## Topic outline

### Topic 1

- Introduction

### Topic 2

**PROJECT GUIDELINES**

PREPARATIONSelecting a topic: Topic should:

-be clear and specific, be feasible to study, be meaningful and enhance understanding of the phenomenonSurvey and questionnaires design:

Keep the questions short, simple and easy. In general, you should keep confidentiality for the respondents; however, for the purpose of this project, you should ask for contact details of the respondents, so the teachers can use them to verify that you have actually carried out the survey.- Using Excel:
- Useful Excel Functions:

Average (to calculate mean), Stdev (to calculate sample standard deviation), Var (to calculate sample variance), Median, Mode, Min, Max, VarP ... WRITING

- Cover page - Table of Contents - List of Tables and Figures:

Main part (6 sections)- 1. Introduction:

(i) describe the research question and its importance

(ii) describe the statistic methods used to answer that question

(iii) very briefly mention the conclusion of the project

- 2. Research methodology:

2.1. Population and sample: Identify the population of interest (eg. all HANU students or FMT students) and a sample

2.2. Questionnaire design: how many questions, the purpose of each question (eg.: to know what students think about the Canteen). Think about how you would analyze the data, and the possible conclusions that could be drawn out of your study.

2.3. Sample size: how

*n*is determined. Take into account the following factors when you decide on a sample size: population size (e.g. if the population has several thousand members, you may need to choose a sample of several hundred, but if the population has only a few hundred members, then the sample size could be several tens), the accuracy of results that you want to achieve (i.e. in a very crucial project, the sample size should be very large so that it can well represent the population), the resources that you have (i.e. human resource, time...), what test you are going to use and its assumptions (some tests may have constraint on the sample size).2.4. Sampling method and data collection:

**Sampling Method (for selecting a sample)**: Describe strategies/ process of sample selection. Choose from the methods that have been introduced in lectures: simple random sampling, systematic random sampling or stratified random sampling. Give reasons to justify your method.**NOTE**: if the sample is not RANDOM, the results may not be accurate, so make sure that you choose a sample RANDOMLY.To choose a sample randomly, you need to have a LIST OF ALL POPULATION MEMBERS. For the Simple Random Sampling Method: you should assign each member in the population a number. As an example, suppose that the population has 200 members, the numbers you assign to the population members may be 1, 2, 3,... 200. After that, use your calculator or Excel to generate a random number. E.g. if you obtain a number 0.050 then multiply this number by 1000, which becomes 50, in this case the population member numbered 50 should be selected. If the number 0.450 is generated, you multiply this number by 1000, and receive the number 450. But in your list, you have only numbers from 1 to 200, in this case you have to generate another random number. Repeat the process until you have selected all sample members.

**Data collection**: describe how you distribute the questionnaires, record and organize the data. Provide a table of the data that have been organized in the Appendix of the report.Describe your strategies/plan in case something unexpected occurs e.g. the person you plan to interview is not available or refuses to cooperate.

2.5. Data processing: use Excel or other softwares (Eviews, SPSS....)? How data is input and processed in Excel. For the purpose of this project, if you don't know how to use Excel, then you have to calculate the necessary statistics and measures by hand.

2.6

: As the main purpose of the project is to carry out a hypothesis test, you should specify an appropriate level of significance. Use only 1 value for significance level.__Significance level of test__ 3.

__Descriptive__Results and Findings:In this section, you provide a descriptive analysis of the data. Draw graphs, charts... Remember to

**describe and analyze the graphs**... Provide summary measures (i.e. mean, median, standard deviation...) as appropriate and**interpret these values**.The emphasis of this section should be on the analysis and discussion of results, not just listing out the facts.

4.

__Results and Findings of the__Hypothesis Test:In this section, you describe the testing process, beginning with restating the question of research.

After that you specify the test you are going to use, and the test assumptions, if any, and demonstrate that these assumptions are met. If the assumption is population is normal, and you have no idea about the shape of the population distribution, then try the following methods:

a. If the sample size is sufficiently large (n at least 30), draw histogram of the sample data (after organizing data into groups). If the histogram resembles a bell-curved distribution then, you can conclude (with some degree of confidence) that the population is approximately normal

b. Use the Chi-squared Goodness of Fit test for Normality to check if the population is normal (optional).

After the assumptions are shown to be met, you continue with formulating Ho, and Ha. Ho, and Ha, will follow from the question you want to answer. Let's discuss the following example: the question we want to answer is "Is the proportion of population greater than 20%" will be translated to: p greater than 0.2. We will test this belief with its opposite: p no more than 0.2. Now you already have two hypotheses, but you should be able to specify which one is Ho, and Ha. First, you need to decide which hypothesis is Ha. Ha is what we want to show statistically, in this case, p > 0.2. This also satisfies the condition that Ha can't contain the equality sign. Ho would be the opposite of Ha: p no more than 0.2 (this satisfies the condition that Ho must contain the equality sign).

The reason Ho must contain the equality sign, is that you have to be able to calculate the test statistic for example Z = (X bar - miu)/sigma/square root of n. If the value of miu (mean of population) is not assumed to be something, you wouldn't be able to convert X bar into Z to compare with the Z critical value to make rejection decision. It can be proved that if Ho is p no more than 0.2, we can write it as Ho: p = 0.2 and this does not affect the result. That means for Ho is p = 0.2, and Ha: p > 0.2, with the same sample, if you reject Ho, then for the hypotheses Ho: p no more than 0.2, and Ha: p greater than 0.2, you will also reject Ho. The same for not rejecting Ho. It is therefore possible to specify Ho as p = 0.2 and the result of test doesn't change.

Now Ho becomes: p = 0.2

Ha become p > 0.2

After that, you continue with the other steps in the testing process...

Remember to make conclusion

**according to Ha**: i.e. there is evidence or no evidence that Ha is true, at the level of significance of... This is because, if you are not able to reject Ho, that doesn't necessarily mean Ho is true. Ho could be true, or false, but we don't know. As you can notice that in the conclusion, we have specified the level of significance, because if you change the level of significance, then the decision may change. E.g. You may reject Ho at level of significance of 0.05 but if you choose a smaller level of significance, say, = 0.01, the decision may become: not rejecting Ho.**It is unfortunate that if we reject Ho, we cannot be absolutely sure that Ha is true**. Rejecting Ho, doesn't mean that we are 100% confident that Ho is false and Ha is true, but it means that because there exists an evidence so strong against Ho, and in favor of Ha, that we**lose confidence in Ho**, and**are inclined to believe that Ha is true**. They have proved that if Ho is true, there is some probability that the hypothesis test will lead to a conclusion of rejecting Ho (this error is called Type I error). And this probability is equal to the level of significance. That is why we need to keep the level of significance small enough.You should also provide a discussion and analysis of the results of the hypothesis test.

5. Project Evaluation:(0.5 page)

Limitations: eg.: small sample, limitation of sampling method used, biases in selecting sample, some survey questions might be confusing or complex, non-responding rate (people refuse to do questionnaire), or any factors that could affect the precision of the answers to the questionnaire. Consider the limitations of the hypothesis test as well (for example if we reject Ho, we can't be 100% sure that Ha is true; also if the test has assumptions, the test for assumptions may not be 100% correct )Implications: how the project benefits some people (eg.: implications for HANU Refectory to improve its food and service)

- 6. Conclusion and Recommendations:

(i) Summarize the findings and interpret them.

(ii) Any further predictions or implications.

(iii) Recommendations.

(iv) Suggestions for further future research.

- References
Appendices

-Collected questionnaires-List of all population members

-Table of organized data

### Topic 3

### Topic 4