DisadvantagesFor a heterogeneous group, purely random sampling might not representatively sample each level or subgroup within the population, particularly when the samples to be usedd are small of the groups to be formed are of unequal sizes.For a homogeneouss group, purely sampling is much easier to do than stratified sampling.

    What is an adeuate sample size?There is no quick and easy answer to this question, but there is little doubt that the large the sample, the better.

  5. The generalizability of a study is the degree to which the results can be said to be meaningful beyond the study. The central question to ask your self in checking generalizability is, What is the largest population that the sample can be said to represent, given the sampling procedures?
  6. The sample is subgorup from the whole population, and the sample can represent the population.
  7. The degree of relationship between two variables;comparing means; comparing frequency.
  8. The important thing to remeber is that siginificance and meaningful are separate issues.
  9. a. If α is < .01 , the experiment is significant.
    b. α is a level, but p is a real data from the experiment.
    c.If α is .01 or less, the researcher can reject the bull hypothesis.
  10. a. The problem, statistical hypotheses, alpha decision level, number of individuals, observed statistic, assumptions, degree of freedom and critical value of the statistic.
    b. (1) Look at the hypothesis.
    (2) Look at the alpha level.
    (3) Compare the observed statistic and the critical value.
    (4) a. If the observed statistic is less than the critical value, it could be due to chance. Therefore, we accept the null hypothesis and stop.
    b. If the observed statistic is greater than the critical value, the probability is .01 or less (alpha level) that it is due to chance. We can , therefore, reject the null hypothesis and continue.
    (5) Decide which alternative hypothesis is more logical.
    (6) Interpret the results in terms of the p level.

Application

A.
  1. According to the steps of "hypothesis testing" in p.121, I examine the example in Chapter 5.
    a. Look at the hypothesis: H1:r>0 and H0:r=0
    b.Look at the alpha level: doesn't give an alpha level
    c.Compare the observed statistic and the critical value:just give the observed statistic (MEAN and SD) , but doesn'y give a critical value.
    d. The null hypothesis or not: because doesn't give a critical value, so the example also lack this step.
    e.Decide which alternative hypothesis is more logical: robs is positive, so, H1 is more logical.
    f.Interpret the results in terms of the p level: he have used two p level(p<.01 and p<.05), so there are 1% or 5% may be null hypothesis, and other is not.
  2. because he does not give us the critical value, we can not find he how to prove the H0 is rejected.
  3. I do not consider the sampling is adequate in the study.I guess the Latin language nations may be generalized.
B. I choose "Standardizing Infra-red Measures of Bone Mineral Crystallinity : an Experimental Approach" by Todd A. Surovell and Mary Stiner as my target article.
  1. the purpose of the article:
        Abstract: They give an paragraph what they want to demonstrate and how to demonstrate it. They want to demonstrate that infra-red spectroscopic measures of bone mineral are partially dependent upon sample preparation method.

        Introduction: They explain why they want to proceed this experiment.Because they find the condition of grinding sample will affect the result in archaeological experiment. In this part, they also cite many literature about archaeological experiments to help explain what the preparation of sampling is the major consern today.

         Method: They proceed three experiments and give method, results and discussion in every experiment part. They mentioned what material they used (a small compact bone fragment od Middle Paleolithic age from Hayonim Cave was ground into a homogeneous powder wuth agate mortar and pestle.)

        Conclusion: They analyze the three results of each experiments in the conclusion parts. Giving their conclusions in the list.

         Future Research: They have an part like "Future Research" which called "Proposed Sample Preparation Method and Standards". They propose some method to help archaeologist can get some replicable results, independent of sample preparer, and with relatively low and consistent levels of error.

        Acknowledgement: They put their acknowledge in the final part to thank some people who help them.

         Reference:They also list the references they used in this article.


  2. According to the steps of "hypothesis testing" in p.121, I examine the target article.
    a. Look at the hypothesis: H1:r>0 and H0:r=0
    ---> Their hypothesis is infra-red spectroscopic measures of bone mineral are partially dependent upon sample preparation method. b.Look at the alpha level: doesn't give an alpha level
    c.Compare the observed statistic and the critical value:just give the observed statistic (MEAN SFand SD) , but doesn'y give a critical value.
    d. The null hypothesis or not: because doesn't give a critical value, so the example also lack this step.
    e.Decide which alternative hypothesis is more logical: robs is positive, so, H1 is more logical.
    f.Interpret the results in terms of the p level: he have used onep level(p<.01 ), so there are 5% may be null hypothesis, and other is not.